Laws doi: 10.3390/laws15030043

Authors: Dustin Gish

Thomas Jefferson, the Author of the Declaration of Independence and the Father of the University of Virginia, considered it a self-evident truth that our rights must be secured through government and that the people themselves are the only safe guardians of their liberty in a republican form of government. The civic education of the people is, therefore, imperative, in his view, if they are to be informed citizens. This article examines the ways that the first States sought to institute public universities, through both constitutional and legislative means, and highlights Jefferson’s vision for civic education against the activity of the States in establishing education. Surveying early State constitutions and university charters reveals, for those States instituting public education, a wide range of approaches, particularly with respect to three aspects: authorizing mode (constitutional or legislative mandates); civic rhetoric; and scope (tiered system or single institution). While several of the States recognize education as important to republican government, their commitments to public civic education vary. Against this backdrop, Jefferson’s views on education appear both comprehensive and constant, from his reform Bill for the More General Diffusion of Knowledge and Notes on the State of Virginia, which envision a three-tiered public system, to his efforts in retirement to pass education reform and establish a new university, with his purpose being explicitly civic. While his State never adopted his full system, Jefferson continued to advocate for ward republics and public instruction throughout his life. The founding of the University of Virginia in 1819 partially fulfilled this pursuit, embodying the keystone in his educational architecture. Yet Jefferson’s broader system—grounded in local participation and universal civic instruction—remained unrealized. This survey further reveals that statesmen in early America did not always agree with Jefferson that States must have an enduring institutional commitment to public civic education, as the best means to inform the people and to secure republican self-government.

Laws doi: 10.3390/laws15030042

Authors: Jungmi Bang

This study conducts a comparative legal analysis of South Korea’s Framework Act on Artificial Intelligence (enacted January 2025, effective January 2026) and the EU AI Act (effective August 2024), focusing on the structural implications of their divergent regulatory philosophies for sustainable digital governance. Employing legal interpretive analysis (textual, systematic, and teleological) and comparative legal methodology, supplemented by risk-based regulation theory and the theory of hardening of soft norms, this paper examines three interconnected dimensions: the conceptual distinction between “high-impact” and “high-risk” AI, the legal nature of self-regulatory structures, and the potential distortion of civil liability attribution. The analysis reveals that Korea’s adoption of the “high-impact” concept, while strategically reducing compliance costs and avoiding stigma effects, generates significant legal gaps, including potential violations of the constitutional principle of clarity, a “liability lightning rod” phenomenon transferring responsibility from AI operators to frontline practitioners, and insufficient institutional prerequisites for effective self-regulation. In contrast, the EU’s ex-ante preventive framework provides greater legal certainty through direct enumeration of high-risk sectors and mandatory conformity assessments. Drawing on the growing body of EU AI Act scholarship, this paper proposes a five-step legislative model for dynamic regulatory adjustment tailored to Korea’s constitutional structure, encompassing statutory core criteria, periodic re-evaluation with parliamentary oversight, phased mandatory enforcement, and a presumption of conformity system, thereby offering a co-regulatory framework that balances innovation promotion with fundamental rights protection.

Laws doi: 10.3390/laws15030041

Authors: Gulnara T. Nurbekova Marco Greggi Lyazat K. Tussupova

In the era of global data exchange, banking secrecy is no longer absolute, becoming part of a more transparent tax administration system. International exchange of tax information has necessitated a legal analysis of issues related to tax secrecy and banking secrecy in Kazakhstan. The authors analyse the relationship between banking, tax and official secrecy, as well as international and national mechanisms for protecting confidentiality in the context of growing demands for tax transparency. The article discusses international initiatives, including CRS, FATCA and the Convention on Mutual Administrative Assistance in Tax Matters (OECD), as well as their impact on the legal framework governing financial information in Kazakhstan. Focusing on international standards, the article highlights the lack of legal clarity in Kazakhstani legislation regarding the mechanism for ensuring banking secrecy when transferring information to tax authorities. Measures are proposed to harmonise regulatory acts aimed at ensuring a balance between the confidentiality of taxpayer information and the obligation of banking organisations to assist the tax authority in performing its tax administration tasks, as well as legal certainty in the handling of confidential information.

Laws doi: 10.3390/laws15030040

Authors: Dominik Skauradszun Paula Moffatt

This paper examines whether the brand-new EU Directive harmonising certain aspects of insolvency law (Insolvency III) with its Title III on asset tracing is fit for the digital age and whether it offers adequate tools for the tracing of digital assets, such as crypto-assets under the EU Markets in Crypto-assets Regulation (MiCAR). The study will demonstrate that Title III on asset tracing has been outdated since the inception of the Insolvency III as it has a ‘blind spot’ on tracing digital assets.

Laws doi: 10.3390/laws15030039

Authors: Bolanle Adebola

The wholesale transplantation of foreign insolvency laws with minimal contextual adaptation—rule convergence—can be inimical to development in African states because it displaces the legal capacity through which insolvency systems become institutionally responsive over time. Situated within a transnational reform order shaped by overlapping developmental and market-integration logics, insolvency reform is frequently promoted through global scripts, technical assistance, and benchmarking regimes that reward rule convergence. This order and the discourse that supports it often operate within an implicit economic-growth development paradigm that treats legal development as achievable through the external supply of ‘best practice’ rules supported by enforcement capacity. This paper challenges that view, advancing a systematic conceptualisation of state legal capacity as the evolving institutional ability to formulate, adapt, interpret, implement, enforce and legitimate legal rules in response to societal legal demand. Applying this framework to corporate and insolvency reform trajectories in East and West Africa, the paper shows how rule transplantation produces capacity displacement, undermining endogenous legal development and development more broadly. Insolvency reform must therefore be understood as a project of legal capacity-building rather than of rule importation, enabling African states to act as co-producers in the evolution of global insolvency norms and models.

--

La présente contribution étudie dans quelle mesure la police est autorisée à rechercher des suspects au sein des services d'urgence hospitaliers à la suite d'émeutes. La question centrale porte sur la légalité d'une telle perquisition policière, laquelle n'est admissible que si elle repose sur des soupçons suffisants, s'appuie sur un mandat de perquisition valable et respecte le principe de proportionnalité. Si les perquisitions illégales entraînent en principe l'irrecevabilité des preuves, l'art. 141 al. 2 CPP permet toutefois leur utilisation afin d'élucider les « infractions graves », y compris lorsqu'elles sont obtenues de manière punissable. Selon les auteur·e·s, cette disposition est incompatible avec les principes de l'État de droit et devrait être, de lege ferenda, abrogée.

Laws doi: 10.3390/laws15030038

Authors: Tung-Hu Tsai

The coronavirus (COVID-19) pandemic necessitated unprecedented regulatory responses that enabled rapid therapeutic deployment. The integrated medico-legal framework—comprising the FD&C Act Section 564 (Emergency Use Authorization/EUA), PREP Act (liability immunity), and CICP (injury compensation)—facilitated emergency response while protecting all stakeholders. This normative legal and policy analysis examines nirmatrelvir/ritonavir (Paxlovid) as a case study, integrating emerging pharmacokinetic evidence demonstrating its passage across the blood–brain and blood–placenta barriers. The EUA-PREP-CICP framework achieved notable results: nirmatrelvir/ritonavir’s authorization enabled deployment approximately 1 year after trials began, demonstrating an 89% reduction in the risk of hospitalization or death and potentially preventing thousands of hospitalizations. The PREP Act enabled focused pharmaceutical development and protected frontline healthcare workers during the crisis, though access barriers and transparency concerns remain areas warranting ongoing attention. The CICP provided administrative compensation for qualifying injuries, with acknowledged limitations in filing timelines and causation standards. Pharmacokinetic studies published after authorization revealed biological barrier crossing, representing normal scientific progress through continued investigation. The EUA-PREP-CICP nexus functioned as an integrated system: EUA enabled rapid evidence-based access, PREP immunity facilitated development and deployment, and CICP provided injury remedy. Based on this experience, this study proposes targeted enhancements to further strengthen this framework: systematic post-authorization surveillance timelines, enhanced special population monitoring through registries, modest procedural refinements to CICP, and improved surveillance infrastructure. These evidence-based improvements would build on the framework’s demonstrated strengths, optimizing performance for future emergencies while preserving the essential functions that helped address the COVID-19 pandemic.

Laws doi: 10.3390/laws15030037

Authors: Mohamed Morsi Abdou Ayman M. Zain Othman Aisha Obaid Alqaydi Mahmoud Fayyad

The AI-supported digitisation of bills of lading has become an important requirement for the maritime transport industry, because it accelerates maritime shipping operations and helps avoid the drawbacks of paper bills of lading. This importance prompted the UAE legislator to introduce a legal provision in the new Maritime Law expressly permitting the use of electronic bills of lading. Despite the significance of this legislative step, this study demonstrates that it suffers from regulatory shortcomings; accordingly, the study aims to bridge the legal gap arising from the deficiency and ambiguity that characterise the rules governing the use of electronic bills of lading. This research fills a gap in the legal literature, as the digitisation of bills of lading under the new UAE Maritime Law has not been deeply explored. It also examines the role of artificial intelligence as an auxiliary instrument in enhancing the efficiency and reliability of this digital transformation. The research adopts an inductive and analytical approach to the provisions of the Maritime Law and related legislation to extract the general legal principles governing dealings in electronic bills of lading. The study shows that the digitisation of maritime bills of lading raises several legal issues resulting from their subjection to more than one legal regime, which may lead to legislative conflict and divergence in judicial approaches. The study concludes that the effective use of electronic bills of lading requires issuance of implementing regulations that explicitly clarify the conditions for their issuance, recognising their possession and electronic negotiability.

This Article documents an emerging practice that has dominated the second Trump administration's approach to unfavorable court rulings, a practice this Article calls legalistic noncompliance. Legalistic noncompliance occurs when the administration uses the language of the law as cover to claim that it is complying with court orders when in fact it is not. Drawing strength from a kind of casuistry, practitioners of legalistic noncompliance deploy an array of specious legal arguments and legalisms in an attempt to conceal what is actually widespread resistance to judicial oversight.

In addition to identifying legalistic noncompliance as an emerging phenomenon, this Article explores its logic and analyzes some of its dangers. This Article suggests that legalistic noncompliance lowers the public salience of noncompliance while, at the same time, signaling to both judges and officials in the executive branch that the administration is not complying with court orders. That dynamic, in turn, could lay the groundwork for outright noncompliance or legalized noncompliance—where courts craft rules that allow judges themselves to say that the administration is complying with court orders when, in reality, the administration is doing little to nothing to carry out a court's ruling.

This Article does not purport to establish a standard for when courts should reject legalistic noncompliance by calling it out as a form of noncompliance. The politics of the moment, more than anything else, may determine if they do. This Article instead aims to surface the developing practice and explore its costs, which have thus far escaped full consideration.

This Article proceeds on the theory that, nevertheless, the Court did not intend to overthrow the government. It offers alternative readings of Arthrex—narrow and broad—to show how the case undermines the existing law of agency design. It then shows how a middle road could reconcile presidentialism with the Constitution, statutory law, and bureaucracy, enabling the Court to advance presidential administration without abandoning traditional principles of administrative law.

This alternative reads Arthrex through the lens of Myers v. United States and its foundational distinction between politics and administration. The approach should be congenial to presidentialists, for whom Myers remains a touchstone. And it offers a principled way to distinguish cases where the president may exercise control of the bureaucracy to realize policy goals from those where the law may appropriately limit the president to mere supervision in the name of good administration.

We sound a note of caution, however: Without care, Arthrex's theory of presidentialism could be far more transformative of the administrative state than many, including perhaps the Justices, are aware.

Healthcare professionals who work in states with criminal abortion bans face an impossible scenario that scholars have termed the “double abortion bind”: the impossible choice between, on the one hand, providing patients with emergency lifesaving abortions and facing potential criminal liability for violating their state’s abortion ban and, on the other hand, not providing emergency abortions and potentially losing their patients and facing medical malpractice claims for violating the standard of care. Litigators nationwide are fighting in state courts to clarify the medical exceptions in abortion bans. While legal scholars have explored criminal liability and the various implications of criminal abortion bans for patients and medical professionals, the civil liability component—including that of medical malpractice lawsuits—remains murkier, in part due to a dearth of such cases. But as more and more patients suffer and die while waiting for emergency medical care or are denied such care altogether, civil cases are beginning to emerge.

This Note asks what role tort claims—specifically, medical malpractice suits—can play in the area of emergency abortion access. Can torts realistically provide useful legal recourse for impacted patients and apply new pressure on healthcare industry actors to widen emergency abortion access? This Note analyzes whether such claims could be used to pressure hospital systems into broadening emergency access to lifesaving abortion care and sketches one potential path forward.

†           The title “Lifesaving Care, Denied” is inspired by the work of If/When/How, specifically, the “Self-Care, Criminalized: August 2022 Preliminary Findings Report,” by Laura Huss, Farah Diaz-Tello, and Goleen Samari. If/When/How: Lawyering for Reprod. Just., Self-Care, Criminalized: August 2022 Preliminary Findings (2022), https://ifwhenhow.org/wp-content/uploads/2023/06/22_08_SMA-Criminalization-Research-Preliminary-Release-Findings-Brief_FINAL.pdf [https://perma.cc/U72M-A49Y].

Table of Contents Show

Introduction

In August 2022, Texas resident Amanda Zurawski was seventeen weeks pregnant.[4] After undergoing eighteen months of fertility treatment, she and her husband were “beyond thrilled.”[5] The couple had decided to name their daughter Willow.[6] On August 23, 2022, Zurawski contacted her obstetrician after experiencing unexpected pregnancy-related symptoms.[7] She was diagnosed with a cervical condition and was told that the loss of her pregnancy was inevitable.[8] Zurawski was sent home, devastated from the news, and her water broke that night.[9] She rushed to the emergency room.[10] Zurawski had experienced previable preterm premature rupture of membranes (PPROM).[11] Because Zurawski was only seventeen weeks pregnant, even if she had gone into labor, the fetus would not have survived.[12]

Abortions in Texas were criminally prohibited at the time and remain so today. At this point in her treatment, and in a different state where abortion was legal, Zurawski would have been offered an abortion.[13] But because of Texas’s abortion laws—which had taken effect a mere two days before Zurawski’s water broke—her healthcare team was terrified that providing her an abortion without signs of acute infection could run afoul of the state’s legal exceptions for a lifesaving abortion.[14] Zurawski was left with no other choice but to wait until she delivered a preterm, nonviable fetus—at the risk of developing a life-threatening infection along the way. Her doctors felt that they could not intervene as long as a fetal heartbeat was present; they would need to wait until she became “sick enough for the ethics board at the hospital to consider [her] life at risk.”[15]

After her doctors had her wait in the hospital overnight hoping she would go into labor on her own, Zurawski was sent home with instructions to monitor herself for signs of infection.[16] After three days of physical and psychological agony, she took a turn for the worse.[17] When she developed a raging fever, dangerously low blood pressure, and became nonresponsive, her husband rushed her to the hospital.[18] She had developed sepsis and was admitted to the Intensive Care Unit (ICU).[19] She delivered Willow, who inevitably did not survive.[20] Zurawski remained in the ICU for three days, battling infections and septic shock.[21] Though she survived, the infections left Zurawski with such severe scar tissue in her uterus and reproductive organs that one of her fallopian tubes remains permanently closed—an outcome that would likely have been avoided had she received a timely abortion.[22] Zurawski longed to have children, and it is now unclear whether she will ever be able to become pregnant again. Unfortunately, Zurawski is not alone: PPROM occurs in approximately 2 to 3 percent of pregnancies in the United States.[23]

Zurawski went on to become the lead plaintiff in the Center for Reproductive Rights’ case against the State of Texas, which asked Texas to clarify the scope of “medical emergency” exceptions under its abortion bans. In May 2024, the Texas Supreme Court unanimously ruled against Zurawski and her nearly twenty co-plaintiffs, holding that the medical exceptions in Texas’s law were broad enough to withstand constitutional challenge and refusing to clarify them.[24]

Nationwide, people facing pregnancy complications who live in (or travel through) states where abortion is prohibited by criminal statutes (banned states) risk permanent injury or death as a direct result of abortion bans.[25] While anti-abortion lawmakers claim that the medical emergency exceptions included in abortion bans (ban statutes) are more than sufficient to protect pregnant individuals,[26] the numerous lawsuits demanding clarification of those exceptions tell a different story.[27] Maternal mortality rates were already rising in the United States before the Dobbs decision.[28] Meanwhile, abortion bans only exacerbate death rates among pregnant people.[29] Moreover, maternal mortality statistics reflect deep inequities nationwide: Outcomes for Black and Native American people are consistently far worse than those of their white counterparts.[30] While many of the plaintiffs in Zurawski ultimately had the resources to travel out of state and receive care, those who remain the most impacted by abortion bans are marginalized communities: Black, Brown, and Native people; low-income individuals; and undocumented people.[31] These groups bear the brunt of the violence that abortion bans inflict, and the stories of these pregnant people are also the least likely to be told in mainstream media. While a handful of tales make national headlines, too many other stories fall through the cracks due to the enormous disparities along lines of race, class, and immigration status. Many of those stories only make their way to mainstream media because of dedicated investigative journalists.[32]

The risks that pregnant people face in banned states are only amplified for pregnant members of marginalized communities. For many of these individuals, access to any kind of healthcare is already often shaky due to a myriad of factors, including lack of health insurance[33] and resulting lack of access to prenatal care,[34] fear of deportation,[35] the need to work several jobs to make ends meet[36] without paid leave,[37] and lack of access to adequate transportation[38] or childcare.[39] Consider the story of Josseli Barnica, a Houston resident and immigrant from Honduras. After miscarrying her second child, Barnica languished in a hospital bed for an agonizing forty hours because doctors refused to provide her with lifesaving care until no fetal heartbeat was detected.[40] She subsequently developed sepsis and died.[41] Her husband, an immigrant from El Salvador, is now tasked with raising the couple’s four-year-old daughter with the help of his late wife’s younger brother.[42]

While the harms that Zurawski suffered due to abortion bans remain horrific and by no means merit minimizing, she is a white woman who had the means to pay for costly fertility treatments and who became a nationally recognized figure after Time named her one of its 2024 Women of the Year.[43] She was also featured prominently in a documentary titled Zurawski v Texas, whose executive producers included the likes of Jennifer Lawrence and Hillary and Chelsea Clinton.[44] Most of the plaintiffs in her case appear to be white women who have American citizenship or valid immigration status, which is of course no accident: Thrusting oneself into the public sphere in a state as hostile to abortion as Texas quite nearly requires having such protections.[45] Yet while Zurawski attends Time’s Women of the Year gala, Barnica’s husband is quietly left to raise a child on his own while he works twelve-hour shifts at a construction site.[46]

While the Zurawski case and its counterparts represent important battles in the fight to save the lives of pregnant people and prevent the needless suffering of future patients, such cases take time to move through the courts and come with no guarantee of success.[47] Though they serve an important function within the court of public opinion, considering alternate routes of litigation is also key—especially as a nationwide assault on basic reproductive healthcare continues.[48]

As access to abortion care,[49] birth control,[50] and comprehensive sexual education[51] diminishes, outcomes for pregnant people in states with existing restrictions grow increasingly dire by the minute.[52] In Texas, after the state banned abortion in 2021, the rate of sepsis for women hospitalized for pregnancy loss in their second trimester spiked by over 50 percent and “dozens more pregnant and postpartum women died in Texas hospitals than had in pre-pandemic years.”[53] While the national maternal mortality rate declined, Texas’s rose substantially.[54] This comes as no surprise; sepsis-linked maternal deaths are a hallmark consequence of abortion criminalization worldwide.[55] And even though many pregnant individuals who have suffered delayed or denied abortions do not end up dead, their injuries are staggering: severe illness with the risk of serious infection;[56] fertility loss from permanent reproductive scarring and loss of reproductive organs;[57] or lasting trauma from, say, being told to wait in a hospital parking lot while sick and bleeding.[58] Given the gravity of this plainly preventable public health emergency, it is imperative that reproductive health advocates consider the potential merits of alternative approaches to litigation. The legal strategies utilized by traditional pro-choice organizations, such as federal impact litigation, may no longer be equipped for the post-Dobbs crisis in American reproductive healthcare.[59]

While tort claims have not traditionally been a weapon in the arsenal of the pro-choice movement,[60] would-be plaintiffs in emergency abortion situations have stories akin to those of traditional medical malpractice cases. Plaintiffs in the most severe instances, like Barnica’s case, may have legitimate wrongful death claims.[61] Within the healthcare space as a whole, medical professionals are clearly grappling with the possibilities of medical malpractice suits related to abortion bans, but medical malpractice insurers have remained largely silent regarding coverage for physician liability stemming from delayed and denied emergency abortions.[62]

Moreover, medical malpractice claims are only just beginning to emerge[63] from the numerous cases of pregnant people nationwide being turned away from lifesaving abortion care.[64] At the time of this Note’s writing, three patient plaintiffs who were forced to suffer denials or delays of lifesaving emergency abortions—Anna Nusslock, Rachel Harrison,[65] and a Jane Roe—have pending lawsuits in California against the hospitals and staff from whom they sought treatment.[66] While all three complaints allege intentional infliction of emotional distress and negligent infliction of emotional distress, only one alleges medical malpractice.[67] The defendants in all three cases are Catholic-affiliated hospital systems, which must adhere to the Ethical and Religious Directives for Catholic Health Services developed by the United States Conference of Catholic Bishops.[68] Those directives state: “Every procedure whose sole immediate effect is the termination of pregnancy before viability is an abortion . . . . Catholic health care institutions are not to provide abortion services . . . .”[69] Litigation is ongoing, but it is likely that more cases of this nature will follow.

This Note explores whether plaintiffs can successfully use tort claims to seek compensation for both physical and emotional harm and to pressure hospital systems, insurance companies, and policymakers to broaden emergency access to lifesaving abortion care. Part I provides an overview of the U.S. medical malpractice landscape generally, including the requirements plaintiffs must meet to successfully bring a medical malpractice suit and the particular challenges and opportunities presented by tort claims relating to lifesaving emergency abortions. Part II explores potential routes for successful tort litigation that injured patients like Zurawski might employ, focusing on the theory of corporate negligence to target institutional healthcare players with resources and power rather than individual physicians caught in the crosshairs. Part III discusses potential challenges with and counterarguments against using tort claims and corporate negligence doctrine as litigation tools. Finally, Part IV explores this proposed legal strategy in the context of both the anti-abortion movement and the reproductive justice movement. This Note concludes that combining medical malpractice torts with a theory of corporate negligence is the strongest path for signaling to hospitals and lawmakers that abortion bans only serve to put pregnant people in harm’s way and threaten the doctor-patient relationship. Litigators interested in abortion access advocacy should pursue creative legal strategies in this space; innovative lawyering is needed now more than ever.

I. U.S. Medical Malpractice Claims & Emergency Abortion Access

A. Medical Malpractice, Generally

U.S medical malpractice law originates from English common law.[70] This field of law offers recourse to patients who have suffered injuries due to negligent medical care. However, medical malpractice suits are contentious: For many in the healthcare field, medical malpractice conjures mental images of plaintiffs filing frivolous claims against physicians who practice medicine in a heavily litigious country.[71] But, at its core, the overarching social goal of medical malpractice remains the improvement of the medical field through (1) deterrence of unsafe practices; (2) compensation of patients injured through negligence; and (3) exacting corrective justice.

The heart of U.S. medical malpractice law is the tort of negligence.[72] Medical negligence claims are defined by a physician’s failure to provide a patient with the degree of care that a reasonable physician in the same position would have afforded the same patient.[73] Medical malpractice law exists in contrast to the nonpunitive patient safety movement, and yet both seek to better the experiences of individuals in the American healthcare system. On one hand, trial attorneys believe that the threat of litigation increases patient safety by creating incentives for healthcare organizations to pay attention to the quality of the healthcare they provide.[74] On the other hand, the patient safety movement espouses a “nonpunitive, systems-oriented, cooperative” approach to advocating for patients.[75] The patient safety movement, which began gaining traction in the late 1990s, advocates for change within the healthcare field with the goal of achieving zero patient deaths caused by preventable patient harm each year.[76] The movement’s values include creating a culture of safety and transparency. It uses a “holistic approach,” working with a variety of stakeholders (such as doctors, patients, and policymakers) to shift practices towards streamlined processes that use evidence-based practice blueprints to help minimize medical errors.[77]

Despite tensions between medical malpractice law and the patient safety movement, the notion of “preventable” error remains central to both: The patient safety movement seeks to prevent harms before they happen, and medical malpractice law seeks to punish and obtain compensation for harms that have already occurred. Yet although the concepts of human error or lack of good judgment lie at the heart of both approaches, the reputation of medical malpractice among healthcare professionals remains generally negative: Many doctors, already stretched thin, are (understandably) hostile to the added burdens of lawsuits and second-guessing of their professional judgment.[78] In considering whether to introduce medical malpractice into the area of lifesaving abortions, it is impossible to avoid the clash between these two schools of thought, in part because of the real impact of both on healthcare professionals and patients.

Proponents of medical malpractice litigation argue that it is a critical means of redress for the injured patient who seeks compensation and offers the ability to hold healthcare professionals accountable for their errors or lack of care.[79] However, some of the field’s critics argue that medical malpractice litigation only incentivizes “secrecy and blame” and criticize its approach as “punitive, individualistic, [and] adversarial.”[80]

Despite these critiques, however, medical malpractice is an unavoidable feature of the healthcare field today. While medical malpractice claims remain rare on the patient side—one 1991 study found that less than 2 percent of individuals harmed by a negligent medical error actually pursued malpractice claims[81]—researchers estimate that around a third of physicians have had a malpractice claim filed against them at least once in their careers.[82] Certain specialties—particularly those involved in crisis management—face more litigation than their peers. For example, more than 75 percent of emergency department physicians will be named in a medical malpractice suit at least once by the end of their careers;[83] OB-GYNs and surgeons are the two categories of specialists most likely to be sued.[84]

The higher likelihood of medical malpractice suits among reproductive health providers compared to other medical professions is likely linked to the longstanding dangers of being pregnant in the United States,[85] which existed well before abortion bans exacerbated those risks and narrowed access to reproductive care. The United States has historically had one of the highest—if not the highest—maternal mortality rates among wealthy countries: In 2022, the United States had approximately 22 maternal deaths for every 100,000 live births, while for the same number of live births, Switzerland had approximately 1 death, and Sweden, the Netherlands, Japan, Germany, and Australia all had between 2 and 4 deaths.[86] Race is also a crucial component of deadly outcomes for pregnant Americans: The rate for Black maternal mortality in the United States in 2022 was a staggering 49.5 deaths for every 100,000 live births, while the corresponding white maternal mortality rate was 19 deaths.[87]

Traditionally, medical malpractice litigation has been left within the regulatory authority of individual states and has been governed by state tort statutes.[88] While medical malpractice claims have existed since the 1800s, they began to surge in the 1960s, likely because of new treatments that bore higher risks for patients, fewer barriers to lawsuits, and the removal of rules that previously shielded charitable institutions from suit.[89] This, in turn, prompted lobbying by healthcare groups to reform the medical malpractice system as a whole.[90] As a result, the landscape of medical malpractice torts varies considerably state by state. According to some analyses, the states that spend the most on medical malpractice cases are New York, California, Pennsylvania, Florida, and New Jersey.[91]

B. Medical Malpractice Claims in Pre-Roe v. Wade Abortions

While Dobbs poses novel issues for the intersection of medical malpractice and emergency abortion care, medical malpractice claims have long existed in the context of negligently performed abortions.[92] Pre-Roe, some women and their families brought cases against those who provided them with abortions, despite the fact that the women themselves could be prosecuted for having the abortions in certain states.[93] Most of these claims were against practicing physicians.

Women and their relatives expected not only competent care when they engaged an abortionist, but, if their abortion provider proved to be unskilled, they assumed they had recourse in the courts. Their victories in court show they were right. Juries granted awards ranging from $200 to $8,000. The average award in these abortion suits was $3,250, comparable to awards granted in other medical malpractice cases.[94]

The scope of monetary awards for abortion cases varied, but plaintiffs often sought recovery of funeral expenses in fatal cases and reimbursement of emergency medical care costs.[95] Courts split into two camps on such cases. Some concluded that women who had received “illegal” abortions were victims and could therefore seek damages from the doctors who had performed the procedures, while other courts treated those women as accomplices who had consented to a crime and therefore held that they were not entitled to bring such claims.[96]

American history and public health scholar Leslie Reagan offers the following commentary on these pre-Roe,abortion-related medical malpractice cases, which can also shed light on the social norms that underlie health-related tort claims today: “[T]he willingness of Americans to sue doctors for badly performed illegal abortions should be read as an indication of Americans’ high expectations of doctors.”[97] Arguably, due to advances in science and technology, public health standards have only improved since the late nineteenth and early twentieth centuries[98] when these cases were filed, and it would be logical for Americans’ high expectations to have increased, or at least to have stayed the same. Pregnant patients today who seek to carry their pregnancies to term suffer similar harms as did pre-Roe plaintiffs. The difference between the plight of pre-Roe plaintiffs and that of modern patients is that today’s harms stem from delayed or denied lifesaving abortions, rather than exclusively those abortions performed upon request to terminate an unwanted pregnancy. As a result, it’s reasonable to infer that healthcare providers today are even more likely to face liability for harms related to abortion-related care than they were in the past.

C. Today: The “Abortion Double Bind”

Today, the practice of “defensive medicine” is a well-established phenomenon among physicians in which healthcare providers act primarily based on a fear of litigation rather than to treat or diagnose a patient.[99] Defensive medicine emerged as a safeguard against potential malpractice suits but has faced criticism as a practice that merely drives up healthcare costs without improving the quality of patient care.[100] One common defensive medicine practice is over-testing: doctors ordering excessive tests and diagnostics for illnesses or conditions that they suspect are unlikely in an effort to protect themselves in a hypothetical future malpractice case through the defense that they did everything right in treating a patient.[101] Defensive medicine can extend to surgeries as well: In the OB-GYN space, high rates of caesarean sections are often considered a signal of a defensive practice.[102]

The situation faced by healthcare professionals providing emergency abortions in banned states adds a twist to the traditional perception of physicians practicing defensive medicine. These physicians find themselves in what health law scholar Dov Fox calls the “abortion double bind,” in which a doctor is caught between potential felony charges if they perform the emergency abortion and potential medical malpractice liability if they do not.[103]Doctors in banned states who are at the front lines of emergency medical situations involving pregnancy thus have an incredibly difficult set of circumstances to consider. In Mississippi, performing an illegal abortion means a criminal penalty of one to ten years in state prison, along with license revocation for nurses and doctors.[104] In Texas, individuals who perform abortions face up to life in prison as well as a $100,000 fine.[105] Doctors, understandably, are terrified: Their own livelihoods and freedom hang in the balance in addition to the lives of their patients.[106] In addition to the heavy responsibility of providing patient care, physicians are now also forced to think about whether doing their jobs will imperil their ability to continue to do their jobs and care for their families.[107] As a result, many physicians have begun to simply stop providing all care that could potentially result in the termination of a pregnancy—often despite assurances from politicians and prosecutors that state authorities will not go after doctors.[108]

The choice at hand for doctors appears to be either (A) uproot their lives and leave the states where they currently practice so that they can work in states where abortion is legal, (B) provide emergency abortion services and risk the ends of their careers as well as potential legal nightmares with no guaranteed end in sight, or (C) wait to provide their patients with care until those patients are truly circling the drain. To the great detriment of pregnant Americans, the barrage of patient stories that this Note discusses suggests that too many physicians are choosing option (C), even in states where they receive assurances that they will not be prosecuted.

While some state government leaders have taken steps to indicate that they will not enforce abortion bans through prosecutions, the climate of fear remains. In 2024, after Arizona courts upheld an 1864 law that bans nearly all abortions, Arizona Attorney General Kris Mayes publicly stated that she would not prosecute any doctors for performing abortions.[109] Despite this, many Arizona physicians simply do not want to take any chances and are opting to not perform the procedure.[110] The stress and risks are also prompting doctors to take drastic measures to ensure their own peace of mind: A 2024 study of OB-GYNs in thirteen states with abortion bans indicated that 60 percent of physicians are thinking about leaving the states in which they currently work and 11 percent have already left and moved their practices to states with protections for abortion.[111]

This climate of fear extends to nurses and other hospital staff. A 2022 survey from Texas that analyzed the impact of Senate Bill 8 (SB8)—the first civil bounty-hunter-style abortion ban,[112] which Texas passed in September 2021—indicated that many physicians with training in dilation and evacuation (D&E)[113] were unable to offer the procedure to patients even for abortions that were permitted under SB8, because nurses and anesthesiologists were concerned “about being seen as ‘aiding and abetting’” abortions and therefore refused to participate.[114] Therefore, even if a physician was willing to perform a dilation and curettage (D&C)[115] or a D&E, the supporting staff required for the procedure may not have been. As a result, some doctors relied on alternate methods to get care for their patients, such as labor induction or even hysterotomy, an incision made into the uterus, “because it might not be construed as an abortion.”[116] However, these procedures carry different risks and are not standard clinical protocol for abortions: Hysterotomy carries increased risk of complications (compared with a D&E) and risks for future pregnancies, just as induction does.[117] The bottom line here is that patients suffer and the standard of care rolls backwards: One OB-GYN described these practices as going “back to doing what they used to do before there was a D&E provider in town.”[118]

While both scholars and members of the medical community have called upon physicians to engage in “medical disobedience”[119] by continuing to provide abortion care, this remains a tall order. The post-Dobbs physicians of today are being compared to some of the pre-Roe doctors who stood up to laws criminalizing abortion, facing arrest and public harassment for openly violating abortion laws.[120] Legal defense funds have opened up to help physicians nationwide who face prosecution.[121] But the chilling effect of abortion bans is expansive, and while some brave doctors and nurses are out there openly flouting them,[122] many healthcare workers remain understandably fearful. One OB-GYN based in Indiana told NPR:

There is no way that I would risk my personal freedom and jail time for providing medical care . . . . I would love to show my children that I am brave in the world, but our society will not allow me to be a civil-disobedient citizen in the way that some of these articles suggest, because I would be imprisoned, I would be fined, I would lose my license and I very well could be assassinated for doing that work.[123]

The specter of criminal liability in the abortion double bind looms so large for doctors that the scales tip in favor of the self-protection of individual healthcare workers and the medical system, and not in the direction of the lifesaving healthcare that patients deserve and so desperately need.[124] As a result, patients pay the price.

While physicians continue to face horrific risks, including criminal liability, for providing basic abortion care, laws limiting emergency access to abortion remain deeply unpopular even with conservative Americans. A recent poll showed that 86 percent of adult Americans support protecting access to abortions for people suffering from pregnancy-related emergencies such as miscarriages.[125] As a result, while bans remain on the books, states feeling public backlash have taken steps to expand such so-called protections—all without paring down their overall criminal prohibitions. These steps, however, have been largely useless. Texas, for instance, instituted an affirmative defense in criminal cases for physicians and pharmacists who perform abortions in cases of ectopic pregnancy or PPROM.[126] However, creating affirmative defenses will likely do little to mitigate the overall chilling effect and culture of intense fear that state abortion bans have sown among healthcare professionals because they still place the onus on physicians to defend themselves after being charged.[127] The clear result of states attempting to “work around” the double bind that abortion bans create is that physicians nationwide continue to fear for their medical licenses, livelihoods, and families if they provide emergency abortions.[128]

D. Who Decides? What Post-Dobbs Patient Experiences Tell Us

What’s happening on the ground in hospital emergency rooms indicates chaos: Doctors, hospital administrators, and legal counsel have been left to scramble to make decisions on a case-by-case basis, frequently on compressed timelines, often leading to horrific outcomes for patients. Because the data is still emerging and because every state ban is different, the situation varies from emergency room to emergency room. The fact patterns emerging from different accounts give us clues about how healthcare systems are operating when it comes to emergency abortions and offer the best insight into how medical staff and administrators are making decisions around who gets care and who doesn’t.

Ultimately, the eerie similarities in so many stories of patients who were either denied care or made to suffer life-threatening delays draw a clear throughline that points to decision-making failures at the top of hospital administrative structures. Because information about how hospital administrations come to these decisions is not public, laying out the internal process leading to circumstances on the ground is difficult and beyond the scope of this Note. However, the anecdotes discussed in this Section suggest that, in cases where hospital staff tell patients that they cannot provide them care, it could be because (A) the hospital has provided no guidance to its employees about state abortion bans; (B) the hospital has instructed staff to intervene only if a patient is imminently close to death; or (C) a hospital has given employees guidance and a promise of protection from liability in an effort to widen access to care, but staff still remain fearful of prosecution and license loss and therefore choose to exercise restraint regardless.

There is no uniform approach to handling medical exception abortion cases that emerges across anecdotal evidence—let alone a standard procedure that is founded upon the goal of saving pregnant people’s lives.

1. Emergency Room Staff

Jaci Statton, a stay-at-home mother of three, was diagnosed with a molar pregnancy[129] with cancerous tissue that required a D&C after coming close to passing out in her kitchen and realizing that her pants were soaked with blood.[130] She was transferred from hospital to hospital in Oklahoma in search of the care she needed; no one would offer her the procedure despite the risks to her life—which all the clinicians she saw recognized.[131] The staff at one hospital told her they could not perform the procedure because it was a Catholic hospital; at another, the University of Oklahoma Medical Center, an ultrasound tech objected to a D&C because he detected fetal cardiac activity.[132] At a third hospital, staff told her she could wait in the parking lot until her condition worsened enough to be treated with an abortion because, in their words, “[w]e can’t touch you unless you’re like crashing in front of us.”[133] They assured her that at that point, staff would be “ready to help [her].”[134] It is unclear whether this policy was the result of individual physicians’ readings of the state’s abortion ban and medical exemptions or the result of a more official hospital policy that staff were enforcing. Statton eventually traveled out of state to get the abortion she needed, and she ultimately chose to get a tubal ligation to prevent future pregnancies after her harrowing experience.[135] After the ordeal, she began taking antidepressants for the first time.[136]

When one of Zurawski’s Texas coplaintiffs, Anna Zargarian, was diagnosed with PPROM, emergency room doctors informed her that they could not offer the D&E they believed to be the safest treatment because doing so would violate Texas abortion bans since fetal cardiac activity was still detectable.[137] Induction was out of the question, and doctors merely typed a generic abortion-finder resource into Zargarian’s phone and offered to admit her for “expectant management,” which essentially meant that she would either wait to go into labor naturally or deteriorate enough so that the hospital could then treat her.[138] Zargarian went home and eventually made the decision to travel to Colorado—despite running the risk of going into septic shock or labor during her trip.[139]

Just as Zargarian was forced to travel to Colorado to get care, so too must patients in other states. In a three month period, six pregnant patients in Idaho—a state known for litigation relating to its abortion ban[140]—were airlifted to out-of-state hospitals in order to receive the stabilizing abortion care they needed.[141] This figure, if annualized, could amount to twenty patients per year needing to be airlifted for out-of-state care.[142] One Idaho physician explained: “When the guessing game gets too uncomfortable, we transfer the patients out at a very high cost to another state where the doctors are allowed to practice medicine.”[143] Indeed, airlifting a single patient hundreds of miles away to Oregon or Washington can cost tens of thousands of dollars.[144] In 2024, some Idaho physicians started advising their pregnant patients—or those trying to become pregnant—to buy memberships with critical care transport companies to avoid facing massive costs in case they end up requiring emergency transport to another state for an abortion.[145] Private insurance is a requirement for membership.[146] However, those who do not have the resources to pay for this service are forced to return home and wait until they get sicker.[147]

2. Hospital Lawyers

The case of Mylissa Farmer shows the extent to which hospital legal departments exert control over the care that pregnant patients so desperately need in emergencies. Like so many others, Farmer experienced PPROM a mere eighteen weeks into her pregnancy.[148] She sought abortion care in two states: Kansas and Missouri. Hospital lawyers at two separate institutions determined that they would not treat her, leaving her, like Statton and Zargarian, to travel elsewhere in order to get care.[149] At Freeman Hospital West’s emergency room in Missouri, Farmer was informed that the hospital’s “legal department refused to allow its doctors to intervene.”[150] The legal department cited the state’s abortion ban as the basis for its decision, despite doctors knowing that this decision threatened Farmer’s life. Freeman physicians informed Farmer that she could either remain in the hospital on IV antibiotics and wait for her body to start laboring on its own or leave.[151] Freeman doctors also later informed Farmer that she was not the first woman in similar circumstances they had turned away and expressed fear that “women like her ‘would die’ as a result.”[152]

One of the physicians treating Farmer advised her to go to an emergency room outside the state, prompting Farmer to drive to Kansas, where she was denied care by a second hospital’s lawyers. At the University of Kansas Hospital, a physician examined Farmer and recommended a surgical procedure that would “resemble an abortion.”[153] Farmer elected for labor induction rather than the surgery, in part because it would give her a chance to hold and say goodbye to her daughter.[154] For a moment, it seemed like Farmer would finally get the care she needed. Twenty minutes later, the same doctor returned and informed Farmer that the hospital legal counsel had decided that “despite [the doctor’s] assessment that Ms. Farmer needed prompt care, [the doctor] could not provide Ms. Farmer with any treatment, including inducing labor, because it would be ‘too risky in this heated political environment to intervene.’”[155] Farmer was once again forced to leave.[156] She eventually received the care she needed at a clinic in Illinois, a state where abortion is not banned.[157]

Some hospitals have begun to take more overt steps to insulate themselves from liability. Attorneys at one Missouri hospital emailed physicians notices stating, “You cannot offer abortion. You cannot refer for abortion out of our hospital, because we’re not going to take on that liability and that responsibility.”[158] The hospital directed its doctors to email its legal team if there were concerns that a pregnancy could not continue.[159] One OB-GYN specializing in high-risk pregnancies, Dr. Betsy Wickstrom, was given a 24/7 legal hotline to call in “dire” situations and seek advice on how to proceed.[160] Yet, despite these resources, Dr. Wickstrom’s hands were often tied. Her hospital’s protocols meant that there was simply more red tape she had to navigate during the most time-sensitive moments of high-risk pregnancies, where, according to Dr. Wickstrom, “time is life.”[161] When entering ectopic pregnancy diagnoses into patients’ medical records, “a large red banner pop[ped] up” asking Dr. Wickstrom to confirm the accuracy of the diagnosis and directing her to consult with an attorney if the fetus in such cases had a heartbeat.[162] As a result, Dr. Wickstrom has sometimes needed to refer patients to providers over state lines, in Kansas.[163]

While some healthcare organizations have made clear that their employees will have the organization’s backing when it comes to defending any staff member, this is not necessarily the case at every institution.[164] In some cases, hospitals have left physicians largely in the dark, neglecting to provide any guidance to doctors about how to act when assessing emergency situations involving abortion.[165] One Idaho OB-GYN described how their hospital’s legal counsel refused to meet with them for over two months, and, when they finally met, the lawyer merely regurgitated the law back to them.[166]

3. Physician Specialist Committees

At other hospitals, informal, physician-driven protocols are beginning to emerge, but, while important, they remain a Band-Aid approach to patching a gaping wound. At Vanderbilt University Medical Center in Nashville, an ethics committee must now approve emergency abortions.[167] When questions around providing patients with lifesaving abortions come up, the committee communicates by email, phone, and text message. A group of six OB-GYNs, five of whom are women, formed the committee in fall 2022 to address the crisis they saw unfolding. Their goal was to back each other up in order to serve as many patients as possible while remaining within the confines of Tennessee law.[168] Often, discussions happen by phone, and sometimes the committee pulls in specialists from other departments to discuss specific cases.[169] Abortion requests “can hit the committee’s inbox at any hour,” and while there are at least two requests per month, the committee can get as many as four per week.[170] Urgently life-threatening complications such as cardiac failure and Stage 3 kidney disease are usually signed off quickly via text message, but other cases—such as stable patients with extremely high-risk delivery parameters due to medical history, who face eventual organ loss and potential death in the delivery process,[171] or cases of fetal anomalies, which put pregnant individuals at risk for uterine rupture and potential death[172]—require lengthy deliberation.[173] The answer in those cases is often “no”: Even physicians on the committee have stated they are not “brave enough” to OK a potentially lifesaving procedure, and they frequently urge restraint due to fear that they might be considered “cavalier” and appear like they’re trying to circumvent state law.[174] Moreover, the committee itself—not Vanderbilt as an institution—has developed critical protocols and blanket policies that allow doctors to provide abortion care in cases of PPROM due to the high risk of sepsis.[175] However, the committee only discusses patients whose doctors sound an alarm and reach out to the committee for guidance; many patients whose physicians don’t reach out are silently denied care.[176]

4. The Takeaway: Zero Transparency

While evidence is beginning to support what these individual accounts tell us, on-the-ground accounts remain the primary source for understanding the processes hospitals are utilizing to make decisions about lifesaving abortions.[177] A recent study, in which individuals contacted Oklahoma hospitals while attempting to obtain information about hospital policies around providing abortions in case of pregnancy complications,[178] yielded findings that indicate the complete lack of a cohesive approach across Oklahoma hospitals:[179]

[N]ot a single hospital appeared to be able to articulate clear, consistent policies for emergency obstetric care that supported their clinicians’ ability to make decisions based solely on their clinical judgment and pregnant patients’ stated preferences and needs. At most hospitals, representatives were unable to tell the callers whether there was an approval process or any other guidelines to help clinicians determine that a pregnancy must be terminated to save the life of the patient, and they were similarly unable to describe what miscarriage management services were available to patients. In a few outlier cases, representatives explained that their hospitals had instituted such policies but refused to provide this information, in some cases citing an inability to share material with those outside of the hospital system, and in others offering no explanation.[180]

The study concluded that the following elements all remained deeply unclear: (1) what hospital policies or guidelines were for clinical decision-making in medical emergencies concerning pregnant people; (2) what criteria clinicians used for determining who could receive an emergency abortion; (3) what the administrative oversight mechanisms for approval were; and (4) how miscarriages were to be handled, particularly in cases of fetal cardiac activity.[181] Given the absence of protocols, the potential for staff to turn patients away or force them to wait dangerously long periods of time—both of which constitute breaches of physicians’ ethical responsibilities to their patients—is high. The absence of protocols, however, provides legal cover for hospitals: Protocols that go too far in one direction could risk breaking a strict abortion ban, while protocols that go too far in the other direction could open the door to negligence liability for the institution itself, as this Note will discuss in more detail in Part II.

II. Paths Towards Institutional Liability

While several paths to liability exist—including negligence claims against individual healthcare providers, administrative complaints, and wrongful death claims—institutional tort liability, through negligence claims filed against hospital systems, presents the most promising path forward to pressuring healthcare systems to provide lifesaving emergency abortion care.

A. A Note on EMTALA

One federal law has significant overlap with questions around medical malpractice liability and failure to provide lifesaving emergency abortions: the Emergency Medical Treatment and Labor Act (EMTALA). EMTALA, a Reagan-era federal law, was originally passed in order to prevent “patient dumping,” a practice in which hospitals either refuse to treat unwanted patients—who are generally uninsured, mentally ill, and indigent—or send them away by inappropriately transferring them to other hospitals.[182] EMTALA provides recourse to victims of patient dumping by creating penalties for hospitals and physicians found to have violated the law by failing to stabilize patients presenting with emergency medical conditions.[183] At the time of this Note’s writing, the Supreme Court has yet to directly address whether EMTALA preempts state statutes that do not adequately protect the health of pregnant people, having conspicuously dodged the question in Moyle v. United States.[184]

Much as with medical malpractice claims, patients themselves can bring claims against hospitals and doctors not in compliance with the law. While EMTALA’s narrow scope makes it an imperfect tool for broader abortion-related litigation,[185] at least two post-Dobbs plaintiffs have brought EMTALA claims against hospitals for alleged violations concerning denials of lifesaving abortion care.[186] One such plaintiff, Mylissa Farmer (who was repeatedly denied an abortion at multiple emergency rooms after being diagnosed with PPROM, as previously discussed in this Note) was successful in her EMTALA claims against two hospital systems in Kansas and Missouri.[187]

Unfortunately, following Donald Trump’s reelection in 2024 and the appointment of Robert F. Kennedy Jr. as secretary of the Department of Health and Human Services (HHS)[188]—the federal administrative body charged with oversight and enforcement of EMTALA[189]—it is unlikely that HHS will be willing to enforce EMTALA in the proactive ways that the Biden-era HHS administration did.[190] The Trump Department of Justice’s decision to dismiss its challenge to Idaho’s ban on emergency abortion care in Moyle (initiated during the Biden administration) reflects quite plainly the likelihood that HHS will be unwilling to enforce EMTALA.[191] At his confirmation hearing, Secretary Kennedy demonstrated a glaring lack of understanding of EMTALA and of HHS’s authority to enforce the law.[192] As a result, it’s unlikely that future plaintiffs will have the same access to recourse as Farmer, or that plaintiffs’ lawyers will rely on EMTALA as a tool for their clients. New legal tools are therefore more important than ever.

B. Most-Applicable Types of Tort Claims

1. Medical Malpractice

For a plaintiff to successfully bring a medical malpractice claim, they must show that they received substandard medical care by proving four elements: (1) a duty that was owed to the patient by the healthcare professional (or institution), (2) breach of that duty, (3) causation, and (4) damages suffered.[193] Injured plaintiffs may bring medical malpractice claims against the individual providers who treated them or against the medical institution where they sought care. The latter may be achieved through vicarious and corporate negligence liability, which will be discussed in more detail in Part II.C. The key consideration in a medical malpractice case is whether the provider violated the standard of care owed to the patient. This issue will be discussed in detail in Part III.B. State court obstacles to medical malpractice claims will be discussed in Part III.A.

While cases involving medical malpractice claims and improperly provided abortion care prior to Dobbs exist, the complaint filed by Rachel Harrison and Marcell Johnson in September 2025 against Dignity Health in California is the first known medical malpractice claim filed by a patient in the context of denied or delayed emergency abortion care after Dobbs.[194]

2. IIED & NIED

Intentional infliction of emotional distress (IIED) and negligent infliction of emotional distress (NIED) are two types of tort claims that seek to hold a defendant liable for conduct causing severe emotional distress to another.[195] IIED requires that a defendant’s conduct be outrageous, that the defendant act recklessly or purposefully, and that the conduct cause severe emotional distress.[196] The standard for NIED differs, requiring only negligent action causing severe emotional distress.[197] Most states allow NIED claims when it was reasonably foreseeable that a defendant’s actions would cause emotional distress.[198]

IIED and NIED claims are an important route to liability for litigators to consider in the context of denied or delayed emergency abortion care. While analysis for these claims is fact-specific, many of the patient stories referenced thus far in this Note already present situations ripe for IIED and NIED claims.[199]

Similarly to medical malpractice claims, IIED and NIED claims can be brought against individual providers who treated a patient or against healthcare institutions where a patient sought treatment.

All three California lawsuits against hospitals feature IIED and NIED claims.[200] The lawsuit filed by Rachel Harrison and Marcell Johnson features an especially horrifying scenario. Rachel Harrison was denied lifesaving abortion care by Dignity Health’s hospital facilities not once but twice, during treatment relating to not one but two different pregnancies.[201] The first time, Harrison waited in the emergency department waiting room for four hours while bleeding, was not given a room or seen by a doctor, and was told by an ultrasound technician that her fetus was the size of an avocado before being discharged to complete a high-risk miscarriage at home without medical supervision.[202]

IIED and NIED claims offer injured patients a viable route towards recourse for the abominable conduct they suffer in hospitals while seeking care. This is particularly important because, while many patients might not suffer long-term physical harms such as loss of fertility, the emotional trauma they endure in these moments cannot be understated and deserves recognition.

3. Wrongful Death

In the most severe cases, the tort of wrongful death could be an option for the families of a pregnant patient who died as a result of denied or delayed care. Just as with malpractice, IIED, and NIED claims, wrongful death claims can be brought against individual providers who treated a patient or against healthcare institutions where a patient sought treatment. Wrongful death requires a finding that the defendant was negligent in causing the individual’s death.[203]

Yennifer Alvarez, an uninsured Texas resident whose family came to the United States from Mexico when she was a child, became pregnant just as Texas’s first abortion ban took effect in 2021.[204] At various points in her pregnancy, she experienced difficulty breathing, unexpected bleeding, and dangerously high blood pressure.[205] She was hospitalized for pulmonary edema but later discharged, and ultimately died after returning to the hospital unable to breathe.[206] The autopsy report read as follows:

Cause of death

Hypertensive cardiovascular disease associated with morbid obesity

Other contributing factors

Pregnancy[207]

Physicians who spoke with The New Yorker stated that, while Alvarez’s high blood pressure and other factors put her at risk of pregnancy-related complications, including the pulmonary edema she developed, an abortion early into her pregnancy would likely have saved her life.[208] The physicians agreed that the outcome Alvarez suffered was “a very preventable maternal death.”[209] And yet pregnancy was noted only as a “contributing factor.”[210]

Alvarez’s story illustrates the need to offer abortions as part of comprehensive maternal-fetal care. While her case differs from emergency abortion situations such as Zurawski’s, which involved PPROM, it underlines just how quickly a pregnancy can become deadly. Moreover, the vast disparity between the levels of news attention Alvarez and Zurawski have received reflects the extent to which many abortion ban-related deaths are likely not making headlines because of how the deaths are classified and the racial and socioeconomic context of a given individual.

Families suing for wrongful death would likely target their claims at the healthcare workers who caused the injury that led to a patient’s death by “wrongful act, neglect, carelessness, unskillfulness, or default.”[211] However, wrongful death statutes vary widely by state, meaning that a plaintiff’s potential recovery would depend on where they file their claims—which itself would depend on where a person died and where they are domiciled. Further research is needed on a state-by-state basis to understand whether the standards state courts apply to the terms “wrongful act, neglect, carelessness, unskillfulness, or default” would support viable negligence claims.[212] Furthermore, damage caps, standing to file a claim and recover damages, statutes of limitations, applicable law, and other questions all depend on each state’s individual laws.[213] In many states, for instance, statutes of limitations hover around two to three years.[214] Finally, depending on the strategies outlined in subsequent sections, the hospital itself may also be named as a defendant under a theory of corporate negligence or respondeat superior.

Wrongful death claims have previously been used in the context of childbirth, so it should be relatively straightforward to extend them to the context of pregnancy and emergency abortions. In 2016, Kira Dixon Johnson, a Black woman in California, died from internal bleeding following a scheduled cesarean section to deliver her son.[215] Johnson languished for hours without appropriate care despite repeated signs of internal bleeding and readmission to the operating room.[216] In 2022, Johnson’s widower filed a wrongful death suit against Cedars Sinai Medical Center in Los Angeles, citing systemic racism as the primary factor in its medical staff’s treatment of patients of color and its negligent treatment of his late wife.[217] While a recent California bill lifted the cap on medical malpractice awards, the timing of Johnson’s claim means he will not benefit from the change, and his noneconomic damage awards remain capped at $250,000.[218] The case also brought national attention to the issue at Cedars Sinai in particular: A year later, the federal government launched a civil rights investigation into the hospital over its treatment of patients of color—particularly Black pregnant patients.[219] Johnson’s claim has since been resolved privately, likely with an undisclosed settlement.[220]

Johnson’s lawsuit exemplifies how to target hospital systems directly by deploying financial pressure, bad publicity, and federal government scrutiny (though such scrutiny is unlikely under the Trump administration). However, because hospitals generally seem to prefer potential malpractice liability to implementing policies that prevent patient harms, it is likely that wrongful death cases would need to be both egregious and numerous to create enough financial pressure to change systemic behavior. Still, cases hold significant weight in the court of public opinion and may offer the bereaved families of pregnant patients some financial compensation through settlements with hospitals. Ultimately, targeting hospitals themselves (rather than nurses and physicians) is likely to yield the greatest impact.

C. Theories of Institutional Tort Liability: Hospitals & Healthcare Systems

Targeting institutions, rather than individual physicians and healthcare workers, offers plaintiffs a chance to sue the corporate actors who are behind the very policies that frightened healthcare workers are required to enact. Targeting hospitals rather than individual physicians would also alleviate issues stemming from the abortion double bind: Instead of imposing liability upon individuals who face the specter of criminal liability, lawsuits would seek to hold responsible the hospital systems who enact—or fail to enact—policy decisions and impose them onto their staff from the top down. Moreover, targeting large corporate actors in the healthcare space with corporate negligence claims could incentivize them to reverse harmful policies, thereby empowering doctors and staff to treat patients according to the standard of care.

1. Managed Care Liability

Managed care liability—a form of liability directed at health insurers—could offer a vehicle for plaintiffs to bring claims against insurance companies in cases where health insurance plans deny coverage for lifesaving abortion care. While this scenario is less common than a denial of care by hospital staff, it is already happening. In 2024, a woman insured under the Federal Health Benefits Program was denied coverage for her D&E when she was diagnosed with PPROM.[221] While an insurance denial may not prevent a patient from receiving care altogether, it could mean a significant delay, which might in turn cause significant injury or death.

In the 1990s, managed care organizations (MCOs) began to spread in the United States, mostly replacing traditional indemnity insurance plans.[222] MCOs are prepaid health plans, such as health maintenance organizations (HMOs) and preferred provider organizations (PPOs), which are essentially networks of doctors and plans that work with certain insurance companies to offer patients care at lower costs.[223] Managed care liability essentially means that an insurance company can be liable for the decisions it makes for patients who are enrolled in its health plans. Managed care liability insurance is a form of liability coverage that is written to cover organizations that deliver medical services on a managed care basis, including the HMOs and PPOs, which are now among some of the most common American health plans.[224]

Managed care liability means that, rather than suing individual physicians, plaintiffs can sue the managed care plan because the plan, not the physician, made a decision regarding a patient’s care.[225] This offers a potentially fruitful route for injured patients, particularly because some data indicates that jurors disfavor corporate defendants compared to individual defendants[226] and because of the enormous wealth that large MCOs possess.[227] Additionally, states with certain tort-related damage caps sometimes do not have limitations for managed care cases, meaning that plaintiffs may be able to obtain higher damage awards.[228] This could all mean increased chances of higher recovery for would-be plaintiffs.

2. Vicarious Liability

Courts can hold hospitals responsible for the negligence of their employees under a theory of respondeat superior, or vicarious liability.[229]

However, many hospitals’ employment schemes may pose obstacles to vicarious liability. Many hospitals employ physicians as independent contractors rather than employees, which reduces institutional liability for a physician’s actions, and some hospitals reinforce this position by including language such as “[p]hysicians are not employees or agents of the hospital” in their conditions of service.[230] However, some courts have still held that hospitals may be liable for the negligent acts of their physicians even when an independent contractor relationship exists.[231]

The issues arising from staff who are independent contractors are particularly relevant in the emergency room context, where many pregnant patients seek emergency care.[232] Three types of physicians work in hospital emergency rooms: staff (employee) physicians, private physicians, and contract physicians.[233] Many hospitals use contracts with third-party vendors who then staff the hospital’s emergency room with physicians.[234] Physicians often choose this employment structure because as contract physicians they have guaranteed incomes, flexible schedules, and can avoid being on call.[235] However, the scope of liability for physicians employed through a third-party contractor is often unclear.[236] In some cases, though, independent contractors may still be held liable under a theory of ostensible agency in situations where a patient could reasonably believe that the physician was an employee of the hospital and the hospital had done nothing to dispel that belief.[237]

Therefore, analysis of a hospital’s liability under a theory of vicarious liability or ostensible agency will depend on the employment details of healthcare workers in a given emergency room and the features of the employment contracts which govern their work.[238] A doctor or nurse’s specific employment scheme may have been disclosed in consent forms that a would-be plaintiff signed upon entering the hospital or may only be revealed through discovery once litigation is already underway.

3. Corporate Negligence Liability

The theory of corporate negligence offers the most promising path to holding hospitals liable for their own negligent institutional conduct, rather than vicariously liable for the actions of their employees.[239]

Under corporate negligence liability, a corporate entity may be directly liable to a plaintiff for violating its own independent duty of care, even when the individual healthcare provider was not negligent.[240] This doctrine emerged as a result of the changing role of physicians and hospitals in American healthcare over the twentieth and twenty-first centuries, originating in a 1965 Illinois case, Darling v. Charleston Community Hospital.[241] In Darling,the Illinois Supreme Court recognized that, because patients now look to hospitals for treatment rather than individual physicians, hospitals themselves owe a duty to their patients.[242] Darling was the first case to hold that a hospital has an independent duty to supervise the doctors who work within it.[243]

In many states, the gradual abolition of the doctrine of charitable immunity—which had previously shielded hospitals from liability—made the holding in Darling possible.[244] Darling’s new theory of liability emerged as a result of a shifting healthcare landscape in which hospitals today “do far more than furnish facilities for treatment,” as they once did.[245] Before the twentieth century, American doctors mostly provided healthcare through house calls.[246] But as the need for hospitals and surgical facilities grew, hospitals began to extend admitting privileges to doctors so that they could use the specialized facilities.[247] While, previously, patients were often treated by a single physician, today that framework is neither possible nor desirable: Modern medicine requires treatment by multiple doctors and healthcare professionals.[248]

As a result, courts across the country have adopted different approaches to corporate negligence, with some choosing more expansive visions of liability and exposure than others. This doctrine could be useful to would-be abortion-denial plaintiffs in many states, but not all: Corporate negligence doctrine varies by state, and not all states have adopted it. As of 1983, courts in fifteen states had already recognized or implemented some form of corporate negligence.[249]Maine, however, is an example of one state that has explicitly declined to adopt corporate negligence doctrine, meaning that the applicability of this doctrine will be limited.[250] A full overview of the states that have and have not adopted corporate negligence is beyond the scope of this Note.

Some states continue to expand corporate negligence doctrine today, thereby increasing its availability. Washington adopted corporate negligence liability in 1984 and has continued to expand the scope of hospital liability for negligence by holding that hospitals have a “duty of care owed . . . directly to patients” regardless of employee negligence.[251] As recently as 2024, the State recognized corporate negligence in the context of emergency medicine: In Estate of Essex v. Grant County Public Hospital District No. 1, the Washington Supreme Court held that hospital licensing statutes and regulations create a nondelegable duty to emergency department patients.[252] The Court declined to “cabin” corporate negligence to the limited set of historically recognized circumstances.[253]

How might corporate negligence doctrine apply in cases of denied or delayed emergency abortion care? For a patient to hold a hospital itself liable for violating its duty of care, the plaintiff would need to argue that the harms they suffered constituted a violation of one of four central duties that the hospital owes to patients: (1) the duty to use reasonable care in the maintenance of safe and adequate facilities and equipment, (2) the duty to select and retain only competent physicians, (3) the duty to oversee all those practicing medicine within the hospital, and (4) the duty to “formulate, adopt, and enforce adequate rules and policies to ensure quality care for the patients.”[254]

The fourth area—failure to formulate, adopt, or enforce adequate rules and policies—is likely to be the most promising in the context of emergency abortion access because it goes directly to a central issue: hospitals’ failures to enact any guidance whatsoever for their physicians.[255] In 2024, Senator Ron Wyden requested that eight American hospitals that had been the subjects of reports of delayed or denied emergency pregnancy care provide documentation regarding their “policies, processes, and procedures related to state abortion laws and emergency reproductive health care.”[256] The hospitals’ responses revealed that many of them continued to rely on guidance created before the very existence of post-Dobbs abortion bans.[257] In other words, these institutions did nothing to create policies or protocols that would be appropriate in light of a changed legal landscape. As a result, a plaintiff could argue that a hospital breached its duty to ensure the safety and care of its patients by failing to formulate policies and processes governing how staff should handle cases that require emergency abortion in a banned state.

The duty to formulate and follow policies was used successfully in Barkes v. River Park Hospital, where the wife of a deceased patient brought a medical malpractice action against a hospital for the death of her husband.[258] Barkes died from a cardiac infection following a visit to the emergency room for arm pain.[259] After initial intake and triage by a paramedic, a nurse practitioner diagnosed him with an overuse arm sprain but did not observe any symptoms suggestive of cardiac issues, so Barkes was discharged without cardiac testing.[260] Two hours later, he was dead.[261] Barkes’ wife sued the hospital, alleging that her husband would still be alive if he had been triaged by a registered nurse rather than a paramedic and then examined by a doctor rather than a nurse practitioner.[262] The jury found the hospital entirely at fault for failing to enforce its written policies and procedures in the course of Barkes’ treatment.[263] The court held that the hospital had breached its duty of care to Barkes by failing to enforce its policy that all patients be seen in the emergency room by an appropriate physician.[264] The court noted that hospital administration is responsible for implementing a system of oversight to enforce its policies and has a duty to “create and maintain effective lines of communication to health care providers practicing within its walls” to ensure their awareness of hospital policies and procedures.[265] The court also observed that Tennessee courts “have long recognized that hospitals owe certain duties of care to patients that come to them for treatment and healing.”[266]

Barkes offers one example of a court recognizing a hospital’s direct duty not only to adopt policies that ensure quality care exists, but also to ensure their implementation. Litigators can use this example to argue that hospitals have an affirmative duty to formulate and enforce policies that offer staff the tools to keep pregnant patients as safe and healthy as possible—especially when a statewide atmosphere of fear and confusion endangers patients. Because the case law around a hospital’s duty to formulate and enforce policies—and within corporate negligence as a whole—is not well developed, litigators have an opportunity to make creative, forward-thinking arguments about hospitals’ duty to create policies that provide patients the standard of care, to communicate those policies with their staff, and to appropriately supervise them.

A key issue in this area of litigation will be whether the hospital knew that its lack of policy or procedures was causing harm to patients. In order for a hospital to be found negligent, the plaintiff must show that the hospital had “actual or constructive knowledge of the defect or procedures which created the harm.”[267] Lawyers for a plaintiff could argue that a hospital had constructive knowledge of harm based on its awareness of even a single instance of delayed or denied care in that hospital. Alternatively, lawyers could argue the hospital was aware simply because of the prevalence of reports and anecdotal evidence about pregnant patients going septic as a result of delayed abortion care.

However, one obstacle for plaintiffs filing corporate negligence claims is gaining access to the documents that would reveal hospital administrators’ actual or constructive knowledge or clarify who makes decisions about enforcing—or declining to enforce—hospital policies. The discussions that those at the helm of healthcare decision-making (most often hospital CEOs and boards of directors) have with their general counsels are likely protected by the cover of attorney-client privilege.[268] Access to the documents that could prove knowledge may therefore be limited. But the constructive aspect of the knowledge requirement offers room for litigators to argue that hospital administrators and directors indeed have knowledge of the defects in their policies (or the absence of policies altogether) that satisfies the knowledge requirement in less restrictive ways.

4. Leveraging Business Incentives

The absence of streamlined procedures for emergency reproductive care suggests that hospitals may be making a deliberate choice to risk the cost of potential malpractice liability over the risk of criminal charges, bad press, pressure from anti-abortion elected officials, and potential loss of employees’ medical licenses. The latter factors all spell bad business, a key consideration in today’s American healthcare landscape.[269] Because of the structuring of American healthcare as a medical-industrial complex, economic motives underlie nearly all decisions in American healthcare.[270] This is especially true as more private equity firms purchase healthcare organizations and turning a profit becomes intertwined with providing medical care.[271] Expanded corporate negligence liability would make prioritizing business over patients more difficult and costly.

Targeting hospitals through a theory of corporate negligence would in many ways be the ideal vehicle for malpractice claims in cases where hospital administrators implement a policy instructing their employees to put patients in harm’s way by explicitly not offering emergency abortions. For instance, a hospital’s blanket policy of turning away abortion patients may constitute corporate negligence and a breach of a duty that the hospital owes its patients. The challenges and downsides inherent in suing individual practitioners, as well as the fraught position of physicians in banned states who are merely trying to do their jobs without going to prison, also make pursuing institutional liability preferable.[272] System-wide change is more likely if the entities feeling the logistical and financial pressure of litigation are the hierarchically highest decision-making bodies themselves—particularly when the hospitals they manage are an important feature of the American economic landscape and are often valued in billions of dollars.[273]

However, a route to system-wide change through individual malpractice claims might also exist if injured plaintiffs bring a high enough volume of cases. If enough of a hospital’s physicians and staff became embroiled in medical malpractice claims that they, in turn, put pressure on a hospital system to change and threatened to find work elsewhere, that hospital’s governing body and leaders might be forced to enunciate clear policies regarding the treatment that staff are permitted to offer in common cases of PPROM. Hospital systems might not be able to change their states’ abortion bans (although the power that hospital systems wield in state politics should not be understated).[274] But hospitals could develop policies that inform employees that the hospital will protect them and fund resources like legal counsel in the event of a prosecution for performing an emergency abortion that is deemed unlawful. While this knowledge might not entirely remedy staff fears when it comes to providing emergency abortion care in banned states, any kind of guarantee of institutional support would be a starting point that presents far more support than currently exists.

Lastly, it is critical to assess whether courts in banned states have expansive or narrow views on corporate negligence liability, which would impact the viability of claims against hospitals with policies that impact would-be plaintiffs. However, because an expansive, state-by-state survey of this case law is beyond the scope of this Note, further research is needed in this area as litigators consider how to bring much-needed claims on behalf of injured plaintiffs.

III. State Law Challenges to Corporate Negligence Liability Claims

While corporate negligence liability, a theory of liability existing within tort law, remains the most promising route forward, several challenges exist on a state-by-state basis.

A. State Procedural Restrictions

Individual state restrictions on medical malpractice claims can significantly impede a would-be plaintiff’s case. These restrictions take two primary forms: screening processes and damage caps.

Several states have initial screening processes that plaintiffs must complete before they can file their claims, creating an additional hoop that injured plaintiffs must jump through. For example, Idaho requires every medical malpractice claim to go through a prescreening process it refers to as “prelitigation screening panels.”[275] Prior to filing a claim, plaintiffs must complete a prelitigation form, which allows them “to request a hearing for prelitigation consideration of a personal injury claim for money damages.”[276] The form requires plaintiff information, the names of all involved physicians and hospitals, and facts to support the claim of medical malpractice.[277] Hearings take place before a panel composed of one licensed Idaho physician, one layman panelist, one resident lawyer, and one person serving as an administrator of a licensed acute care general hospital (in cases of claims against hospitals).[278] A panel chairman and attorneys appointed by the Idaho State Bar oversee prelitigation hearings, which are described by local practitioners as “fast moving and informal.”[279] The hearing acts like a mini-trial, during which the plaintiff must prove that “it was more likely than not that the named defendants failed to meet the applicable standard of health care practice of the community in which the care was or should have been provided.”[280]

After hearing arguments from both sides and deliberating, the panel issues an advisory opinion on the claim’s merit.[281] Though the screening is required, the panel’s opinion is not binding, and plaintiffs are free to file their claims after the hearing.[282] This panel creates an additional barrier to plaintiffs pursuing their claims, which adds delays and requires additional resources.[283] It is also unclear whether the panel’s opinion influences the decision on any later claim it finds to be without merit. Utah has a nearly identical process to Idaho, which the state argues offers assistance to parties “in evaluating the strengths and weaknesses of their cases.”[284]

Medical malpractice damage caps vary widely by state, limiting the potential total sum a plaintiff can hope to collect in the aftermath of negligent care. While Connecticut, New York, Vermont, and Delaware—states where abortion is legal—have no caps, Idaho, Montana, and Texas—states that all restrict abortion—each have a cap of $250,000 for noneconomic damages.[285] The degree to which these caps vary makes an enormous difference in a plaintiff’s potential recovery, which is highly significant in matters concerning reproductive health and future fertility: Two hundred and fifty thousand dollars is little disincentive for a large corporate hospital system whose annual revenue totals in the billions of dollars. Take Dignity Health, the defendant in one of the California lawsuits: Its 2022–2023 revenue was $18.3 billion.[286] Its CEO earned nearly $30 million that year.[287]

For comparison, damage awards for future pain and suffering totaling in the millions of dollars are common for medical injuries which do not involve death in uncapped states. In 2014, a New York state court jury granted a man who eventually required the amputation of his leg after twelve surgeries a $9.1 million medical malpractice award, with noneconomic damages totaling $2 million for past pain and suffering and $4 million for future pain and suffering.[288] These are sums that, while still small compared to the overall revenue of America’s largest hospital systems, are more likely to draw the attention of the corporate leadership at a hospital system like Dignity Health. Damage caps, however, preclude them. Therefore, where a claim is filed may play a pivotal role in how much a plaintiff can hope to recover in damages—which also opens the door to questions about an Erie choice-of-law analysis.[289]

A potential route to dealing with limitations such as screening processes and damage caps could be to try and bring state medical malpractice claims in federal court through diversity claims. Further research is needed in this area, but a plaintiff may be able to bring a diversity claim against a hospital or physician in a case where the plaintiff resides in a different state and was, say, traveling through a banned state when they needed emergency abortion care. Alternatively, a plaintiff could target a healthcare system whose headquarters are in a different state than the one where a plaintiff or physician resides. Questions of whether state limitations such as damage caps and screening processes are procedural law would determine the viability of this route.

B. The Standard of Care in Banned States

Another key challenge to negligence claims for denied or delayed abortion involves the relevant standard of care post-Dobbs, which remains untested. Bringing a successful medical malpractice claim requires proving that a hospital or medical professional breached their duty of care to a patient, which centers around the question of what the applicable standard of care owed to that patient was.[290] In other words, what norm of care did the physician diverge from? Was the course of conduct the same as “what reasonably prudent similar healthcare providers are doing under similar circumstances”?[291] Whether the standard of care in banned states is different from the standard of care in states where abortion remains legal is a central question affecting the viability of tort claims for emergency abortion.

As medical malpractice law evolved in the late nineteenth and early twentieth centuries, courts began comparing physicians’ practices to those of similarly situated professionals in their geographical area, developing the “locality rule.”[292] The locality rule compares physicians to other medical professionals in a specific geographical area, in part to offer a fairer level of comparison to rural doctors, who operated with far fewer resources than doctors in bigger cities.[293]

However, the locality rule was established before the standardization of medical training and certification, which many argue “obviated the need” for such a rule.[294] Many states turned away from the rule after the establishment of national accreditation bodies and other organizations.[295] In 2015, only five states still followed a version of the locality rule: Arizona, Arkansas, Idaho, New York, and Pennsylvania.[296] Elsewhere, a national standard of care presupposes that physicians have the same training and are able to exercise the same “level of judgment and diligence” as peer practitioners in other states and parts of the country.[297] Therefore, it is likely that practitioners in states with abortion bans could not argue for a standard of care lower than the national standard, other than in the remaining states that still use the locality rule.

Under a locality rule analysis, physicians facing lawsuits could argue that doctors in banned states operate under a different set of circumstances than physicians in states where abortion is legal—that courts cannot apply the same abortion standard of care that they do elsewhere because of the prohibited status of the procedure. Furthermore, even physicians in states that do not recognize the locality rule could advocate for its revival, arguing that diverging criminal frameworks result in differences more relevant than geography.

Ultimately, without clarification of which standard of care takes precedence in situations with conflicting laws, determining the standard of care in banned states remains an open question. Therefore, the viability of medical malpractice litigation in state courts may hinge on resolving whether a state can have a “local” standard of care that is predicated upon criminal prohibitions—rather than a standard of care analogous to the one in a state where abortion remains legal.

Truly determining which standard of care should apply in states with abortion bans will require research into any potentially analogous situations of criminalization and/or prohibition of medical tools or procedures (if any exist). Further research must determine how courts have handled competing standards of care, and whether standards of care change in light of such external circumstances. Moreover, unearthing any potential correlations between banned states and states which have enacted limitations on medical malpractice claims, whether procedural or substantive, would help to further understand the state-by-state landscape for tort litigation that this Note discusses.

IV. The Viability of Torts Within the Larger Reproductive Justice Movement

This Part will explore several issues inherent to using torts as a litigation tool for plaintiffs in the post-Dobbs landscape, particularly the possibility that such claims would inflame tensions more than they help pregnant patients. This Part will argue that, though litigation tools may disrupt the current alignment between physicians and the reproductive rights[298] movement, using torts centers patients’ stories, harms, and choices in a way that more aligns with the principles of reproductive justice.[299]

A. Following the Anti-Abortion Playbook?

An initial and major concern with pursuing medical malpractice claims is that such action might merely create distractions for the reproductive healthcare access movement and follows the agenda of the anti-abortion movement. Legislators and other conservative politicians in banned states are transparent about their strategy of shifting blame for the health crisis facing pregnant people from themselves to doctors. In fact, they directly advocate medical malpractice claims as the legitimate forum for abortion litigation.[300]

In Zurawski, the State claimed during oral argument that medical malpractice litigation would define the scope of lifesaving exceptions in Texas’s abortion ban.[301] This was a clear example of conservatives attempting to shift blame away from abortion bans and towards doctors. Moreover, two prominent anti-abortion advocacy groups, the Charlotte Lozier Institute[302] and the Alliance for Hippocratic Medicine,[303] filed an amicus brief in Zurawski in which they argued the same point: The legal standard included in Texas’s abortion ban has long existed in Texas and has never been found vague.[304] The amicus brief was filed in an effort to support Texas’s ban and its emergency exemption, claiming that Texas physicians have long exercised “reasonable medical judgment” in treating pregnancy complications.[305] It further claims that Texas laws do not require doctors to abandon pregnant patients’ treatment in cases of emergencies, and that “refusing to treat such a patient is likely malpractice, given the established standard of care.”[306] In other words: Texas’s statute is not at fault, doctors are, and plaintiffs should take up any issues they encounter with their doctors rather than the state.

This strategy has been on full display in Zurawski and other cases regarding emergency abortion access. At a July 2023 hearing for Zurawski, the State repeatedly asked plaintiffs: “Did Attorney General Ken Paxton tell you you couldn’t get an abortion? Did anyone, working in any capacity for the state, tell you that you couldn’t get an abortion?”[307] The subtext of this argument is clear: Plaintiffs, blame your doctor and their poor judgment; they denied you the abortion you needed, but our hands as legislators and members of government are clean. This strategy is also deployed in other states. Tennessee Attorney General Jonathan Skrmetti stated that parties other than the state (including physicians) were behind pregnant persons’ harms, citing “doctors’ independent choices not to provide permissible abortions.”[308] Moreover, Republican politicians are trying to cast doubt on physicians more generally: Idaho Attorney General Raúl Labrador has publicly questioned the veracity of doctors’ accounts of abortion emergencies, including the need to airlift pregnant people out of state in Idaho to receive care.[309]

This Note’s goal is not to propose ways to make the lives of beleaguered physicians any more difficult than they already are, nor to exacerbate the already tense and fragile relationship between plaintiffs and their doctors in situations of life and death. Rather, this Note seeks to analyze the benefits and drawbacks of alternate litigation tools available to the increasing number of injured plaintiffs with very little recourse currently available to them. The fact that conservative, anti-abortion states and organizations are proposing medical malpractice claims should certainly serve as a caution and warning sign to anyone who seeks to advance reproductive health and access.[310] However, this Note posits that tort claims, which target large and powerful healthcare systems—and not individual doctors—are the most promising path forward. Corporate negligence liability would allow plaintiffs to apply pressure on organizations and corporations with actual political lobbying power,[311] which could, in turn, lead to legislative change—and, ultimately, access to care for pregnant patients.

B. Disrupting the Physician-Reproductive Rights Coalition

Historically, the modern American reproductive rights movement has involved a close partnership with doctors and the overall healthcare system in the United States. The case law contemporaneous with Roe v. Wade centered the right to privacy and the right of an individual and their healthcare provider to make decisions around abortion: The modern conception of abortion access has long framed the procedure as a private choice between a pregnant person and their physician, classically framing the procedure as a choice between a woman and her doctor.[312] Indeed, many of the plaintiffs in the foundational Supreme Court jurisprudence on abortion have been doctors.[313] They remain central to reproductive rights litigation today: The Zurawski complaint featured a claim on behalf of physicians and argued that under the Texas Constitution’s Section 19, “Rights of Physicians,” the state’s ban was depriving doctors of their right to practice their profession freely.[314] Physicians have therefore long been essential actors in the struggle for access to reproductive care. Doctors and other healthcare professionals who provide abortion care have often borne the brunt of the anti-abortion movement’s crusade to put an end to the procedure—sometimes paying with their lives.[315] As scholars C.E. Joffe, T.A. Weitz, and C.L. Stacey point out, the very text of Roe noted the centrality of physicians to the decision:

The decision vindicates the right of the physician to administer medical treatment according to his professional judgment up to the points where important state interests provide compelling justifications for intervention. Up to those points, the abortion decision in all its aspects is inherently, and primarily, a medical decision, and basic responsibility for it must rest with the physician.[316]

Without a doubt, plaintiffs suing their doctors for negligence in cases involving denied or delayed emergency abortion care would make the doctor-patient relationship more adversarial. Merely discussing the prospect of medical malpractice claims threatens one of the core elements of the modern reproductive rights movement: the alliance between physicians and legal abortion advocacy groups and the doctor-patient relationship itself.[317] And yet, who is to say that this relationship is not already becoming adversarial and hostile, given the harrowing stories coming out of American emergency rooms? While this might not be the fault of physicians, the introduction of abortion bans has already created the double bind which produces a conflict of interest for physicians.[318]

Historically, physicians have also not always been allies to the movement for reproductive autonomy. Physician mobilization around abortion first began in the mid-19th century, when physicians were a leading group campaigning for the criminalization of abortion in an effort to differentiate licensed doctors from other healthcare and abortion providers such as nurses and midwives.[319] The American Medical Association (AMA), formed in 1847, made criminalization of abortion “one of its highest priorities, a move based not only on moral objections to abortion, but rather because the issue served so well as the centre of the new organisation’s professionalising project.”[320] By 1880, the AMA campaign was successful in its goal of criminalizing abortion to the extent that physicians could control the terms under which the only “authorized” abortions took place; in many states, that meant “when there was a threat to the life of the mother, or a serious threat to her health as determined by a physician.”[321] But of course there was no “uniform agreement on which conditions posted a true threat to the woman’s life or what degree of threat to her health merited an authorised abortion.”[322] This phrase describes the circumstances on the ground today in banned states.

The modern-day partnership between healthcare professionals and abortion advocacy groups developed in the twentieth century, after abortion was criminalized. But, even now, not all individual physicians or healthcare institutions support abortion access. The American Association of Pro-Life OBGYNs (AAPLOG)—a “special interest group” within the American College of Obstetricians and Gynecologists from 1973 until 2013 that now operates as an independent organization—offers a “Pro-Life Directory” of its members online.[323] AAPLOG encourages its members to work with what it refers to as Pregnancy Resource Centers (also known as “crisis pregnancy centers”) and other programs “helping women with ‘crisis pregnancies,’” which are primarily intended to discourage pregnant individuals from getting abortions.[324] Not all physicians are allies to the movement to widen and increase access to abortion.[325]

As a result, while medical malpractice claims do threaten the carefully maintained alliance between physicians and reproductive rights organizations, and there certainly may be pitfalls to this shakeup, there are also opportunities to consider. Tort claims that target healthcare providers and institutions could create a more patient-centered approach, rather than one that safeguards the interest of larger organizations, who may prioritize institutional litigation goals and movement partnerships. Individual attorneys who only represent one plaintiff who was denied care are likely to have different agendas and obligations when compared to a nonprofit that specializes in impact litigation and has many different clients, some of whom may be physicians. Interestingly, attorneys not traditionally deemed as central to the reproductive rights movement—personal injury and tort lawyers—may be better situated to litigate from a patient-centered mindset. As a result, these lawyers could develop litigation strategies that are more anchored in the approach of the reproductive justice movement.

While this Note advocates the pursuit of institutional liability as opposed to individual claims against physicians, tort claims can also be used against physicians and healthcare institutions whose unequivocal policy is to never provide lifesaving abortion care. A 2020 study showed that “35.3% of U.S. counties, where 38.7% of U.S. women of reproductive age live, have a high Catholic hospital market share.”[326] Many doctors at these hospitals, as well as physicians working at non-Catholic hospital systems, may fall on the side of the “strict constructionists,” individuals who are morally opposed to abortion and who wish for their medical colleagues to follow the most rigid interpretations of abortion ban statutes.[327] Medical malpractice claims against such doctors could instill an understanding that consequences may follow a refusal to treat patients or a harmful delay in treatment. This approach would offer a departure from the current status quo, in which healthcare institutions and their employees can delay or refuse treatment because the only consequences of such actions fall on patients. The California lawsuits against Dignity Health and St. Joseph Health seek to do just that: Both are Catholic-affiliated hospital systems that follow the Ethical and Religious Directives for Catholic Health Services developed by the U.S. Conference of Catholic Bishops—and whose policy is to never provide abortions.[328]

Ultimately, medical malpractice claims and tort litigation against doctors threaten a core concept behind modern reproductive rights jurisprudence: that the doctor is central and indispensable to the procedure.[329] In a post-Dobbs world, advocates have an opportunity to move beyond this paradigm and towards one which is pregnant-person centered. This opportunity is particularly pressing in an era of increased medication abortion thanks to the decentralization of abortions that is taking place through the critical operations of abortion funds, pill providers, and other grassroots organizing.[330] Medication abortion offers patients more power to choose, echoing the premonition of Dr. Robert Hall, an American obstetrician and early advocate for the liberalization of U.S. abortion laws: “When it comes to the doctor, I think he is eventually going to be no more than a technician. This may be humiliating to him. But it is his unavoidable plight if we are to grant women their inherent right to abortion.”[331]

In a world based on the principles of reproductive justice, patients and their needs must be at the center of decision-making. Tort law is a tool in the arsenal of patients, which allows them to seek compensation for harm through monetary damages. Tort law also ideally protects future patients from suffering the same violence by allowing injunctive relief or prompting policy change through the collective influence of patient lawsuits. In this sense, tort law is a mechanism that offers plaintiffs the promise of compensation for past harms and the hope to correct future policy.

C. The Need to Bring Together Diverse Stakeholders

The current healthcare crisis around emergency abortion access makes clear the pressing need to bring together various stakeholders from a range of areas. Though patients and their family members, insurance companies and underwriters, hospital administrators and general counsel, and plaintiff-side medical malpractice and personal injury lawyers may not have the same set of interests or priorities, the shared goal of reducing preventable injuries and even deaths of pregnant people—and perhaps litigious nightmares which might ensue—is one with benefits for all.

Offering patients forums in which to tell their stories and discuss treatment, informed consent options, and more with all stakeholders in the medical system, while also empowering them to come to the table with the ability to sue the medical systems they rely on, could generate real movement within healthcare spaces. It could also force hospitals to develop real procedures that protect (or at least strive to protect) pregnant patients. Such forums might be imperfect, but they would still offer an improvement from the current state of affairs and a step forward.

Meanwhile, hospitals have untapped power to influence abortion policy nationwide. Hospital systems regularly use their lobbying resources in state legislatures when policies threaten their bottom line, often spending millions of dollars.[332] Additionally, while four federal agencies account for the 629 regulatory requirements that health systems, hospitals, and other care providers must comply with,[333] the American Hospital Association is a professional organization that includes over 5,000 hospitals in its membership.[334] The organization conducts legislative and regulatory advocacy efforts, and while it has supported challenges to abortion bans such as Idaho’s through amicus briefs, it could do more to protect patients.[335]

Bringing diverse stakeholders together to at least begin discussing how to create systems that truly provide basic healthcare to all, including pregnant people, is a critical step to take as soon as possible. And we should put patients at the center of the conversation. As long as abortion bans remain, more and more pregnant people will find themselves in harm’s way when they experience emergencies and end up in hospital emergency rooms—or in the parking lots outside.

Conclusion

While the post-Dobbs landscape is messy and complex, pregnant patients facing medical emergencies desperately and urgently need change in the policies on which their lives depend. Tort claims could spark important shifts yet remain largely unexplored in the context of emergency abortion care. Despite the real challenges and obstacles discussed in this Note, opportunities remain for litigation that could both compensate injured plaintiffs and put pressure on healthcare systems to increase immediate access to care for pregnant patients facing medical emergencies. Tort claims that target medical institutions through a theory of corporate liability offer a litigation route with room for creative lawyering in which individual patients sit at the center of advocacy efforts.

Now, more than ever before, pregnant individuals nationwide are at risk for serious bodily injury and death because of inaction by the very people tasked with their lifesaving care. They deserve justice for the physical and emotional harms they needlessly suffer. Litigators should consider using new tools to advocate on behalf of pregnant patients who were denied the lifesaving care they so desperately needed. Action is critical, particularly under a second Trump administration and its continued onslaught on access to reproductive care. While more research at the state level is needed to fully understand the different obstacles that could limit the success of these cases, the tort claims and theories of liability discussed in this Note offer unique mechanisms to put pressure on healthcare organizations and hospital systems. Because those organizations have far more influence and power than individual doctors and nurses, lawsuits targeting hospitals have the potential to create forward-looking change, in addition to providing compensation for individual harms suffered in the past. While nothing can erase the violence inflicted on pregnant people who have suffered needlessly thus far, such claims could help protect future patients from preventable death and harm. Pregnant people deserve far better, and litigators should use every tool available to advocate on their behalf.

Copyright © 2026 Ortal Isaac, J.D. 2025, University of California, Berkeley, School of Law. This Note reflects developments through October 2025. This Note would not exist without the encouragement of Professors Alexa Koenig and Eric Stover, under whose tutelage I began my research in Spring 2024. I am immensely grateful to Professor Koenig for continuing to share her wisdom with me ever since, as well as to Professors Khiara Bridges, Mridula Raman, and Ty Alper for their mentorship, guidance, and care. Special thanks to the participants of the Spring 2024 Human Rights and Social Justice Writing Workshop, my friends in the April 2025 Death Penalty Clinic (Backyard) Writing Workshop, and Jordan Hefcart for their sharp insights on drafts of this piece. Thank you to Jacob Bendicksen for endlessly thinking about this Note with me and supporting me throughout in more ways than I can count. Finally, my sincere thanks to the California Law Review editors who truly went above and beyond in their thoughtful edits.

[1]. “Sepsis is a life-threatening condition that happens when the body’s immune system has an extreme response to an infection . . . . The body’s reaction causes damage to its own tissues and organs . . . . Common signs of sepsis include fever, fast heart rate, rapid breathing, confusion and body pain. It can lead to septic shock, multiple organ failure and death.” Sepsis, World Health Org. (May 3, 2024), https://www.who.int/news-room/fact-sheets/detail/sepsis [https://perma.cc/X9RX-3N6Q].

[2]. See Lizzie Presser, Andrea Suozzo, Sophie Chou & Kavitha Surana, Texas Banned Abortion. Then Sepsis Rates Soared., ProPublica (Feb. 20, 2025), https://www.propublica.org/article/texas-abortion-ban-sepsis-maternal-mortality-analysis?utm_source=substack&utm_medium=email [https://perma.cc/LPD2-42L4]; Kavitha Surana, Abortion Bans Have Delayed Emergency Medical Care. In Georgia, Experts Say This Mother’s Death Was Preventable., ProPublica (Sep. 16, 2024), https://www.propublica.org/article/georgia-abortion-ban-amber-thurman-death [https://perma.cc/SS37-9A9E]; Molly Castle Work, Catholic Hospital Offered a Bucket and Towels to Woman It Denied Abortion, California AG Says, 19th (Oct. 4, 2024), https://19thnews.org/2024/10/catholic-hospital-offered-bucket-towels-woman-abortion/ [https://perma.cc/A3YA-W2S6]; Catherine Lucey, The Tragic Pregnancy Stories Filling the Abortion Campaign Airwaves, Wall St. J. (Apr. 27, 2024), https://www.wsj.com/politics/policy/the-tragic-pregnancy-stories-filling-the-abortion-campaign-airwaves-74c4fa41 [https://perma.cc/M6KY-RBCY]; Selena Simmons-Duffin, In Oklahoma, a Woman Was Told to Wait Until She’s ‘Crashing’ for Abortion Care,NPR (Apr. 25, 2023), https://www.npr.org/sections/health-shots/2023/04/25/1171851775/oklahoma-woman-abortion-ban-study-shows-confusion-at-hospitals [https://perma.cc/ANZ2-UDFH]; Nadine El-Bawab, Tennessee Woman Gets Emergency Hysterectomy After Doctors Deny Early Abortion Care, ABC News (May 31, 2023), https://abcnews.go.com/US/tennessee-woman-gets-emergency-hysterectomy-after-doctors-deny/story?id=99457461 [https://perma.cc/BR34-VEHR]; Selena Simmons-Duffin, Her Miscarriage Left Her Bleeding Profusely. An Ohio ER Sent Her Home to Wait, NPR (Nov. 15, 2022) [hereinafter Simmons-Duffin, Her Miscarriage Left Her Bleeding Profusely], https://www.npr.org/sections/health-shots/2022/11/15/1135882310/miscarriage-hemorrhage-abortion-law-ohio [https://perma.cc/AVU6-ZVBF]; Rosemary Westwood, Bleeding and in Pain, a Pregnant Woman in Louisiana Couldn’t Get Answers, KFF Health News (Jan. 12, 2023), https://kffhealthnews.org/news/article/bleeding-and-in-pain-a-pregnant-woman-in-louisiana-couldnt-get-answers/ [https://perma.cc/8ZRD-4XME]; Petula Dvorak, She Never Wanted an Abortion, Then She Needed One to Save Her Life, Wash. Post (Apr. 25, 2024), https://www.washingtonpost.com/dc-md-va/2024/04/25/abortion-ban-emergency-medical-care-supreme-court/ [https://perma.cc/652T-8YAQ].

[3]. In 2024, Anna Nusslock sought care at Providence St. Joseph Hospital, a Catholic-affiliated hospital in Eureka, California, after her water broke fifteen weeks into her pregnancy. Because fetal heart tones could still be detected, a doctor at the hospital informed Nusslock that the institution’s policy prohibited providing her with an abortion unless her life was at risk. Her husband drove her to a nearby hospital, where she arrived “hemorrhaging and passing a blood clot the size of an apple” and was rushed into surgery. Pam Belluck, California Sues Hospital for Denying Patient an Emergency Abortion, N.Y. Times (Sep. 30, 2024), https://www.nytimes.com/2024/09/30/health/california-abortion-lawsuit-st-joseph-hospital.html [https://perma.cc/PB39-7XFP]. Nusslock said: “I thought I would be safe here [in California] from things like this.” Id. Rachel Harrison was also denied a lifesaving emergency abortion not once, but twice by Dignity Health—over the course of two different pregnancies. “‘While publicly touting their hospitals’ qualifications as reliable emergency services centers, Dignity Health prioritized its own religious directives over the best interests of Rachel’s health and well-being,’ the lawsuit alleges.” Clara Harter, Emergency Abortion Denials by Catholic Hospitals Put Woman in Danger, Lawsuit Claims,L.A. Times (Sep. 30, 2025), https://www.latimes.com/california/story/2025-09-30/california-emergency-abortion-lawsuit [https://perma.cc/67YG-MYJ7].

[4]. Plaintiffs’ Original Petition for Declaratory Judgment and Application for Permanent Injunction ¶¶ 9, 11, Zurawski v. State, No. D-1-GN-23-000968 (Tex. Dist. Ct. Aug. 4, 2023) [hereinafter Plaintiffs’ Original Petition for Declaratory Judgment and Application for Permanent Injunction].

[5]. The Assault on Reproductive Rights in a Post-Dobbs America: Hearing Before the S. Comm. on the Judiciary, 118th Cong. 6 (2023) (testimony of Amanda Zurawski).

[6]. Id.

[7].Plaintiffs’ Original Petition for Declaratory Judgment and Application for Permanent Injunction, supra note 4, ¶ 11; The Assault on Reproductive Rights in a Post-Dobbs America: Hearing Before the S. Comm. on the Judiciary, supra note 5.

       [8]. Plaintiffs’ Original Petition for Declaratory Judgment and Application for Permanent Injunction, supra note 4, ¶ 9.

[9]. Id. ¶ 11.

[10]. Id. ¶ 12.

[11]. Id.

[12]. See Facts Are Important: Understanding and Navigating Viability, Am. Coll. Obstetricians & Gynecologists, https://www.acog.org/advocacy/facts-are-important/understanding-and-navigating-viability [https://perma.cc/QXB2-98PU] (explaining that deliveries before twenty-three weeks have a 5−6 percent survival rate, and “significant morbidity is universal (98–100%) among the rare survivors”).

[13]. Plaintiffs’ Original Petition for Declaratory Judgment and Application for Permanent Injunction, supra note 4, ¶ 13; see also Ayesha Rascoe, In States with Abortion Bans, Hospital Ethics Boards Have the Power to Make Exceptions, NPR (Mar. 12, 2023), https://www.npr.org/2023/03/12/1162917337/in-states-with-abortion-bans-hospital-ethics-boards-have-the-power-to-make-excep [https://perma.cc/FA5N-H73G] (interviewing Dr. Anne Lyerly, professor and OB-GYN: “For example, unfortunately, there has widely been reported situations in which women have had ruptured membranes prior to viability of their pregnancies, which means sort of ruptured membranes several weeks before the baby would be able to survive outside the womb. The standard of care in these cases is to provide a timely abortion to prevent harm to the pregnant woman. There is nothing to be gained. There’s no chance that the pregnancy would be able to continue or that the baby would be able to survive if the pregnancy is prolonged. And in fact, waiting would be against the standard of care and is ethically problematic”).

[14]. The Assault on Reproductive Rights in a Post-Dobbs America: Hearing Before the S. Comm. on the Judiciary, supra note 5, at 6; Plaintiffs’ Original Petition for Declaratory Judgment and Application for Permanent Injunction, supra note 4, ¶ 13.

[15]. The Assault on Reproductive Rights in a Post-Dobbs America: Hearing Before the S. Comm. on the Judiciary, supra note 5, at 6.

[16]. Plaintiffs’ Original Petition for Declaratory Judgment and Application for Permanent Injunction, supra note 4, ¶¶ 12, 15.

[17]. The Assault on Reproductive Rights in a Post-Dobbs America: Hearing Before the S. Comm. on the Judiciary, supra note 5, at 7.

[18]. Plaintiffs’ Original Petition for Declaratory Judgment and Application for Permanent Injunction, supra note 4, ¶¶ 19–20.

[19]. Id. ¶¶ 21, 23.

[20]. Id. ¶ 22.

[21]. Id. ¶¶ 23–24.

[22]. See id. ¶ 25.

[23]. Id. ¶ 144.

[24]. State v. Zurawski, 690 S.W.3d 644, 653–54 (Tex. 2024).

[25]. See generally Plaintiffs’ Original Petition for Declaratory Judgment and Application for Permanent Injunction, supra note 4.

[26]. See, e.g., Pooja Salhotra & Eleanor Klibanoff, Amid Support from Doctors Group, Bill to Clarify Texas’ Abortion Ban Does Little to Save Lives, Critics Say, Tex. Trib. (Mar. 27, 2025), https://www.texastribune.org/2025/03/27/texas-abortion-bill-senate-31/ [https://perma.cc/F5Z7-JPDC].

[27]. See generally, e.g.,Plaintiffs’ Original Petition for Declaratory Judgment and Application for Permanent Injunction, supra note 4; Isabella Volmert, Lawsuit Challenging Indiana Abortion Ban Survives a State Challenge,Associated Press (Apr. 4, 2024), https://apnews.com/article/indiana-appeals-court-religious-freedom-law-abortion-4da0cd6d585e69ede87bea2ee2da2896 [https://perma.cc/8W6V-VHD6]; Oklahoma Call for Reproductive Justice v. Drummond (Oklahoma),Ctr. for Reprod. Rts., https://reproductiverights.org/case/post-roe-state-abortion-ban-litigation/oklahoma-call-for-reproductive-justice-v-oconnor-2/ [https://perma.cc/LV4U-GQT6]; Sabrina Talukder, In Idaho v. United States, the Supreme Court Must Reckon with the Post-Dobbs Reality It Created, Ctr. Am. Progress (Feb. 26, 2024), https://www.americanprogress.org/article/in-idaho-v-united-states-the-supreme-court-must-reckon-with-the-post-dobbs-reality-it-created/ [https://perma.cc/SY3K-VTTV].

[28]. Kavitha Surana, Maternal Deaths Are Expected to Rise Under Abortion Bans, but the Increase May Be Hard to Measure, ProPublica (July 27, 2023), https://www.propublica.org/article/tracking-maternal-deaths-under-abortion-bans [https://perma.cc/D447-L82L].

[29]. See Katha Pollitt, Maternal Mortality Is Rising, and Pro-Lifers Don’t Care, Nation (July 21, 2023), https://www.thenation.com/article/society/maternal-mortality/ [https://perma.cc/JJ3N-TXMR]; Eugene Declercq, Ruby Barnard-Mayers, Laurie C. Zephyrin & Kay Johnson, The U.S. Maternal Health Divide: The Limited Maternal Health Services and Worse Outcomes of States Proposing New Abortion Restrictions, Commonwealth Fund (Dec. 14, 2022), https://www.commonwealthfund.org/publications/issue-briefs/2022/dec/us-maternal-health-divide-limited-services-worse-outcomes [https://perma.cc/AQQ3-UQV8].

[30]. See Latoya Hill, Alisha Rao, Samantha Artiga & Usha Ranji, Racial Disparities in Maternal and Infant Health: Current Status and Efforts to Address Them, KFF (Oct. 25, 2024), https://www.kff.org/racial-equity-and-health-policy/issue-brief/racial-disparities-in-maternal-and-infant-health-current-status-and-efforts-to-address-them/ [https://perma.cc/F5Q5-AQSW]. See generally Donna L. Hoyert, CDC, Maternal Mortality Rates in the United States, 2021 (2023), https://www.cdc.gov/nchs/data/hestat/maternal-mortality/2021/maternal-mortality-rates-2021.htm [https://perma.cc/Q59L-V2GD].

[31]. See Liza Fuentes, Inequity in US Abortion Rights and Access: The End of Roe Is Deepening Existing Divides,Guttmacher (Jan. 17, 2023), https://www.guttmacher.org/2023/01/inequity-us-abortion-rights-and-access-end-roe-deepening-existing-divides [https://perma.cc/FMS2-V79S]; Noel Lyn Smith & Maddy Keyes, Indigenous Women Navigate Abortion Access Hurdles Post-Roe, Source NM (Aug. 29, 2023), https://sourcenm.com/2023/08/29/indigenous-women-navigate-abortion-access-hurdles-post-roe [https://perma.cc/PA6P-M9DN]. The impact of abortion bans is also deeply intertwined with the risk factors that put pregnant people at higher risk for obstetric complications. See Plaintiffs’ Original Petition for Declaratory Judgment and Application for Permanent Injunction, supra note 4, ¶¶ 159–61 (“Racial and ethnic disparities in pregnancy-related health outcomes are well-documented throughout the medical literature. Research has shown that, as compared to non-Hispanic white women, Black women in the U.S. are considerably more likely to experience obstetric complications like hypertensive disorders and preterm birth and to die from complications like preeclampsia, eclampsia, obstetric embolism, hemorrhage, and postpartum cardiomyopathy. Additionally, Black people in the United States are more likely to have preexisting conditions that may be exacerbated by pregnancy such as high blood-pressure, asthma, diabetes, sickle cell disease, and lupus. The Texas Maternal Mortality Report further notes that ‘delay in referring or access to treatment,’ ‘lack of standardized policies/procedures,’ ‘failure to screen/inadequate assessment of risk,’ ‘lack of continuity of care,’ and ‘lack of access/financial resources’ are all contributing factors in maternal deaths in Texas. Barriers such as these disproportionately impact Black patients. Black patients face significant barriers to quality, equitable health care, including delays in care, systemic discrimination, and implicit biases in their interactions with health care providers. Black women in Texas also face disproportionate poverty: 19.5% of Black Texans lives in poverty compared to 10.5% of white Texans. And 15.5% of Texan women live in poverty compared to 13% of Texan men.” (footnotes omitted)).

[32]. See Cassandra Jaramillo & Kavitha Surana, A Texas Woman Died After the Hospital Said It Would Be a Crime to Intervene in Her Miscarriage, Tex. Trib. (Oct. 30, 2024), https://www.texastribune.org/2024/10/30/texas-abortion-ban-josseli-barnica-death-miscarriage/ [https://perma.cc/5DN7-UW8H].

[33]. Seegenerally Fabricio J. Alarcon, The Migrant Crisis and Access to Health Care,8 Del. J. Pub. Health 20 (2022).

[34]. Seegenerally Kim Korinek & Ken R. Smith, Prenatal Care Among Immigrant and Racial-Ethnic Minority Women in a New Immigrant Destination: Exploring the Impact of Immigrant Legal Status, 72 Soc. Sci. & Med. 1695 (2011).

[35]. Priscilla Alvarez, ICE Has Deported Nearly 200k People Since Trump Returned to Office, on Track for Highest Level in a Decade,CNN (Aug. 28, 2025), https://www.cnn.com/2025/08/28/politics/ice-deportations-immigrants-trump [https://perma.cc/W4YD-VSFK].

[36]. Lauren Kaori Gurley, Federica Cocco & Andrew Ackerman, More Americans Take on a Second Job or Side Hustle. They Come at a Cost., Wash. Post (Nov. 2, 2025), https://www.washingtonpost.com/business/2025/01/28/americans-work-two-jobs-market-side-hustle/ [https://perma.cc/AUP2-C28Q].

[37]. Elise Gould, Over 60% of Low-Wage Workers Still Don’t Have Access to Paid Sick Days on the Job, Econ. Pol’y Inst.: Working Econ. Blog (Sep. 23, 2022, at 14:47 PT), https://www.epi.org/blog/over-60-of-low-wage-workers-still-dont-have-access-to-paid-sick-days-on-the-job/ [https://perma.cc/KK8B-23PV].

[38]. Steven Ross Johnson, Millions of Americans Lack Reliable Transportation. It May Affect Their Health., U.S. News (Jan. 11, 2024), https://www.usnews.com/news/health-news/articles/2024-01-11/millions-of-americans-lack-reliable-transportation [https://perma.cc/B9PZ-92FJ].

[39]. Jonathan Koltai, Jess Carson & Tyrus Parker, Childcare Remains out of Reach for Millions in 2021, Leading to Disproportionate Job Losses for Black, Hispanic, and Low-Income Families, Univ. of N.H. (Dec. 23, 2021), https://carsey.unh.edu/publication/childcare-remains-out-reach-millions-2021-leading-disproportionate-job-losses-black-hispanic-low [https://perma.cc/34C8-NRUA].

[40]. Jaramillo & Surana, supra note 32.

[41]. Id.

[42]. Id.

[43]. Charlotte Alter, Amanda Zurawski Won’t Give up the Fight for Reproductive Rights, Time (Feb. 20, 2025), https://time.com/7216391/amanda-zurawski-texas-reproductive-rights/ [https://perma.cc/AY89-H6HG].

[44]. Team, Zurawski v Texas, https://www.zurawskivtexas.com/team [https://perma.cc/E5YC-24CB].

[45]. Together Films, Zurawski v Texas – Official Trailer (YouTube, Oct. 21, 2024), https://youtu.be/M3QdzIgJis0?si=VAWeJxmwY_EmU8zU (on file with the California Law Review).

[46]. Video posted by TIME (@time), Instagram (Feb. 27, 2025), https://www.instagram.com/time/reel/DGlsv9sM5H6/ (on file with the California Law Review); Jaramillo & Surana, supra note 32.

[47]. See William Melhado, Texas AG Appeals Judge’s Order That Allows Women with Complicated Pregnancies to Get Abortions,Tex. Trib. (Aug. 4, 2023), https://www.texastribune.org/2023/08/04/texas-abortion-ban-lawsuit/ [https://perma.cc/K2VH-9VH2].

[48]. Floriane Borel, Samira Damavandi & Irum Taqi, Six Months in: How the Trump Administration Is Undermining Sexual and Reproductive Health and Rights Globally, Guttmacher (Aug. 1, 2025), https://www.guttmacher.org/2025/08/six-months-how-trump-administration-undermining-sexual-and-reproductive-health-and-rights [https://perma.cc/LZ7H-SRY3].

[49]. See generallyInteractive Map: US Abortion Policies and Access After Roe,Guttmacher, https://states.guttmacher.org/policies/ [https://perma.cc/AM5A-B3J9].

[50]. See Mary Clare Jalonick, Republicans Block Bill to Protect Contraception Access as Democrats Make Election-year Push, Associated Press (June 5, 2024), https://apnews.com/article/contraception-senate-abortion-biden-trump-reproductive-rights-3f9e8546624a3acf8e64d1138fcb84b1 [https://perma.cc/8EZ7-RJSS]; see also Ryan Cooper, The Republican Party Is Coming for Birth Control, Am. Prospect (June 6, 2024), https://prospect.org/health/2024-06-06-republican-party-coming-for-birth-control/ [https://perma.cc/JXP5-QW49] (arguing that, following the Supreme Court’s decision to overturn Roe v. Wade despite the principle of stare decisis, conservatives are mounting support to overrule constitutional precedent protecting the right to contraception).

[51]. See Nick Reynolds, Republicans Declare War on Sex Education, Newsweek (May 10, 2023), https://www.newsweek.com/republicans-declare-war-sex-education-seek-restrictions-public-schools-1777650 [https://perma.cc/79MD-PJTW]; Hannah Fingerhut, Some GOP-led States Are Chipping Away at Sex Education in K-12 Schools, Associated Press (Oct. 6, 2023), https://apnews.com/article/sex-education-us-laws-kentucky-florida-mississippi-bd6fffcb3a9a25babc32a143001f4ae1 [https://perma.cc/V6F7-C67S].

[52]. See Presser et al., supra note 2 (analyzing Texas’s hospital discharge data to calculate the impact of the state’s abortion ban on maternal mortality).  ProPublica chronicled the CDC’s failure to ask states to track deaths linked to abortion bans and to require state maternal mortality review committees to examine the role of abortion bans in statewide maternal mortality, which has led to a dearth of information about post-Dobbs maternal mortality. Some states, including Texas, have forgone federal funding to avoid sharing maternal death data with the CDC. Scientists have likened this to “pushing [the impact of abortion bans on maternal mortality] under the rug”: without data about the issue, it’s far simpler to pretend it does not exist. See Kavitha Surana, Robin Fields & Ziva Branstetter, The CDC Hasn’t Asked States to Track Deaths Linked to Abortion Bans, ProPublica (Dec. 20, 2024), https://www.propublica.org/article/abortion-ban-deaths-cdc-maternal-health-care [https://perma.cc/6T2C-6R4C].

[53]. Presser et al., supra note 2.

[54]. Id.

[55]. See Amy Littlefield, “She Had a Heartbeat Too”: Waiting for One Dead Woman, Nation (Mar. 14, 2023), https://www.thenation.com/article/society/texas-abortion-lawsuit/ [https://perma.cc/G7XA-V9CQ]. (“In Ireland, her name was Savita Halappanavar. She was a dentist. Her water broke at 17 weeks, like Amanda Zurawski’s. Doctors in Ireland told her that they could not end her pregnancy because the fetus was protected under Ireland’s Eighth Amendment as long as it still had a heartbeat. She begged for an abortion. Like Zurawski, she developed sepsis. Then she died. She was 31. Her death ignited a political revolution that liberalized Ireland’s abortion laws. Thousands of people rallied in the streets holding banners with Savita’s portrait that read ‘Never again.’ Six years later, Irish voters repealed the Eighth Amendment in a referendum. Under the right circumstances, one death is enough.”).

[56]. See Plaintiffs’ Original Petition for Declaratory Judgment and Application for Permanent Injunction, supra note 4, ¶¶ 19, 23, 71, 233, 235.

[57]. See El-Bawab, supra note 2.

[58]. See Simmons-Duffin, supra note 2 (discussing how patients who arrive at the emergency room with heavy bleeding and other severe symptoms are not treated by hospital staff “unless [they] are crashing in front of [the doctors] or [their] blood pressure goes so high that [they] are fixing to have a heart attack”); see also Caroline Kitchener & Dan Diamond, She Filed a Complaint After Being Denied an Abortion. The Government Shut Her Down,Wash. Post (Jan. 19, 2024), https://www.washingtonpost.com/politics/2024/01/19/oklahoma-abortion-emtala/ [https://perma.cc/GS2H-ZEFJ] (chronicling how the Biden administration rejected patient Jaci Statton’s complaint that her case was handled in violation of the Emergency Medical Treatment and Labor Act (EMTALA)).

[59]. See generally David S. Cohen, Greer Donley & Rachel Rebouché, Rethinking Strategy After Dobbs,75 Stan. L. Rev. Online 1 (2022).

[60]. Thus far, torts have been utilized far more by anti-abortion activists to make abortion look extremely unsafe—a claim that is contrary to medical evidence. For example, in 2023, a family filed a wrongful death claim against a hospital who treated a woman who died not long after undergoing a legal abortion at a different facility. See David Wilson, Lawsuit: Pahrump Woman Died After Not Being Treated Properly Following Abortion,Pahrump Valley Times (Sep. 27, 2023), https://pvtimes.com/uncategorized/lawsuit-pahrump-woman-died-after-not-being-treated-properly-following-abortion-124871/ [https://perma.cc/VDB9-YMQ2].

[61]. See Jaramillo & Surana, supra note 32.

[62]. Chad Follmer, an insurance broker based in San Francisco, told Business Insurance that most insurers are “remaining silent” on questions regarding emergency abortions and abortion criminal bans, and that insurance companies are “trying to get their arms around it[.]” “Everyone is ‘kind of waiting for someone else to blink.’” Judy Greenwald, Dobbs Decision Casts Shadow on Med Mal, Bus. Ins. (Oct. 5, 2022), https://www.businessinsurance.com/dobbs-decision-casts-shadow-on-med-mal-dobbs-vs-jackson-women%C2%92s-health-organiza [https://perma.cc/A66Q-RCME].

[63]. See Harris Meyer, Malpractice Lawsuits over Denied Abortion Care May Be on the Horizon,KFF Health News (June 23, 2023), https://kffhealthnews.org/news/article/malpractice-lawsuits-denied-abortion-care/ [https://perma.cc/6EJV-UHT9]. See generally Complaint, Harrison v. Dignity Health, No. CGC-25-629618 (Cal. Super. Ct. S.F. Cnty. Sep. 26, 2025) [hereinafter Harrison Complaint].

[64]. The first would-be medical malpractice tort action targeting a hospital was announced in October 2024, after delayed abortion care caused the death of Amber Thurman in Georgia. Months later, however, no complaint has emerged, suggesting that the family may be in the process of settling with the hospital system it announced it planned to sue. See Attorney Ben Crump and Family of Amber Thurman to Demand Justice, Address Maternal Health Crisis in Georgia, BenCrump.com (Oct. 1, 2024), https://bencrump.com/press/attorney-ben-crump-and-family-of-amber-thurman-to-demand-justice-address-maternal-health-crisis-in-georgia/ [https://perma.cc/Y3RV-387C].

[65]. Rachel’s partner, Marcell Johnson, is a coplaintiff in the lawsuit. Harrison Complaint, supra note 63, at 2.

[66]. See generally Harrison Complaint, supra note 63; Complaint, Nusslock v. St. Joseph Health of N. Cal., LLC, No. CV2500674 (Cal. Super. Ct. Humboldt Cnty. Apr. 1, 2025) [hereinafter Nusslock Complaint]; Complaint for Damages and Equitable Relief, Roe v. St. Joseph Health N. Cal., LLC, No. CV2402362 (Cal. Super. Ct. Humboldt Cnty. Dec. 12, 2024) [hereinafter Roe Complaint].

[67].Harrison Complaint, supra note 63, ¶¶ 131–38.

[68]. Harrison Complaint, supra note 63, ¶ 37; Roe Complaint, supra note 66, ¶ 22.

[69]. Roe Complaint, supra note 66, ¶ 22 (quoting the U.S. Conf. of Catholic Bishops, Ethical and Religious Directives for Catholic Health Care Services 18 (6th ed. 2018), https://www.usccb.org/resources/ethical-religious-directives-catholic-health-service-sixth-edition-2016-06_0.pdf [https://perma.cc/UCL6-P5MB]).

[70]. Joseph S. Kass & Rachel V. Rose, Medical Malpractice Reform: Historical Approaches, Alternative Models, and Communication and Resolution Programs,18AMA J. Ethics 299, 299 (2016).

[71]. See David M. Studdert, Michelle M. Mello & Troyen A. Brennan, Medical Malpractice,350 NEJM 283, 283 (2004).

[72]. Id. at 299–300.

[73]. Michael J. Bono, Harrison R. Wermuth & John E. Hipskind, Medical Malpractice, in StatPearls (2022) (ebook), https://www.ncbi.nlm.nih.gov/books/NBK470573/ [https://perma.cc/K2WG-R6MB].

[74]. See, e.g., Carolyn Caccese, The Relationship Between Malpractice Litigation & Patient Safety, Salenger Sack Kimmel & Bavaro LLP: SSKB News (Sep. 16, 2024), https://sskblaw.com/the-relationship-between-malpractice-litigation-patient-safety [https://perma.cc/QGE3-YWZE] (“When healthcare organizations are held accountable through litigation, they are more likely to review their practices and address any deficiencies, ultimately improving patient safety for future patients.”).

[75]. Studdert, Mello & Brennan, supra note 71, at 287.

[76]. Our Movement, Patient Safety Movement, https://psmf.org/psmf-our-movement[https://perma.cc/RYC9-NGLV]. The patient safety movement gained national attention particularly with The Institute of Medicine’s 2000 publication of To Err Is Human: Building a Safer Health System, a report that highlighted the numbers relating to patient harms and deaths stemming from preventable human error. The report estimated that nationally, at least 44,000 and as many as a potential 98,000 Americans died each year because of medical errors. See To Err Is Human: Building a Safer Health System 1(Linda T. Kohn, Janet M. Corrigan & Molla S. Donaldson eds., 2000).

[77]. Our Movement, supra note 76; see alsoActionable Evidence-Based Practices, Patient Safety Movement, https://psmf.org/actionable-evidence-based-practices/page/2/?et_blog [https://perma.cc/5EA9-ZX7S]; Early Detection and Treatment of Sepsis, Patient Safety Movement (Apr. 17, 2024), https://psmf.org/aebp-publications/early-detection-and-treatment-of-sepsis/ [https://perma.cc/2HML-J5W7].

[78]. See Andrew Jay McClurg, Fight Club: Doctors vs. Lawyers–A Peace Plan Grounded in Self-Interest, 83 Temp. L. Rev. 309, 313–22, 329 n.130 (2011).

[79]. Peter P. Budetti, Kaiser Fam. Found., Medical Malpractice Law in the United States 14 (2005), https://www.kff.org/wp-content/uploads/2013/01/medical-malpractice-law-in-the-united-states-report.pdf [https://perma.cc/RSN8-W6UP].

[80]. See, e.g., Medical Liability: New Ideas for Making the System Work Better for Patients:Hearing Before the S. Comm. on Health, Educ., Lab., & Pensions, 109th Cong. 19 (2006) (“Tort law’s punitive, individualistic, adversarial approach is antithetical to the nonpunitive, systems-oriented, cooperative strategies espoused by patient safety leaders. Litigation entails secrecy and blame, whereas modern quality improvement strategies demand transparency and focus on systems of care, not individuals.”).

[81]. Michael D. Frakes, The Surprising Relevance of Medical Malpractice Law, 82 U. Chi. L. Rev. 317, 327 n.19 (2015).

[82]. José R. Guardado,AMA, Policy Research Perspectives: Medical Liability Claim Frequency Among U.S. Physicians 6 (2017), https://www.ama-assn.org/sites/ama-assn.org/files/corp/media-browser/public/government/advocacy/policy-research-perspective-medical-liability-claim-frequency.pdf [https://perma.cc/TGH3-V63M].

[83]. Kelly E. Wong, P. Divya Parikh, Kwon C. Miller & Mark R. Zonfrillo, Emergency Department and Urgent Care Medical Malpractice Claims 2001-15, 22 W.J. Emergency Med. 333, 334 (2021).

[84]. Guardado, supra note 82, at 1.

[85]. Roosa Tikkanen, Munira Z. Gunja, Molly FitzGerald & Laurie C. Zephyrin, Maternal Mortality and Maternity Care in the United States Compared to 10 Other Developed Countries, Commonwealth Fund (Nov. 18, 2020), https://www.commonwealthfund.org/publications/issue-briefs/2020/nov/maternal-mortality-maternity-care-us-compared-10-countries [https://perma.cc/J42V-TXBY] (concluding that the United States had the highest rate of maternal mortality among developed countries even before Dobbs).

[86]. See generally Munira Z. Gunja, Evan D. Gumas, Relebohile Masitha & Laurie C. Zephyrin, Insights into the U.S. Maternal Mortality Crisis: An International Comparison, Commonwealth Fund (June 4, 2024), https://www.commonwealthfund.org/publications/issue-briefs/2024/jun/insights-us-maternal-mortality-crisis-international-comparison [https://perma.cc/DSA7-L5UF].

[87]. Id. at 5.

     [88]. Budetti, supra note 79, at 1, 4.

     [89]. Kass & Rose, supra note 70, at 301.

[90]. Id.

     [91]. Sarah Edwards, States Most Affected by Medical Malpractice, Forbes: Advisor (Dec. 18, 2023), https://www.forbes.com/advisor/legal/medical-malpractice/medical-malpractice-cases-by-state/ [https://perma.cc/EH7J-S5RR].

[92]. See generally Leslie Reagan, Victim or Accomplice: Crime, Medical Malpractice, and the Construction of the Aborting Woman in American Case Law, 1860s-1970, 10 Colum. J. Gender & L. 311 (2001).

[93]. Id. at 317.

[94]. Id. at 317–18.

[95]. Id. at 319.

[96]. Id. at 320.

[97]. Id. at 319.

[98]. See generally Inst. of Med., The Future Of Public Health (1988).

[99]. See generally Michael J. Saks & Stephan Landsman, The Paradoxes of Defensive Medicine, 30 Health Matrix 25 (2020).

[100]. Id.; see also Michael Rothberg, Joshua Class, Tara F. Bishop, Jennifer Friderici, Reva Kleppel & Peter K. Lindenauer, The Cost of Defensive Medicine on Three Hospital Medicine Services, 174 JAMA Internal Med. 1867(2014).

[101]. See generally Daniel P. Kessler & Mark McClellan, Do Doctors Practice Defensive Medicine? (Nat’l Bureau of Econ. Rsch., Working Paper No. 5466, 1996); Mario Plebani, Defensive Medicine and Diagnostic Testing, 1 De Gruyter 151 (2014).

[102]. See Jake Miller, Does Defensive Medicine ‘Work’?, Harv. Med. Sch. (Nov. 4, 2015), https://hms.harvard.edu/news/does-defensive-medicine-work-0 [https://perma.cc/GWG6-XSYV]; see also Anupam B. Jena, Lena Schoemaker, Jay Bhattacharya & Seth A. Seabury, Physician Spending and Subsequent Risk of Malpractice Claims: Observational Study, 2015 BMJ 1, 3 (using rate of caesarean sections as an indicator of physician response to malpractice liability, finding that “[g]eographic areas with greater [malpractice liability exposure] have been associated with higher caesarean delivery rates”).

[103]. See generally Dov Fox, The Abortion Double Bind, 113 Am. J. Pub. Health 1068 (2023); Christian De Vos, Michele Heisler, William Jaffe, Payal Shah, Tamya Cox-Touré, Priya Desai, Nimra J. Chowdhry, Risa Kaufman & Rabia Muqaddam, Physicians for Hum. Rts, Okla. Call for Repro. Just. & Ctr. Repro. Rts., No One Could Say: Accessing Emergency Obstetrics Information as a Prospective Prenatal Patient in Post-Roe Oklahoma 4(2023), https://reproductiverights.org/wp-content/uploads/2023/04/OklahomaAbortionBanReport_Full_SinglePages-NEW-4-27-23.pdf [https://perma.cc/7VYF-BRNL] (“American Medical Association President Jack Resneck, Jr. has decried the ‘chaos’ into which health care has been thrust since the Dobbs decision, describing physicians as ‘caught between good medicine and bad law,’ struggling to ‘meet their ethical duties to patients’ health and well-being, while attempting to comply with reckless government interference in the practice of medicine that is dangerous to the health of . . . patients . . . . Physicians and other health care professionals must attempt to comply with vague, restrictive, complex, and conflicting state laws that interfere in the practice of medicine.’” (alterations in original)).

[104]. Miss. Code Ann. §§ 41-41-45, 97-3-3.

[105]. See Eleanor Klibanoff, Texans Who Perform Abortions Now Face up to Life in Prison, $100,000 Fine, Tex. Trib. (Aug. 25, 2022), https://www.texastribune.org/2022/08/25/texas-trigger-law-abortion/ [https://perma.cc/L5SJ-BPJP]; see alsoInterim Update:Abortion-Related Crimes After Dobbs, Tex. Dist. & Cnty. Att’ys Ass’n (June 24, 2022), https://www.tdcaa.com/legislative/dobbs-abortion-related-crimes/ [https://perma.cc/3L2T-P3HD].

[106]. See generally Erika L. Sabbath, Samantha M. McKetchnie & Kavita S. Arora, US Obstetrician-Gynecologists’ Perceived Impacts of Post-Dobbs v Jackson State Abortion Bans, 7 JAMA Network Open 1 (2024); see also Kavitha Surana, Their States Banned Abortion. Doctors now Say They Can’t Give Women Potentially Lifesaving Care,ProPublica (Feb. 26, 2024), https://www.propublica.org/article/abortion-doctor-decisions-hospital-committee [https://perma.cc/EKN7-K4XA] (discussing an OB-GYN’s experience as a physician in a banned state: “‘We are on the front lines,’ he said. ‘At the end of the day, the patients are staring right in our faces.’ Late last year, he sighed heavily as he counseled the woman whose baby was developing without a skull and gently told her what he tells all of his patients in her position: that he had the training to help her, but because of Tennessee’s laws, he might face prosecution and jail time if he did. He had a baby at home and couldn’t take that risk, he explained. Instead he would refer her to options outside the state”).

[107]. See Surana, supra note 106 (“No doctor wants to be the first to stand trial. ‘I don’t know how you can overinterpret the law when you are looking at jail time,’ said Dr. Dawn Bingham, an OB-GYN in South Carolina. ‘A prudent person would hear that and go, “Well I guess I will interpret that to be as safe as possible.”’”).

[108]. See Sabbath, McKetchnie & Arora, supra note 106 (“One participant recounted the visceral nature of this fear: ‘I’m in the [operating room] dry heaving. I’m not dry heaving because of this surgery. I know how to do this surgery. I trained for this surgery. I trained for the stress of treating an unstable . . . ectopic pregnancy. I did not train for, I am not ready for thinking about, “Is this the case that’s gonna make me a felon?”’”).

[109]. Morning Edition, Arizona Attorney General Says She Won’t Enforce a 164-Year-Old Abortion Law,NPR(Apr. 12, 2024), npr.org/2024/04/12/1244265593/arizona-attorney-general-says-she-wont-enforce-a-164-year-old-abortion-law [https://perma.cc/ZM7M-54VJ].

[110]. See Jessica Boehm, “I have no intention to break an established law”: Arizona Doctors Unlikely to Perform Illegal Abortions, Axios Phx.(Apr. 18, 2024), https://www.axios.com/local/phoenix/2024/04/18/arizona-abortion-access-doctors-illegal-prosecution [https://perma.cc/TS4S-QDBE] (“‘I have no intention to break an established law just because an attorney says they’re not going to prosecute,’ Ronald Yunis, a physician at Acacia Women’s Center, told Axios Phoenix.”).

[111]. Sabbath, McKetchnie & Arora, supra note 106.

[112]. Chelsea Tejada, Texas’ Bounty Hunter Abortion Ban Is a Dire Warning of What Lays Ahead for Our Reproductive Rights, ACLU (Mar. 17, 2022), https://www.aclu.org/news/reproductive-freedom/texas-bounty-hunter-abortion-ban-is-a-dire-warning-of-what-lays-ahead-for-our-reproductive-rights [https://perma.cc/TD8R-3DCH].

[113]. D&E is the standard procedure for abortion after fifteen weeks of gestation. See Dilation and Evacuation (D&E),Kaiser Permanente, https://healthy.kaiserpermanente.org/health-wellness/health-encyclopedia/he.dilation-and-evacuation-d-e.tw2462 [https://perma.cc/LR9T-QPWU]; Whitney Arey, Klaira Lerma, Anitra Beasley, Lorie Harper, Ghazaleh Moayedi & Kari White, A Preview of the Dangerous Future of Abortion Bans — Texas Senate Bill 8, 387 NEJM 387, 388–90 (2022).

[114]. Arey et al., supra note 113, at 389.

[115]. D&C is a procedure in which tissue is removed from a uterus. Dilation and Curettage (D&C),Mayo Clinic (Nov. 7, 2023), https://www.mayoclinic.org/tests-procedures/dilation-and-curettage/about/pac-20384910 [https://perma.cc/67NN-AFL4].

[116]. Arey et al., supra note 113, at 390.

[117]. Id.

[118]. Id.

[119]. See generally Dov Fox, Medical Disobedience, 136 Harv. L. Rev. 1032 (2023).

[120]. Selena Simmons-Duffin, Doctors Who Want to Defy Abortion Laws Say It’s Too Risky,NPR (Nov. 23, 2023), https://www.npr.org/sections/health-shots/2022/11/23/1137756183/doctors-who-want-to-defy-abortion-laws-say-its-too-risky [https://perma.cc/23JX-BF53].

[121]. See, e.g.,Noel León, Introducing the Abortion Defense Network, Nat’l Women’s L. Ctr. (Mar. 8, 2023), https://nwlc.org/introducing-the-abortion-defense-network/ [https://perma.cc/CG5E-PHYF].

[122]. See Alan Braid, Opinion, Why I Violated Texas’s Extreme Abortion Ban,Wash. Post (Sep. 18, 2021), https://www.washingtonpost.com/opinions/2021/09/18/texas-abortion-provider-alan-braid/ [https://perma.cc/C3W9-5NQX]; see also Surana, supra note 106 (“Doctors described the position they’ve been put in — denying abortions to high-risk patients who are begging for them — as ‘distressing,’ ‘untenable’ and ‘insane.’ Speaking out about the broken system felt like the only way to not be complicit, Osmundson said. ‘It’s going to take physicians coming together and saying: “We’re not going to participate in this. We’re going to do what we think is right for patients.”’”).

[123]. Simmons-Duffin, supra note 120.

[124]. See Sabbath, McKetchnie & Arora, supra note 106 (“Many participants described needing to delay medically necessary care until patients were at risk of death or permanent impairment, or the fetal heart stopped spontaneously. One said, ‘The way our legal teams interpreted it, until they became septic or started hemorrhaging, we couldn’t proceed . . . [it] puts women in a very challenging, risky position. Is a 5% risk of death enough? Does it take 20%? Does it take 50%? What is enough legally? And the legal people seem to have a different definition that also just feels horrible, to say until you’re at a greater than likely chance of dying, you can’t make a decision.’”).

[125]. See Maria Godoy, Most Americans Support Abortion for Pregnancy-Related Emergencies, NPR (Mar. 7, 2024), https://www.npr.org/sections/health-shots/2024/03/07/1236344392/abortion-politics-poll-miscarriage-pregnancy-emergencies-kff# [https://perma.cc/NFT7-RFFE].

[126]. Abortion Laws,Tex. State L. Libr.(Aug. 21, 2025), https://guides.sll.texas.gov/abortion-laws/civil-penalties [https://perma.cc/G9PT-XQDT].

[127]. See Brittni Frederiksen, Usha Ranji, Ivette Gomez & Alina Salganicoff, KFF, A National Survey of OBGYNs’ Experiences After Dobbs (2023), https://www.kff.org/womens-health-policy/report/a-national-survey-of-obgyns-experiences-after-dobbs/ [https://perma.cc/XX8T-WUEU] (“Over four in ten (42%) OBGYNs report that they are very or somewhat concerned about their own legal risk when making decisions about patient care and the necessity of abortion. This rises to more than half of OBGYNs practicing in states with gestational limits (59%) and abortion bans (61%).”); see also Mariel Padilla, Abortion Bans Are Causing ‘Chilling Effect’ for OBGYNs, Study Says, 19th (June 21, 2023), https://19thnews.org/2023/06/obgyns-abortion-miscarriages-study/ [https://perma.cc/QMY5-L6WE] (discussing the KFF National Survey of OBGYNs’ Experiences After Dobbs to show a “chilling effect” among doctors in states with abortion bans); Sabbath, McKetchnie & Arora, supra note 106, at 5–6 (finding that abortion bans have created heavy emotional, moral, and physical burdens on OB-GYNs who have to choose between upholding their duty of care and violating the law and explaining that “[i]n addition to widespread distress and anxiety, the state of hypervigilance we observed in several participants—constantly worrying about potential consequences of providing care or counseling—increases the risk for longer-term physical and mental health problems”).

[128]. J. David Goodman, Abortion Ruling Keeps Texas Doctors Afraid of Prosecution, N.Y. Times (Dec. 13, 2023), https://www.nytimes.com/2023/12/13/us/texas-abortion-doctor-prosecution.html [https://perma.cc/2AUJ-BK9G].

[129]. “Molar pregnancy happens when a fertilized egg has too many chromosomes. It does not develop into a viable fetus. It is usually a benign condition, but in about 15% of cases, like Jaci’s, it is cancerous. Her doctor told her she was at risk of hemorrhage and even death, but that she couldn’t get treated there. The treatment for a patient in her condition is a dilation and curettage or D&C [—] an abortion procedure that clears pregnancy tissue from the uterus.” Selena Simmons-Duffin, supra note 4.

[130]. Pam Belluck, Legal Actions Seek Guarantee of Abortion Access for Patients in Medical Emergencies, N.Y. Times (Sep. 12, 2023), https://www.nytimes.com/2023/09/12/health/abortion-rights-lawsuits.html [https://perma.cc/6G83-U2EH].

[131]. Selena Simmons-Duffin, How Hospitals Decide What Qualifies as a Life Threatening Emergency to Allow Abortion,NPR (Apr. 25, 2023), https://www.npr.org/2023/04/25/1172005589/how-hospitals-decide-what-qualifies-as-a-life-threatening-emergency-to-allow-abo [https://perma.cc/34DS-8N8E].

[132]. Selena Simmons-Duffin, ‘I’ll Lose My Family.’ A Husband’s Dread During an Abortion Ordeal in Oklahoma, NPR (May 1, 2023), https://www.npr.org/sections/health-shots/2023/05/01/1172973274/oklahoma-abortion-ban-exception-life-of-mother-molar-pregnancy [https://perma.cc/AT9N-2MBG].

[133]. Belluck, supra note 130.

[134]. Kitchener & Diamond, supra note 58.

[135]. Id. Statton also filed an Emergency Medical Treatment & Labor Act (EMTALA) complaint, which was rejected. Her medical records noted that she “desired discharge home”—a fact that she found infuriating: “‘I was not okay, I was not fine,’ said Statton, who ultimately traveled out of state to get an abortion. ‘I remember being mad whenever I read [the medical records].’” Id.

[136]. Belluck, supra note 130.

[137]. Plaintiffs’ First Amended Verified Petition for Declaratory Judgment & Application for Temporary & Permanent Injunction ¶¶ 85, 87, Zurawski v. State, No. D-1-GN-23-000968 (Tex. Dist. Ct. Aug. 4, 2023).

[138]. Id. ¶¶ 88–89.

[139]. Id. ¶¶ 91–92.

[140]. See generally Moyle v. United States, 144 S. Ct. 2015 (2024).

[141]. Julie Luchetta, As Emergency Airlifts for Pregnant Patients Increases in Idaho, U.S. Supreme Court Abortion Case Starts, Boise St. Pub. Radio (Apr. 24, 2024), https://www.boisestatepublicradio.org/health/2024-04-24/moyle-idaho-supreme-court-airlift-abortion-emergency [https://perma.cc/E6AQ-JLJ2].

[142]. Id.

[143]. Id.

[144]. SeeThis Is How Much Air Medical Transport Costs,ABA Ins., https://www.abainsurance.com/resource-center/education-center/this-is-how-much-air-medical-transport-costs/ [https://perma.cc/VUU5-RFVL] (“[T]he average air ambulance cost for a 52 mile trip falls somewhere between $12,000 and $25,000 per flight which can reach as high as $6 million depending on the medical equipment and maintenance. . . . A study conducted by the Montana Legislature shows that air ambulance costs are generally divided into two main parts: the liftoff fees and per-mile charges. Now, while the former ranges between $8,500 to $15,200, the latter can be anywhere from $26 to $133 per mile.”). With these averages, a 300-mile trip from one hospital to another could easily run upwards of $50,000.

[145]. Payment of membership costs between $60 and $85 per year and private insurance is required to qualify. Previously, these policies were recommended to people who recreate outdoors in remote areas or who ride motorcycles, but now pregnancy has been added to the list of statuses and activities with additional recommendations for caution and airlift access. As of April 2024, one company has already seen an increase in membership requests. See Kelcey Mosley-Morris, Loss of Federal Protection in Idaho Spurs Pregnant Patients to Plan for Emergency Air Transport, Idaho Cap. Sun (Apr. 23, 2024),  https://idahocapitalsun.com/2024/04/23/loss-of-federal-protection-in-idaho-spurs-pregnant-patients-to-plan-for-emergency-air-transport/ [https://perma.cc/6R3K-VVLX].

[146]. Id.

[147]. Id.

[148]. Complaint at 1, Farmer v. Univ. of Kan. Health Sys., No. PA1594-23 (Kan. Hum. Rts. Comm’n Dec. 28, 2022).

[149]. Administrative Complaint ¶¶ 20, 29, 37–39, Farmer v. Freeman Health Sys. (U.S. Dep’t of Health & Hum. Servs. Nov. 8, 2022), https://nwlc.org/wp-content/uploads/2022/11/2022.11.08-Mylissa-Farmer-EMTALA-complaint.pdf [https://perma.cc/D3ZY-TQTR].

   [150]. Complaint of Discrimination at 1, Farmer v. Freeman Hosp. W. (Miss. Comm’n on Hum. Rts. Dec. 19, 2022), https://nwlc.org/wp-content/uploads/2022/11/Missouri-Commission-on-Human-Rights-Complaint-Against-Freeman-Hospital-West-2.pdf [https://perma.cc/R9B5-C6ES].

[151]. Id. at 2.

[152]. Id.

[153]. Complaint, supra note 148, at 2.

[154]. Id.

[155]. Id.

[156]. Id.

[157]. Id.

   [158]. Anna Spoerre, ‘Care Delayed and Care Denied’: Doctor Recalls 30 Months Under Missouri Abortion Ban,Mo. Indep.(Dec. 2, 2024), https://missouriindependent.com/2024/12/02/dr-betsy-wickstrom-ob-abortion-ban-missouri-amendment-3/ [https://perma.cc/VCP2-KHVT].

[159]. Id.

[160]. Id.

[161]. Id.

[162]. Id.

[163]. Id.

[164]. See Claire Wilkinson, Abortion Ruling Complicates Med Mal Insurance Environment, Bus. Ins. (Feb. 7, 2023), https://www.businessinsurance.com/abortion-ruling-complicates-med-mal-insurance-environment/ [https://perma.cc/V8LC-75ME] (“Any medical professional who interacts with pregnant mothers is potentially exposed, said Eric J. Gardzina, Nashville, Tennessee-based senior vice president, risk management, at Ob Hospitalist Group, which employs more than 1,300 obstetrician-gynecologists in 37 states under 250 hospital contracts. OB-GYNS, emergency department physicians and pharmacists are clearly at risk, but other medical professionals such as technicians, in vitro fertilization providers, oncologists, facilitators and vendors could also be exposed, he said. . . . Ob Hospitalist Group decided to defend its employees through the process, with ‘corporate dollars or whatever it takes,’ he said.”).

[165]. “Physicians [quoted in a 2024 Senate Finance Committee report] described hospital lawyers who ‘refused to meet’ with them for months, were ‘pretty much impossible’ to reach during ‘life or death’ scenarios and offered little help beyond ‘regurgitating’ the law . . . . Doctors described how other doctors gave out wrong and potentially harmful information, saying that patients could not legally choose their own course of treatment and that doctors could not legally treat ectopic pregnancies, potentially fatal complications in which an embryo develops outside the uterine cavity.” Kavitha Surana, Report: Hospitals Rarely Advise Doctors on How to Treat Patients Under Abortion Bans, ProPublica (Dec. 19, 2024), https://www.propublica.org/article/abortion-ban-deaths-report-ron-wyden [https://perma.cc/LDG9-SRCC]; see also Staff of S. Fin. Comm., 118th Cong., Practicing Amid “a Minefield”: Emergency Reproductive Health Care Post-Dobbs 13–16 (2024).

   [166]. Staff of S. Fin. Comm., supra note 165, at 15.

   [167]. Surana, supra note 106.

[168]. Id.

[169]. Id.

[170]. Id.

[171]. Id.

[172]. Id.

[173]. Id.

[174]. Id. When maternal-fetal medicine specialist Dr. Sarah Osmundson wrote to her colleagues on the abortion approval committee on behalf of a woman who was fourteen weeks pregnant and whose fetus was developing without a skull, her colleagues “urged her to consider the optics” of providing an abortion and replied that they were not “brave enough.” Id. The fetus’s condition put the patient at increased risk for “severe buildup of amniotic fluid, which could cause her uterus to rupture and possibly kill her.” Id.

[175]. Id.

[176]. Id.

[177]. See generally Staff of S. Fin. Comm., supra note 165.

[178]. De Vos et al., supra note 103, at 13 (“At the start of every call, the callers asked hospital staff, ‘What are your hospital’s procedures if I were to face a medical emergency while pregnant that jeopardizes my life?’”).

[179]. Id. at 12; see also Simmons-Duffin, supra note 131.

[180]. De Vos et al., supra note 103, at 12.

[181]. Id. at 13–16.

[182]. Henna K. Pithia, Patient Dumping: The Cobra that Never Struck, 24S. Cal. Rev. L. & Soc. Just. 109, 110, 116, 119 (2014).

[183]. Understanding EMTALA, Am. Coll. Emergency Physicians, https://www.acep.org/life-as-a-physician/ethics--legal/emtala/emtala-fact-sheet [https://perma.cc/YN7E-ZX28].

[184]. 144 S. Ct. 2015, 2016–17 (2024).

[185]. Kitchener & Diamond, supra note 58 (“EMTALA is an imperfect tool for the post-Roe landscape, several government officials and abortion rights advocates said, in part because of the narrowness of the law. ‘If you were starting from scratch on protecting abortion rights in this country, you’d never build [the strategy] around EMTALA,’ said an administration official who spoke on the condition of anonymity because of litigation around the Biden administration’s abortion guidance. ‘That we’re here . . . is a testament to the failings of Congress and the realities of the courts.’”).

[186]. NWLC Applauds CMS’ Action on Mylissa Farmer’s Emergency Abortion Complaint, Finding Hospitals Violated Federal Law, Nat’l Women’s L. Ctr. (May 1, 2023), https://nwlc.org/press-release/nwlc-applauds-cms-action-on-mylissa-farmers-emergency-abortion-complaint-finding-hospitals-violated-federal-law/ [https://perma.cc/PBF7-XZ89]; Erica Cerutti, OU Health Did Not Violate EMTALA in Abortion Case: HHS,Becker’s Hosp. Rev. (Jan. 22, 2024), https://www.beckershospitalreview.com/legal-regulatory-issues/ou-health-did-not-violate-emtala-in-abortion-case-hhs.html [https://perma.cc/YK9B-2FYY].

[187]. NWLC Applauds CMS’ Action on Mylissa Farmer’s Emergency Abortion Complaint, Finding Hospitals Violated Federal Law, supra note 187.

[188].Press Release, U.S. Dep’t Health & Hum. Servs., Robert F. Kennedy, Jr. Sworn in as 26th Secretary at HHS, President Trump Signs Executive Order to Make America Healthy Again (Feb. 13, 2025), https://www.hhs.gov/about/news/2025/02/13/robert-kennedy-jr-sworn-26th-secretary-hhs-president-trump-signs-executive-order-make-america-healthy-again.html [https://perma.cc/A5GZ-YTUN].

[189]. The Emergency Medical Treatment and Labor Act (EMTALA), U.S. Dep’t Health & Hum. Servs.: Off. of Inspector Gen. (Sep. 11, 2024), https://oig.hhs.gov/reports/featured/emtala/ [https://perma.cc/VMJ9-PFHJ].

[190]. SeeCMS Rescinds Biden-era Guidance on EMTALA, Ctr. for Repro. Rts.(June 3, 2025), https://reproductiverights.org/news/cms-rescinds-biden-era-guidance-on-emtala/ [https://perma.cc/BQ8X-ATFP] (detailing CMS’ rescinding of Biden-era guidance documents on EMTALA because the prior guidance does not ‘reflect the policy’ of the Trump administration); see alsoTimeline of Trump’s Attacks on Our Rights, Planned Parenthood Votes,https://www.plannedparenthoodaction.org/electoral/trump-timeline-of-attacks [https://perma.cc/6QJ8-KKUT] (chronicling President Trump’s use of administrative agencies to hamper access to abortion during his first term).

[191]. See Press Release, ACLU, ACLU Statement on DOJ Abandoning Fight to Protect Emergency Abortion Care for Pregnant Patients (Mar. 4, 2025),https://www.aclu.org/press-releases/aclu-statement-on-doj-abandoning-fight-to-protect-emergency-abortion-care-for-pregnant-patients [https://perma.cc/9QLF-SVLY].

[192]. See Press Release, Catherine Cortez Masto, Sen. for Nev., At Confirmation Hearing, Cortez Masto Spars with RFK Jr. on Commitment to Protecting Abortion Access, Lowering Drug Prices(Jan. 29, 2025), https://www.cortezmasto.senate.gov/news/press-releases/at-confirmation-hearing-cortez-masto-spars-with-rfk-jr-on-commitment-to-protecting-abortion-access-lowering-drug-prices/ [https://perma.cc/4KXD-DBJR] (“Senator Cortez Masto asked Mr. Kennedy to acknowledge that a woman having a heart-attack had the right, under a 40-year old federal law, to be given emergency care in hospitals that receive Medicare payments. Mr. Kennedy agreed that she did. Then, Cortez Masto asked ‘A pregnant woman with life-threatening bleeding from an incomplete miscarriage goes to the ER, and her doctor determines that she needs an emergency abortion. But, she’s in a state where abortion is banned. You would agree also . . . that federal law protects her right to that emergency care, correct?’ ‘Uh . . . I don’t know,’ Mr. Kennedy responded.”). When Senator Cortez Masto pressed Mr. Kennedy about his authority to enforce EMTALA, he also demonstrated a glaring lack of understanding of his enforcement power of EMTALA as secretary of HHS. SeePressed by Lawmakers, R.F.K. Jr. Fumbles Details of Key Health Programs, N.Y. Times (May 5, 2025),https://www.nytimes.com/live/2025/01/29/us/rfk-jr-health-senate-hearing (on file with the California Law Review).

[193]. See Bono, Wermuth & Hipskind, supra note 73, at 1.

[194]. Harrison Complaint, supra note 63, at ¶¶ 131–38. This claim is made to the best of the author’s knowledge and research in this area at the time of writing.

[195]. Intentional Infliction of Emotional Distress, Corn. L. Sch.: Legal Info. Inst., https://www.law.cornell.edu/wex/intentional_infliction_of_emotional_distress [https://perma.cc/V6QX-H94W]; NIED, Corn. L. Sch.: Legal Info. Inst., https://www.law.cornell.edu/wex/nied [https://perma.cc/AP63-EZCF].

[196]. Intentional Infliction of Emotional Distress, supra note 195.

[197]. NIED, supra note 195.

[198]. Id.

[199]. See e.g.,Belluck, supra note 130 (arguing that Jaci Statton’s experience of almost passing out in her kitchen and realizing that her pants were soaked with blood yet being told that she should wait in the hospital parking lot until her conditioned worsened enough to be treated, by which point she was “crashing in front of [hospital staff],” presents a strong case for an IIED or NIED claim).

[200]. Roe Complaint, supra note 66, ¶¶ 107–21; Harrison Complaint, supra note 63, ¶¶ 111–30; Nusslock Complaint, supra note 66, ¶¶ 83–98.

[201]. Harrison Complaint, supra note 63, ¶¶ 1–8.

[202]. Id. ¶¶ 42–52.

[203]. Wrongful Death, Corn. L. Sch.: Legal Info. Inst., https://www.law.cornell.edu/wex/wrongful_death [https://perma.cc/6UV9-ZYV6].

[204]. Stephania Taladrid, Did an Abortion Ban Cost a Young Texas Woman Her Life?, New Yorker (Jan. 8, 2024), https://www.newyorker.com/magazine/2024/01/15/abortion-high-risk-pregnancy-yeni-glick [https://perma.cc/L8Y7-A9NQ].

[205]. Id.

[206]. Id.

[207]. Id.

[208]. Id.

[209]. Id.

[210]. Id.

[211]. See Tex. Civ. Prac. & Rem. Code Ann. § 71.002(c)(2).

[212]. See id.

[213]. See generally Wilson Elser Moskowitz Edelman & Dicker LLP, Wrongful Death Claims: 2023 50-State Comparative Law Review (2023), https://www.wilsonelser.com/publications/wrongful-death-claims-a-50-state-comparative-law-review-2023 [https://perma.cc/S3ET-VGPP] (surveying the variance in requirements and limitations for wrongful death claims).

[214]. Id.

[215]. Lawsuit Says a Black Patient Bled to Death Because of a Hospital’s Culture of Racism, NPR (May 5, 2022), https://www.npr.org/2022/05/05/1096833756/racism-lawsuit-cedars-sinai-medical-center-wife-death [https://perma.cc/CN4P-GVX7].

[216]. Id.

[217]. Id.

[218]. Id.

[219]. Deena Zaru & Brittany Gaddy, Cedars-Sinai Medical Center Facing Federal Probe Over Treatment of Black Mothers, ABC News (July 12, 2023), https://abcnews.go.com/US/cedars-sinai-medical-center-facing-federal-probe-treatment/story?id=101165260 [https://perma.cc/N82T-RMPS]; Family of Black Woman Who Died After ‘Sloppy’ C-section Sues Hospital for Racism,Today (May 9, 2022), https://www.today.com/parents/parents/black-woman-died-c-section-racism-lawsuit-cedars-sinai-rcna28029 [https://perma.cc/AKT3-TX4Y]; Governor Newsom Signs Legislation to Modernize California’s Medical Malpractice System, Governor Gavin Newsom (May 23, 2022), https://www.gov.ca.gov/2022/05/23/governor-newsom-signs-legislation-to-modernize-californias-medical-malpractice-system/ [https://perma.cc/WL52-77KY].

[220]. See Zaru & Gaddy, supra note 219.

[221]. Morning Edition, Insurance Refused to Pay for Her Abortion, Even When Her Life Was at Stake, NPR (Aug. 26, 2024), https://www.npr.org/sections/shots-health-news/2024/08/26/nx-s1-5068276/abortion-ban-exception-health-insurance [https://perma.cc/6R5E-BJ3A].

[222]. Jennifer Arlen & W. Bentley MacLeod, Malpractice Liability for Physicians and Managed Care Organizations, 78 N.Y.U. L. Rev. 1929, 1932 (2003).

[223]. Managed Care, CDC: Nat’l Ctr. for Health Stat. (Aug. 12, 2022), https://www.cdc.gov/nchs/hus/sources-definitions/managed-care.htm [https://perma.cc/82ZY-A8V4].

[224]. Managed Care Liability Insurance, IRMI, https://www.irmi.com/term/insurance-definitions/managed-care-liability-insurance [https://perma.cc/3AET-PW8H]. PPO plans are the most common type of health plan in group health insurance that employers offer, with 47 percent of covered workers enrolled in PPOs and 13 percent of covered workers enrolled in HMOs in 2023. 2023 Employer Health Benefits Survey, Kaiser Fam. Found. (Oct. 18, 2023), https://www.kff.org/report-section/ehbs-2023-section-5-market-shares-of-health-plans/#figure51 [https://perma.cc/XWV8-S5G5].

[225]. Arlen & MacLeod, supra note 222, at 1942, 1944.

[226]. See, e.g., Craig C. New & Chris Dominic, Us and Them: A Comparison of Juror Attitudes in Oregon and the Northwest, Or. state barhttps://www.osbar.org/publications/bulletin/03dec/jurors.html[https://perma.cc/2FHX-6JRS] (“Respondents were asked which side they would favor if a person was suing a corporation. To this question, 60 percent of the respondents reported favoring the plaintiff, 24 percent favored neither party and 16 percent favored the defendant.”).

[227]. See, e.g.,Helen Santoro, Health Insurers’ $371 Billion Windfall, Lever (Dec. 11, 2024), https://www.levernews.com/health-insurers-371-billion-windfall/ [https://perma.cc/3CLL-XULQ].

[228]. See, e.g., Patient Protections in Managed Care: Hearing Before the Subcomm. on Health of the H. Comm. on Ways and Means, 107th Cong. 20 (2001) (“The only noneconomic damage cap in Texas applies in statutorily created medical malpractice actions for wrongful death. That cap is adjusted for inflation and the 2000 cap amount is $1,410,000. This cap does not apply to a cause of action against a managed care insurer.”).

[229]. See generally Arthur F. Southwick, Vicarious Liability of Hospitals, 44 Marq. L. Rev. 153 (1960); see also Bolton v. Willis-Knighton Med. Ctr., 116 So. 3d 76, 83–84 (La. Ct. App. 2013) (“It is well settled that a hospital is liable for its employee’s negligence, including its doctors and nurses, under the respondeat superior doctrine. In a medical malpractice claim against a hospital, the plaintiff is required to prove by a preponderance of the evidence, as in any negligence action, that the defendant owed the plaintiff a duty to protect against the risk involved (or the applicable standard of care), and the injury was caused by the breach. Whether an emergency room physician is an employee or an independent contractor is a factual issue turning on the control exercised by the hospital over his activities. The distinction between employee and independent contractor status is a factual determination to be decided on a case-by-case basis.” (citations omitted)).

[230]. See Magallanes v. Doctors Med. Ctr. of Modesto,295 Cal. Rptr. 3d 828, 830, 834–35 (Ct. App. 2022) (holding that, because the patient did not rely on an apparent agency relationship between the hospital and the physician to receive care, the hospital was not liable for medical negligence).

[231]. A. Shenoy, G.N. Shenoy & G.G. Shenoy, Respondeat Superior in Medicine and Public Health Practice: The Question Is – Who Is Accountable for Whom?, 17 Ethics Med. & Pub. Health 1, 4 n.17 (2021).

[232]. See, e.g., Simmons-Duffin, Her Miscarriage Left Her Bleeding Profusely, supra note 2.

[233]. Nancy R. Levin, Hospital’s Liability for Independent Emergency Room Service, 22 Santa Clara L. Rev. 791, 794 (1982).

[234]. Id. at 795.

[235]. Id. at 793.

[236]. Id. at 795.

[237]. Id. at 799.

[238]. See Joseph Wood, Key Contract Issues for Emergency Physicians, Am. Acad. of Emergency Med. RSA, https://www.aaemrsa.org/education/residents/key-contract-issues [https://perma.cc/LP3Q-PYVM].

[239]. See generally Mitchell J. Nathanson, Hospital Corporate Negligence: Enforcing the Hospital’s Role of Administrator, 28 Tort & Ins. L.J. 575, 578 (1993);Erika L. Amarante, Recent Trends and Defense Strategies: Corporate Liability for Hospitals, 2016 For the Defense 8.

[240]. Amarante, supra note 239, at 9–10.

[241]. See Nathanson, supra note 239, at 578; 211 N.E.2d 253 (Ill. 1965).

[242]. Nathanson, supra note 239, at 578; see alsoDarling, 211 N.E.2d at 257 (“Present-day hospitals, as their manner of operation plainly demonstrates, do far more than furnish facilities for treatment. They regularly employ on a salary basis a large staff of physicians, nurses and internes, as well as administrative and manual workers, and they charge patients for medical care and treatment, collecting for such services, if necessary, by legal action. Certainly, the person who avails himself of ‘hospital facilities’ expects that the hospital will attempt to cure him, not that its nurses or other employes will act on their own responsibility.” (citation omitted)).

[243]. Nathanson, supra note 239, at 579.

[244]. Id. at 576 n.2. (“Prior to the abolishment of charitable immunity, many courts rationalized their reluctance to pin liability on hospitals in part on the notion that they were not health care providers at all. Rather, they were considered merely buildings that provided the equipment and facilities necessary to allow physicians to practice.”).

[245]. Est. of Essex v. Grant Cnty. Pub. Hosp. Dist. No. 1, 546 P.3d 407, 411 (Wash. 2024) (quoting Adamski v. Tacoma Gen. Hosp., 579 P.2d 970, 974 (1978)).

[246]. Id.

[247]. Id.

[248]. Nathanson, supra note 239, at 575.

[249]. Gary F. Loveridge & Betsy S. Kimball, Hospital Corporate Negligence Comes to California: Questions in the Wake of Elam v. College Park Hospital, 14 Pac. L.J. 803, 809 n.47 (1983).

[250]. Gafner v. Down E. Cmty. Hosp., 735 A.2d 969, 979–80 (Me. 1999).

[251]. See Pedroza v. Bryant, 677 P.2d 166, 172 (Wash. 1984); see also David B. Robbins, Cara Wallace, Malori Basye, Juliana Lehua Bennington & Adrianna Simonelli, Expansion of Hospital Tort Liability in Washington, Perkins Coie (Apr. 18, 2024), https://perkinscoie.com/insights/update/expansion-hospital-tort-liability-washington [https://perma.cc/B6WB-KLPG].

[252]. 546 P.3d 407, 411 (Wash. 2024).

[253]. Id. at 414.

[254]. Thompson v. Nason Hosp., 591 A.2d 703, 707 (Pa. 1991)

[255]. See Surana, supra note 165.

[256]. Id. (quoting Staff of S. Fin. Comm., supra note 165, at 2).

[257].Surana, supra note 165(“[M]any of the hospitals were relying on guidance created before the existence of abortion bans, the report said. In most cases, physicians were given basic EMTALA guidance that didn’t discuss how to handle new abortion restrictions and were told to contact legal or ethics counsel for questions. Only a few hospitals had created proactive guidance to help their providers navigate the new landscape, and only two of those explicitly discussed conflicts that exist between abortion bans and EMTALA and how to handle them. It was not always clear if these directives were created before media reports of denied care.”).

[258]. See generally 328 S.W.3d 829 (Tenn. 2010) (holding that hospital was liable for failing to follow its internal policy requiring all emergency room patients to be seen by a physician, as a result of patient’s death from cardiac infection following discharge with a diagnosis of a sprain). See also Amarante, supra note 239, at 3.

[259]. Barkes, 328 S.W.3d at 830.

[260]. Id. at 831.

[261]. Id.

[262]. Id. at 832.

[263]. Id.

[264]. Id. at 835.

[265]. Id.

[266]. Id. at 836.

[267]. Thompson v. Nason Hosp., 591 A.2d 703, 708 (Pa. 1991).

[268]. See generally Upjohn Co. v. United States, 449 U.S. 383 (1981) (expanding the availability of attorney-client privilege in the corporate context beyond merely those who play a substantial role in deciding and directing a corporation’s legal response).

[269]. See generally Arnold S. Relman, The New Medical-Industrial Complex,303 NEJM 963 (1980).

[270]. See generally id.; see also Steffie Woolhandler, David U. Himmelstein, Adam W. Gaffney & Danny McCormick, Opinion, The U.S. Experiment with Profit-Driven Health Care Has Failed, STAT (Oct. 21, 2025), https://www.statnews.com/2025/10/21/health-care-system-profit-failed/ [https://perma.cc/4DXL-NHPZ].

[271]. See Maya Brownstein, Private Equity’s Appetite for Hospitals May Put Patients at Risk, Harv. Sch. of Pub. Health (Dec. 16, 2024), https://hsph.harvard.edu/news/private-equitys-appetite-for-hospitals-may-put-patients-at-risk/ [https://perma.cc/6Q3M-LCVY].

[272]. See supra Part I.C.

[273]. See Alan Condon, 65 Health Systems Ranked by Annual Revenue, Becker’s Hosp. Rev. (May 16, 2025), https://www.beckershospitalreview.com/finance/38-health-systems-ranked-by-annual-revenue/ [https://perma.cc/R6YR-65BD] (listing health systems by their annual revenue, between $115.8 billion and $737 million); see also Jeremia Kimelman, A Record Amount Went to Lobbying California’s Government. Who Were the Biggest Spenders?, Cal Matters (Feb. 12, 2024), https://calmatters.org/politics/2024/02/california-lobbying-state-government/ [https://perma.cc/NN55-9LTG] (showing that the California Hospital Association, a trade group representing over 400 hospitals and healthcare systems, spent over $3.5 million on lobbying in 2023).

[274]. See, e.g.,Mike Dennison, Hospitals’ Lobbying Frustrates Montana Lawmakers Who Sought to Boost Oversight, KFF Health News (Apr. 23, 2025), https://kffhealthnews.org/news/article/hospital-lobbying-montana-state-legislators-medicaid-expansion-oversight-community-benefit/ [https://perma.cc/78CG-HZTK].

[275]. See generally Gary T. Dance, Medical Malpractice: Prelitigation Screening Panels in Idaho, 19 Idaho L. Rev. 31 (1983).

[276].Idaho Div. Occupational & Pro. Licenses, Medical Malpractice Prelitigation Claim Form1, https://apps-dopl.idaho.gov/IBOMPortal/LINKS/prelitigation/prelit_form.pdf [https://perma.cc/G6HF-4WQX].

[277]. Id.

[278].Idaho Div. Occupational & Pro. Licenses, We’re Searching for Volunteers, https://apps-dopl.idaho.gov/IBOMPortal/LINKS/prelitigation/volunteers_wanted.pdf [https://perma.cc/5VVD-6R3V].

[279]. Idaho Medical Malpractice Laws, Gilman & Bedigian, LLC, https://www.gilmanbedigian.com/idaho-medical-malpractice-laws/ [https://perma.cc/GYC8-RM6C].

[280]. Id.

[281]. Id.; Idaho Div. Occupational & Pro. Licenses, supra note 276.

[282]. Idaho Medical Malpractice Laws, supra note 279.

[283]. Id.

[284]. Prelitigation: General Information, Utah Com.: Div. of Pro. Licensing, https://dopl.utah.gov/prelitigation/general-information/ [https://perma.cc/2D5K-H8AG].

[285].Nat’l Ass’n of Benefits & Ins. Pros., Malpractice Damage Caps by State, https://nabip.org/media/8331/medical_malpractice_cap.pdf [https://perma.cc/Q29D-M49Z].

[286]. Harrison Complaint, supra note 63, ¶ 15.

[287]. Id.

[288]. James Staas, Former Tonawanda Dispatcher Awarded $9.1 Million in Medical Malpractice Suit, Buffalo News (Feb. 17, 2014), https://buffalonews.com/news/article_3892fb83-1c50-5876-8ae7-f32723e7d7e6.html [https://perma.cc/J9FK-246M].

[289]. See generally Erie R. Co. v. Tompkins, 304U.S. 64 (1938). In January 2026, the Supreme Court decided Berk v. Choy, which discusses state-level medical malpractice requirements and choice-of-law questions. 146 S. Ct. 546 (2026). However, because substantive edits to this Note were completed in October 2025, this Note does not explore the impact of this decision.

[290]. See, e.g., Thompson v. Nason Hosp., 591 A.2d 703, 707 (Pa. 1991).

[291]. Donna Vanderpool, The Standard of Care, Innov. in Clin. Neurosci., Summer (July–Sep.) 2021, at 50, 51.

[292]. Brian K. Cooke, Elizabeth Worsham & Gary M. Reisfield, The Elusive Standard of Care, 45 J. Am. Acad. Psychiatry & L. 358, 359 (2017).

[293]. Id. at 359–60.

[294]. Id. at 361.

[295]. Id.

[296]. Id.

[297]. Id.

[298]. “The reproductive rights framework is largely centered on achieving women’s individualistic reproductive freedom through the legal system and has historically focused largely on the pro-choice and pro-life debate, sex education, and family planning.” Danielle M. Pacia, Reproductive Rights vs. Reproductive Justice: Why the Difference Matters in Bioethics, The Petrie-Flom Ctr.: Bill of Health (Nov. 3, 2020), https://blog.petrieflom.law.harvard.edu/2020/11/03/reproductive-rights-justice-bioethics/ [https://perma.cc/66NH-2D2S].

[299]. “SisterSong defines Reproductive Justice as the human right to maintain personal bodily autonomy, have children, not have children, and parent the children we have in safe and sustainable communities.” Reproductive Justice, SisterSong, https://www.sistersong.net/reproductive-justice [https://perma.cc/Z2HR-YRL9].

[300]. See Saul Elbein, Texas AG’s Office Argues Women Should Sue Doctors – Not State – Over Lack of Abortion Access, Hill (Nov. 28, 2023), https://thehill.com/policy/healthcare/4331412-texas-ags-office-argues-women-should-sue-doctors-not-state-over-lack-of-abortion-access/ [https://perma.cc/F5QN-C97D].

[301]. See Carmel Shachar, Susannah Baruch & Louise P. King, Opinion, Whose Responsibility Is It to Define Exceptions in Abortion Bans?, 331 JAMA 559, 559 (2024).

[302]. The Charlotte Lozier Institute is the nonprofit science and research arm of the Susan B. Anthony Pro-Life America Group, one of America’s most powerful anti-abortion organizations. It has been criticized for producing junk science that is designed to be used in anti-abortion litigation—several of its studies have been retracted by leading science journals. The retractions came due to uncleared conflicts of interest involving several of the papers’ authors, some of whom are plaintiffs or expert witnesses in a case questioning the Food and Drug Administration’s approval of mifepristone, one of the most commonly used medications in medication abortion worldwide. See Sofia Resnick, Retracted Studies the Latest in a Decadeslong Abortion-Science Fight,Colo. Newsline (Feb. 25, 2024), https://coloradonewsline.com/2024/02/25/retracted-studies-abortion-science-fight/ [https://perma.cc/X8NE-V8SW]. Researchers from the Charlotte Lozier Institute have also testified in support of criminal abortion bans in states like Kansas, arguing that felony crimes for abortion “would provide important protections for pregnant women in Kansas.” Sherman Smith, Kansas Abortion ‘Coercion’ Bill Touted by Out-of-State Think Tank That Produced Retracted Research, Kan. Reflector (Mar. 19, 2024) (citation omitted), https://kansasreflector.com/2024/03/19/kansas-abortion-coercion-bill-touted-by-out-of-state-think-tank-that-produced-retracted-research/ [https://perma.cc/GV4M-54NW].

   [303]. The Alliance for Hippocratic Medicine is an organization that was hastily created in 2022 which has been described by watchdog groups as a shoddy paper cover for Alliance Defending Freedom (ADF), one of the most prominent Christian legal advocacy groups in the United States today. The Alliance for Hippocratic Medicine was the lead plaintiff in a major Supreme Court mifepristone case, FDA v. Alliance for Hippocratic Medicine, 144 S. Ct. 1540 (2024). Melissa Gira Grant, Who Exactly Is Behind the Supreme Court’s Big Mifepristone Case?, New Republic (Mar. 7, 2024), https://newrepublic.com/article/179626/mifepristone-abortion-supreme-court-alliance-hippocratic-medicine [https://perma.cc/93LZ-XX44].

[304]. See Brief for Amici Curiae Charlotte Lozier Institute and Alliance for Hippocratic Medicine Supporting Appellants at 1–2, State v. Zurawski, 690 S.W.3d 644 (Tex. 2024) (No. 23-0629).

[305]. Id.

[306]. Id. at 2.

[307]. Eleanor Klibanoff, Kate Cox’s Case Reveals How Far Texas Intends to Go to Enforce Abortion Laws, Tex. Trib. (Dec. 13, 2023), https://www.texastribune.org/2023/12/13/texas-abortion-lawsuit/ [https://perma.cc/JBJ5-VFAM].

[308]. Kimberlee Kruesi, Asked to Clear up Abortion Bans, GOP Leaders Blame Doctors and Misinformation for the Confusion, Associated Press (Mar. 11, 2024), https://apnews.com/article/abortion-bans-exceptions-doctors-lawmakers-7cefca4a248076268c789ea05b367f1a [https://perma.cc/XHG2-NT2M].

[309]. “‘I would hate to think that St. Luke’s or any other hospital is trying to do something like this just to make a political statement, which I think is actually dangerous,’ Labrador said. In a written statement to the Statesman, St. Luke’s spokesperson Christine Myron said the hospital system stands by Souza’s statements. ‘We do not have any way of knowing who Attorney General Labrador spoke to related to out-of-state patient transfers for pregnancy complications, but what we can share with confidence is our data,’ Myron said.” Nicole Blanchard, Idaho Attorney General Labrador Questions Doctors’ Accounts of Abortion Emergencies, Idaho Statesman (Apr. 25, 2024), https://www.idahostatesman.com/news/politics-government/state-politics/article287981125.html [https://perma.cc/7RQT-9ZKT].

[310]. See Shachar, Baruch & King, supra note 301, at 560.

[311]. See Kimelman, supra note 273 (showing that the California Hospital Association, a trade group representing over 400 hospitals and healthcare systems, spent over $3.5 on million on lobbying in 2023).

[312].City of Akron v. Akron Ctr. for Reprod. Health, 462 U.S. 416, 427 (1983) (“The Court also has recognized, because abortion is a medical procedure, that the full vindication of the woman’s fundamental right necessarily requires that her physician be given ‘the room he needs to make his best medical judgment.’ The physician’s exercise of this medical judgment encompasses both assisting the woman in the decisionmaking process and implementing her decision should she choose abortion.” (citations omitted)), overruled by Planned Parenthood of Se. Pa. v. Casey,503 U.S. 833 (1992); Colautti v. Franklin, 439 U.S. 379, 387 (1979) (“Roe stressed repeatedly the central role of the physician, both in consulting with the woman about whether or not to have an abortion, and in determining how any abortion was to be carried out.”), abrogated by Dobbs v. Jackson Women’s Health Org., 142 S. Ct. 2228 (2022); Doe v. Bolton, 410 U.S. 179, 192 (1973) (“Whether, in the words of the Georgia statute, ‘an abortion is necessary’ is a professional judgment that the Georgia physician will be called upon to make routinely. We agree with the District Court that the medical judgment may be exercised in the light of all factors—physical, emotional, psychological, familial, and the woman’s age—relevant to the well-being of the patient. All these factors may relate to health. This allows the attending physician the room he needs to make his best medical judgment. And it is room that operates for the benefit, not the disadvantage, of the pregnant woman.” (citation omitted)), abrogated byDobbs, 142 S. Ct. 2228;United States v. Vuitch,402 U.S. 62, 72 (1971) (“[W]hether a particular operation is necessary for a patient’s physical or mental health is a judgment that physicians are obviously called upon to make routinely whenever surgery is considered.”).

[313]. See, e.g., Casey, 505 U.S. at 845 (noting abortion healthcare providers as parties), overruled byDobbs, 142 S. Ct. 2228; Rust v. Sullivan, 500 U.S. 173, 181(1991) (doctor was a plaintiff); Hodgson v. Minnesota, 497 U.S. 417, 429 (1990) (two doctors and four clinics among plaintiffs); Webster v. Reprod. Health Servs., 492 U.S. 490, 501–02 (1989) (five state-employed health professionals and two private nonprofit corporations providing abortion services); Thornburgh v. Am. Coll. of Obstetricians & Gynecologists, 476 U.S. 747, 752 (1986), overruled byCasey, 505 U.S. 833; Harris v. McRae, 448 U.S. 297, 302(1980) (hospital was a plaintiff); Bellotti v. Baird, 443 U.S. 622, 626 (1979) (doctor was a plaintiff); Singleton v. Wulff, 428 U.S. 106, 109 (1976) (two physician plaintiffs); see also Roe v. Wade, 410 U.S. 113, 120 (1973) (noting that James Hubert Hallford, a licensed physician, was an intervenor inRoe’s action, alleging he had been arrested for violating Texas’s abortion statutes), overruled byDobbs, 142 S. Ct. 2228.

[314]. Plaintiffs’ Original Petition for Declaratory Judgment and Application for Permanent Injunction, supra note 4, ¶¶ 327–34.

[315]. See generally David A. Grimes, Jacqueline D. Forrest, Alice L. Kirkman & Barbara Radford, Clinical Opinion, An Epidemic of Antiabortion Violence in the United States, 165 Am. J. Obstet. & Gynec. 1263 (1991). See alsoThe Murder of Barnett Slepian: 25 Years Later, Buffalo News (Oct. 23, 2023), https://buffalonews.com/news/local/the-murder-of-barnett-slepian-25-years-later/collection_7fffe252-7018-11ee-974a-ff166aeeb020.html [https://perma.cc/2YA9-SKEX] (collecting news articles regarding the murder of an abortion provider). For more recent accounts of violence, see Christine Hauser, California Man Gets 9 Years in Firebombing of Planned Parenthood Clinic,N.Y. Times (Apr. 16, 2024), https://www.nytimes.com/2024/04/16/us/planned-parenthood-firebomber-sentenced.html [https://perma.cc/G4JW-7NJX].

[316]. Roe, 410 U.S. at 165–66; C.E. Joffe, T.A. Weitz & C.L. Stacey, Uneasy Allies: Pro-Choice Physicians, Feminist Health Activists and the Struggle for Abortion Rights, 26 Socio. Health & Illness 775, 781 (2004) (citing Roe, 410 U.S. at 166).

[317]. See Emma Peterson & Daniel Martinez Hosang, How Doctors Came to Play a Key Role in the Abortion Debate,TIME (June 20, 2024), https://time.com/6989501/physician-abortion-access/ [https://perma.cc/K5AL-YSDG].

[318]. See supra Part I.C.

[319].Joffe, Weitz & Stacey, supra note 316, at 776–77.

[320]. Id. at 777.

[321]. Id.

[322]. Id. at 778.

[323]. About Us, AAPLOG, https://aaplog.org/about-us/ [https://perma.cc/5DH7-KVWX]; Pro-Life Directory, AAPLOG, https://aaplog.org/pro-life-directory/ [https://perma.cc/BV8M-Y59R].

[324]. About Us, supra note 323; What are Crisis Pregnancy Centers?, Planned Parenthood (Nov. 4, 2021), https://www.plannedparenthood.org/blog/what-are-crisis-pregnancy-centers [https://perma.cc/6L9S-RTNZ].

[325]. See generallyAbout Us, supra note 323. See also Bracey Harris, Meet the Anti-Abortion Group Using White Coats and Research to Advance its Cause, NBC NEWS (June 7, 2024), https://www.nbcnews.com/news/us-news/charlotte-lozier-institute-anti-abortion-research-influence-rcna151277 [https://perma.cc/LGR6-KLDY].

[326]. Coleman Drake, Marian Jarlenski, Yuehan Zhang & Daniel Polsky, Market Share of US Catholic Hospitals and Associated Geographic Network Access to Reproductive Health Services, 3 JAMA Network Open 1 (2020), https://pmc.ncbi.nlm.nih.gov/articles/PMC6991305/ [https://perma.cc/WG3U-FDAK].

[327]. Joffe, Weitz & Stacey, supra note 316, at 778.

[328]. See Harrison Complaint, supra note 63, ¶¶ 9, 37; Roe Complaint, supra note 66, ¶ 22.

[329]. See Joffe, Weitz & Stacey, supra note 316, at 780.

[330]. See Stephania Taladrid, The Post-Roe Abortion Underground, New Yorker (Oct. 10, 2022), https://www.newyorker.com/magazine/2022/10/17/the-post-roe-abortion-underground [https://perma.cc/L4XG-SPQB].

[331]. Joffe, Weitz & Stacey, supra note 316, at 780 (citing 2 Robert E. Hall, Abortion in a Changing World 109 (1970)).

[332]. See Avik Roy, How the Hospital Lobby Fights Against Affordable Health Care, Found. for Rsch. on Equal Opportunity, https://freopp.org/whitepapers/how-the-hospital-lobby-fights-against-affordable-health-care/ [https://perma.cc/5TR7-X9N4]; Kimelman, supra note 273.

[333].Am. Hosp. Ass’n, Federal Agencies with Regulatory or Oversight Authority Impacting Hospitals (2017), https://www.aha.org/system/files/2018-01/info-regulatory-burden-federal-agencies.pdf [https://perma.cc/LB96-VFLA].

[334]. About the AHA, Am. Hosp. Ass’n, https://www.aha.org/about [https://perma.cc/2K3C-MGAD].

[335]. See, e.g., AAMC, AHA Amicus Brief Supports U.S. Challenge to Idaho Abortion Law, AAMC (Aug. 18, 2022), https://www.aamc.org/advocacy-policy/washington-highlights/aamc-aha-amicus-brief-supports-us-challenge-idaho-abortion-law [https://perma.cc/66LJ-BMWB].

Table of Contents Show

Introduction

They’re making money off us like we’re not even human. But we’re human beings, and we deserve respect.

—A Hunger Strikers Handbook[1]

On March 7, 2014, 1,200 detainees of the Northwest Detention Center (NWDC)[2] stopped eating.[3] The hunger strike, while fluctuating in size, lasted fifty-six days.[4] Protestors cited a range of grievances, including spoiled and insufficient food, exorbitant commissary prices, and payment of one single dollar for each eight-hour day of work maintaining the facility.[5] Despite the prison-like conditions, none of the strikers were serving criminal sentences. Instead, they were awaiting resolution of their civil immigration proceedings.

NWDC is a sprawling structure built along a toxic waste site in Tacoma, Washington.[6] The 1,575 bed facility is one of the largest immigration detention centers in the United States.[7] A private, for-profit corporation obtained financing from the state to build the facility with the promise of creating local jobs.[8] And yet for years, the company[9] has avoided employing community members by relying on the captive labor force to clean, maintain, and operate the facility.[10] In the facility’s kitchen, The GEO Group (GEO) employed thirteen full-time outside workers but relied on nearly one hundred detainees every day to prepare meals, cook and serve food, and wash dishes.[11] The detainees also staffed the facility as janitors, barbers, and librarians—all for just one dollar per day.[12] Meanwhile, in 2018, GEO made $18.6 million in annual profits from the facility.[13]

Three and a half years after the 2014 facility-wide hunger strike, the Attorney General for the State of Washington filed a lawsuit against GEO alleging the company violated the state minimum wage and unjustly enriched itself.[14] In parallel, detained workers[15] brought their own private class action lawsuit.[16] In 2021, a jury found against GEO in both lawsuits on all counts.[17] Recently, both the Washington Supreme Court and the Ninth Circuit upheld these rulings.[18]

GEO met these victories with resistance. First, GEO appealed the jury’s decision. While the Ninth Circuit denied GEO’s petition for panel or en banc rehearing of its split decision,[19] GEO is already seeking review from the Supreme Court.[20] Second, GEO responded to the jury’s decision by ending the work program at NWDC entirely rather than paying detained workers a lawful wage.[21] Not only did GEO refuse to compensate detained workers fairly, but it also raised the already exorbitant commissary prices for essentials such as toilet paper.[22] Without near-free labor to maintain the facility, it fell into disrepair.[23] GEO then barred the Washington Department of Labor and Industries from inspecting the facility until a court order mandated compliance.[24]

GEO is fighting back against detainees’ labor rights inside and outside the courtroom because NWDC is not an anomaly. This dollar-per-day program is the standard at all immigration detention centers, which are largely run by private companies like GEO.[25] The economic model of profit-driven immigration detention depends on this near-free, captive workforce.[26] Requiring companies to comply with labor laws by challenging these programs in court could disrupt the massive private immigration detention industry. More broadly, recognizing detainees as workers empowers them with agency to assert their rights—not only in court but through organizing and leveraging state and federal agencies.[27] Litigation is one pathway to legally claim worker status and rights.[28] And these parallel lawsuits at NWDC are only two of the various actions challenging labor conditions in immigration detention centers with promising results.[29]

Litigating detained workers’ rights requires knowledge at the intersection of employment, immigration, and constitutional law. This Note provides an overview of the core issues in these legal areas and encourages collaboration in bringing this trans-substantive litigation, synthesizing all available causes of action. This Note relies on doctrinal research, including statutory law and jurisprudence, supplemented by anecdotal interviews. This Note analyzes past and pending litigation, existing academic literature, governmental and non-governmental reports, and investigative reporting. This research is supplemented by interviews with more than a dozen attorneys at private plaintiff-side firms, advocacy groups, and government agencies involved in litigating relevant cases. Finally, this Note draws upon interviews with two organizers at the helm of these efforts: one formerly detained, named plaintiff and one formerly undocumented activist.

Part I introduces private immigration detention, detainee labor, and the federal causes of action to address detainee labor conditions and analyzes case law developments. Part II reviews federal causes of action to address detainee labor conditions and explores recent case law developments. Part III analyzes existing and emerging opportunities to bring state causes of action. Commentators have not examined state law claims as thoroughly as federal claims, despite their success at trial and on appeal. Part III also analyzes the latest cases raising state law claims, including recent appellate decisions,[30] and identifies opportunities and obstacles when bringing these lawsuits. The Note concludes with an overview of both the challenges and possibilities of tackling the injustices of privatized immigration detention through labor lawsuits.

I. Private Immigration Detention, Detainee Labor, and the Role of Litigation

NWDC is only one of 129 detention facilities across the country.[31] As further discussed below, the private operation of NWDC is the norm in immigration detention. Detainees in these majority-private facilities labor under a work program established by U.S. Immigration and Customs Enforcement (ICE). All detainees are in custody of the federal government, but different local and federal government actors and agencies simultaneously challenge and uphold these work arrangements. Affirmative civil litigation, while imperfect, offers one path to addressing these labor conditions.

This Part discusses the massive scale and legal nuances of immigration detention in the United States. It outlines the privatization and consolidation of the for-profit immigration detention industry. It identifies the legal basis for working while detained and describes the lived realities of detainees. It sorts through the government’s role in both maintaining and reforming these labor dynamics. Finally, this Part introduces civil litigation as a central strategy in protecting detained workers’ rights.

A. U.S. Immigration Detention

I felt like my life was violated in every way, like my life didn’t matter.

—Jose Ruben Hernandez Gomez, formerly detained organizer and named plaintiff in Hernandez Gomez v. GEO Group.[32]

The United States operates the world’s largest immigration detention system.[33] On average, ICE detains over 250,000 people annually.[34] The second Trump administration has promised to more than double those numbers.[35] Asylum-seekers,[36] legal permanent residents,[37] and even U.S. citizens[38] are among those held in immigration detention. On average, their detention lasts 46.9 days.[39]

The United States has detained immigrants throughout its history. Immigration detention was essential in implementing racist American projects such as Chinese exclusion and Japanese internment.[40] In this way, detention facilitates the racial project of “sift[ing] . . . [and] separat[ing] the desirable from the undesirable immigrants.”[41]

The present mass scale of immigration detention in the United States is a more recent phenomenon.[42] Spurred by antiterrorism, anticommunism, and the “War on Drugs,” various statutory reforms in the late 80s and 90s increased reliance on detention within the U.S. immigration apparatus.[43] These laws imposed mandatory detention of entire categories of noncitizens, including asylum-seekers,[44] regardless of circumstances.[45] Laws like the Anti-Drug Abuse Act, Antiterrorism and Effective Death Penalty Act and the Illegal Immigration Reform and Immigrant Responsibility Act broadened a wide swath of crimes that trigger mandatory detention and deportation.[46] The Laken Riley Act is the latest expansion of this trend, mandating detention without bond for noncitizens merely suspected of committing crimes such as shoplifting.[47] These statutes not only facilitate the mass detention of immigrants but also the “convergence of criminal law and immigration law.”[48]

In many ways, immigration detention is nearly identical to criminal incarceration.[49] Many immigration detention centers are located in former jails and prisons or even in “shared-use” facilities where individuals awaiting immigration court are held in the same facilities as those convicted of criminal charges.[50] Additionally, local law enforcement collaborates with ICE to detain people released from the criminal justice system, even if charges are dropped, in what is known as the “arrest-to-deportation pipeline.”[51] Both detainees and prisoners are “subject to complete control in a confined setting within a system that has a disparate impact on communities of color.”[52]

Despite the intersections and similarities, immigration detention and criminal incarceration “exist within two separate justice systems.”[53] Unlawful presence in the United States is a civil violation adjudicated in administrative court proceedings.[54] According to the Supreme Court, deportation is not a form of criminal punishment.[55] Similarly, the Court held that immigration detention is “not imprisonment in a legal sense” because it is not punishment for a crime but rather serves to ensure presence for immigration hearings or deportation orders.[56] This distinction ensures those facing immigration detention and deportation do not receive the same constitutional safeguards promised to criminal defendants, such as attorney assistance or a speedy trial.[57]

Plaintiffs in the labor lawsuits described in this Note face retaliation particular to the immigration system, including solitary confinement, deportation, and even physical violence from immigration officials.[58] While some legal claims may create limited opportunities for immigration relief, many do not.[59] Some litigators have partnered with advocacy groups and immigration attorneys as co-counsel in these lawsuits to ensure that plaintiffs are also supported in their immigration proceedings.[60]

B. For-Profit Immigration Detention Industry

Private companies say that ICE is responsible for immigrant human rights, and ICE says that private companies are responsible. Since most people detained by ICE are held in for-profit cages, this accountability merry-go-round means that nobody takes responsibility for detention conditions.

—Unjust Enrichment Zine[61]

Detained workers are formally held in custody of the federal government. But immigration detention is dominated by private companies. Over 90 percent of people detained in ICE custody are held in facilities owned or operated by private corporations.[62]

Recent labor lawsuits target two publicly traded companies: GEO and CoreCivic. These corporations constitute a duopoly in immigration detention.[63] GEO brought in more than one billion dollars in revenue in 2022 from ICE contracts alone, over 43 percent of its total revenue.[64] In 2022, CoreCivic made over half a million dollars in revenue from ICE contracts, approximately 30 percent of its total revenue.[65] President Trump’s mass immigration enforcement agenda only promises greater profits for these companies.[66] After his election, GEO and CoreCivic’s stock prices increased by 105 percent and 50 percent respectively.[67]

Like all government contractors, private immigration detention operators promise to provide the same services for cheaper, but there is “little room to improve the efficiency of incarceration.”[68] Technology and innovation can only go so far to more efficiently guard, feed, and care for detainees.[69] Instead, private facilities must cut costs in order to generate profit.[70] Labor and related costs represented 66 percent of GEO’s operating costs in 2023.[71] Similarly, salaries and benefits made up 60 percent of CoreCivic’s expenses.[72] Paying detained workers one dollar per day to run their facilities saves GEO and CoreCivic anywhere from thirty to over seventy million dollars each per year.[73]

C. “Voluntary” Work Program

We just wanted to be respected as workers. Look, we’re actually doing something—we shouldn’t be working for you for one dollar a day. We’re already getting exploited as it is.

—Jose Ruben Hernandez Gomez[74]

In the Aurora Detention Facility in Colorado, Demetrio A. Valerga worked six days a week stripping and waxing kitchen floors from eleven at night to six in the morning.[75] While Mr. Valerga was paid one dollar per day for this work, he was also required to clean common areas for no payment under threat of being thrown into solitary confinement if he refused.[76] In the Stewart Detention Center in Georgia, Wilhen Hill Barrientos occasionally earned the somewhat exceptional[77] rate of eight dollars per day—when he worked twelve-hour double-shifts in the facility’s kitchen.[78] Meanwhile, the federal minimum wage guaranteed more than ten times Mr. Barrientos’s rate of $0.66 per hour.[79] Mr. Barrientos was twenty years old at the time, awaiting a decision on his asylum claim after fleeing gang violence in Guatemala.[80]

ICE and private detention operators claim these work arrangements are permissible in accordance with national detention policies. Private facilities that contract or subcontract with ICE are subject to the Performance-Based National Detention Standards (PBNDS).[81] Section 5.8 of the PBNDS defines the policy at the center of these labor disputes: the Voluntary Work Program (VWP).[82]

According to the PBNDS, the purpose and scope of the VWP is to “provid[e] detainees opportunities to work and earn money while confined.”[83] The stated expected outcomes include enhancing the facilities’ operations and services through “detainee productivity” as well as reducing the “negative impact of confinement” by decreasing “idleness.”[84] The policy prohibits a detained worker “employed full time” from working more than eight hours per day, forty hours per week.[85] Section K states that compensation, while determined in accordance with facility policy, is “at least $1.00 (USD) per day.”[86]

The PBNDS cites no legal authority to support the creation of the VWP.[87] But this dollar-per-day program implemented in private facilities across the country is based on a 1950 statute.[88] Codified in 8 U.S.C. § 1555(d), the bill authorizes the “payment of allowances (at such rate as may be specified from time to time in the appropriation Act involved) to aliens, while held in custody under the immigration laws, for work performed.”[89] Congress set that rate at $1 per day in 1950,[90] the equivalent of $13.82 in today’s dollars.[91] The dollar rate remained until 1979, when the program line item disappeared entirely from the appropriations budget.[92]

Despite the measly wages, the program provides detainees a way to afford otherwise inaccessible basic necessities. Immigration detention facilities serve moldy and expired food,[93] withhold basic hygiene items such as soap and toilet paper,[94] and charge detainees for phone calls to loved ones.[95] Detainees must purchase edible food and toiletries through the detention centers’ commissary store, often at exorbitant prices,[96] “or go without.”[97] Litigators, advocates, and academics call this a “deprivation scheme,” which pressures detainees into participating in work programs only to hand their meager wages back over to the facility to pay for their necessities.[98]

Under these conditions, approximately half of those detained by ICE for more than a few days accept one-dollar-per-day wages.[99] Some even sign onto waitlists or work for free on a trial basis until they secure a paid position.[100] Some detainees report working to break up the tedium and stress of detention.[101] But as Washington activist Maru Mora Villalpando identified at NWDC, this tedium is a natural outgrowth of immigration detention’s lengthy and uncertain timelines.[102] Within the criminal justice system, the Constitution guarantees defendants basic procedural protections, including timely resolution of their cases.[103] Within the civil immigration system, however, detainees are often stuck waiting indefinite periods of time, dependent on backlogs in immigration court and their own willingness to remain in detention to fight their case.[104] Poor conditions in immigration detention drive detainees to “voluntary” departures and deportations as well as “voluntary” work arrangements.[105]

D. The Government as Policymaker and Party

How do we make sure that every elected official and state agency recognizes they have a role to play in the detention center?

—Maru Mora Villalpando, activist at NWDC.[106]

Federal and state government appear on both sides of legal challenges to detained labor conditions. This equivocation is best exemplified by the “Brief for the United States as Amicus Curiae in Support of Neither Party” filed in one now-settled case in 2019.[107] As the title suggests, the federal government made points favoring both detained workers as well as the private detention center company CoreCivic, notwithstanding the avowed commitment of President Trump and then-Attorney General William Barr to aggressively crackdown on immigration.[108] By contrast, President Biden’s administration filed an amicus brief opposing the Washington Attorney General’s lawsuit against GEO at NWDC.[109]

Government agencies and officials are the natural defendants in lawsuits challenging exploitative work conditions in immigration detention.[110] But suing the government invites a myriad of well-known challenges, including sovereign immunity, both absolute and qualified.[111] Additionally, these lawsuits often require proving detained workers are government employees, which introduces laws and regulations surrounding public employment.[112]

It is unsurprising, then, that all recent lawsuits contemplated in this Note target the private detention center operators rather than federal agencies. The Department of Homeland Security (DHS) and ICE have refused to intervene in these lawsuits.[113] However, even when exclusively targeting private contractors, the federal government’s role is constantly looming over efforts to challenge working conditions in detention. Its shadow generally arises through a variety of defenses rooted in the Constitution’s Supremacy Clause,[114] including derivative sovereign immunity, preemption, and intergovernmental immunity.[115]

Even so, detained workers and advocates are leveraging state and federal agencies to challenge labor violations. Some state agencies have found that detained workers are, in fact, employees with enforceable rights.[116] The Washington Attorney General’s lawsuit against GEO serves as a model for how state government officials are well positioned to challenge these conditions.[117]

Despite these successes, anti-immigrant sentiment is rising, even in supposedly liberal states and localities.[118] This impacts state and federal political willpower to enact remedial legislation,[119] bring lawsuits,[120] and take other agency enforcement actions[121] on behalf of detained workers. In this way, private litigation is a powerful tool in addressing unjust labor conditions in immigration detention.

E. Affirmative Civil Litigation: One Tool in the Toolkit

Outside of Los Angeles, Raul Novoa worked four hours a day and up to seven days a week as a janitor.[122] Later, he worked as a barber up to ten hours a day, every day of the week.[123] Regardless of position, Mr. Novoa was paid one dollar per day.[124] When he complained, his supervisors entered his living quarters, threw his belongings and papers around the room, and forced him to relocate.[125] They threatened to confine him in isolation if he complained further, stopped working, or encouraged others to stop working.[126] Mr. Novoa became the lead plaintiff in a class-action lawsuit against his alleged employer: GEO, the private corporation that detained him at the Adelanto ICE Processing Center.[127] As demonstrated by the suit’s class certification, Mr. Novoa’s experience is not unique.[128]

Civil lawsuits are one avenue to challenge these labor schemes in private immigration detention. Labor lawsuits seek to punish and deter labor violations and may even drive these for-profit detention corporations out of the industry.[129] In a series of email communications obtained through Freedom of Information Act (FOIA) litigation in 2019, GEO officials asked ICE to cover the estimated ten to fifteen million dollars in legal expenses and tens of millions more in potential damages from these cases.[130] ICE declined.[131] These legal expenses and remedies cut directly into the profit margins of private immigration detention corporations.[132]

The courthouse is not always the most effective instrument to remedy injustice. Civil litigation in the United States is slow, costly, and inaccessible to many.[133] Fortunately, litigation is not the only method to address injustice.[134] Detained workers and their advocates have effectively harnessed various other methods to challenge labor practices in privatized immigration detention. People inside and outside detention centers have organized to pressure ICE and the private companies they contract with to change their practices.[135] Detained workers and advocates have found creative ways to pursue remedies through state and federal administrative agencies.[136] Both state and federal legislators have developed policy solutions to address the injustices of private immigration detention.[137] While these non-litigation methods have been successful in various ways, they also have limitations. Detained workers face severe repression and retaliation for their organizing.[138] Administrative remedies, such as penalties from workplace health and safety agencies, remain complicated procedures that still require attorney support and depend on precarious legislation and executive enforcement.[139] And legislation limiting private immigration detention at both the state and federal level has been stymied or overturned, in part thanks to extensive lobbying efforts by the private immigration detention companies.[140]

Litigation, therefore, remains a core effective strategy in upholding and elevating detained workers’ rights, but it is best understood as “just one of the many tools that we use” rather than “the solution,” as described by organizer Maru Mora Villalpando.[141] It is most effective in conjunction with organizing efforts, administrative advocacy, and legislative reforms. For example, litigation can expose information otherwise inaccessible to advocates. While the lawsuit Barrientos v. CoreCivic ultimately settled, the detained workers’ attorneys at the Southern Poverty Law Center obtained many documents throughout discovery and successfully moved to make them publicly available.[142]

1. Summary Chart of Lawsuits Claiming Labor Violations in Private Immigration Detention

2. Common Litigation Obstacles

Most defenses discussed in this Note are particular to the type of claim raised. For example, the civic duty exception generally will only arise if Thirteenth Amendment claims are brought.[143] Supremacy Clause defenses are only invited by raising state law claims.[144] Accordingly, these defenses are analyzed in more detail under the sections below covering those particular claims. But immigration detention companies raise certain defenses regardless of the type of claim. This Section outlines two challenges facing any lawsuit alleging labor violations on behalf of detained workers: the Immigration Reform and Control Act and derivative sovereign immunity.

a. Immigration Reform and Control Act (IRCA)

The Immigration Reform and Control Act (IRCA) is a federal statute that prohibits knowingly employing noncitizens without work authorization.[145] Both GEO and CoreCivic argue that IRCA protects them from liability to detainees for employment-related violations.[146] They claim “it would be absurd” to recognize detained workers as employees when IRCA prevents hiring undocumented people.[147]

This argument is based on the assumption that all detained workers are undocumented. But ICE detains legal permanent residents as well as U.S. citizens.[148] Legal permanent residents, also known as green card holders, are authorized to work and their status remains until they receive a final order of removal in their immigration proceedings.[149] Professor Jacqueline Stevens argues that all detainees, except those waiting for execution of their final deportation orders, are waiting for a determination of their legal status.[150] Accordingly, their work status is inherently unknown, defeating IRCA’s mens rea requirement.[151] The Western District of Washington seemingly acknowledged this when it declined to decide whether IRCA conflicted with Washington State’s definition of employee “because [IRCA] relies on factual determinations about the status of detainees.”[152]

Even if hiring individuals unauthorized to work were an IRCA violation, this should only make companies like GEO and CoreCivic liable for both IRCA and relevant labor violations.[153] Any other outcome would insulate violators from liability, creating perverse incentives for employers to hire and exploit undocumented workers.[154]

IRCA remains a lingering obstacle to successfully bringing claims against labor conditions in detention. Detained workers narrowly overcame this claim in a motion to dismiss in one lawsuit[155] but failed to defeat the affirmative defense on a motion for summary judgment in another case, allowing it to proceed to trial for a jury’s determination.[156] Additionally, where detained workers raise state law claims, defendants argue IRCA preempts state law, a defense discussed further below.[157]

b. “Derivative Sovereign Immunity”

Under the long-standing principle of sovereign immunity, the United States cannot be sued without explicit authorization from Congress.[158] The Supreme Court has found some of that protection extends to federal contractors “authorized and directed” by the government.[159] Contractors commonly refer to this principle as “derivative sovereign immunity.”[160] But the Supreme Court recently ruled that this principle is merely a merits defense, not an immunity.[161]

GEO has repeatedly claimed “derivative sovereign immunity” protects it from liability.[162] Detained workers have successfully defeated this argument at various procedural postures at the trial and appellate court level.[163] The workers argue that the Voluntary Work Program as “directed” by the government does not actually violate any of the claims outlined below. Instead, private operators like GEO and CoreCivic violate these laws through their discretionary execution of the program. As the U.S. government wrote in one amicus brief, ICE only directed the contractors to provide voluntary opportunities to work, not to coerce labor.[164] Additionally, the government only directs these private companies to pay at least one dollar per day through the PBNDS.[165] Accordingly, the government “set a payment floor, not ceiling” and did not direct the private contractors to violate, for example, minimum wage laws.[166]

The principle of derivative sovereign immunity is sometimes called “government contractor immunity”—but this is arguably a distinct defense emerging out of Boyle v. United Technologies Corporation.[167] The government contractor test under Boyle overlaps significantly with preemption,[168] as it requires showing a “significant conflict” between an identifiable “federal policy or interest and the operation of state law.”[169] Detained workers have survived both a motion to dismiss and motion for summary judgment under Boyle.[170]

II. Federal Causes of Action—Literature Review and Case Law Developments

A federal cause of action appears the most obvious path to address violations in federal immigration custody. This Part explores three federal causes of action—the Thirteenth Amendment, the Trafficking Victims Protection Act (TVPA), and the Fair Labor Standards Act (FLSA)—that detained workers have brought in court with varying degrees of success. Existing legal literature has explored each of these federal paths.[171] This Part seeks to summarize their conclusions, incorporate the latest developments from ongoing lawsuits, and highlight paths forward for litigators, advocates, and detained workers.

A. The Thirteenth Amendment

Considering immigration detention is almost entirely privatized, detained workers face an uphill battle raising constitutional claims against these non-state actors.[172] While private state facilities may be liable for constitutional violations through a statutory cause of action,[173] private federal facilities may not.[174] But the Thirteenth Amendment offers a path to overcome this obstacle.

The Thirteenth Amendment contains two sections. The first prohibits both “slavery” and “involuntary servitude” in the United States.[175] The second grants Congress the authority to enforce this provision through legislation.[176] Enacted after the Civil War, the Amendment constitutionally grounded abolition after the Emancipation Proclamation.[177] But the earliest judicial interpretations affirm it “comprises much more than the abolition or prohibition of African slavery.”[178] Instead, the Amendment applies to slavery and involuntary servitude “in all forms, and in all degrees.”[179] Like most constitutional guarantees, noncitizens are also protected by the Thirteenth Amendment.[180] Additionally, Section One of the Thirteenth Amendment is “undoubtedly self-executing without any ancillary legislation,” allowing for a direct private right of action.[181] The Thirteenth Amendment is also the only constitutional protection that restrains private actors, such as GEO and CoreCivic.[182]

Academics and experts cite the Thirteenth Amendment as directly applicable for detained workers facing coercive labor conditions.[183] And yet, plaintiffs have yet to prevail on this claim.[184] Detained workers face various obstacles to make successful Thirteenth Amendment claims. But these challenges are not entirely insurmountable.

First, the Thirteenth Amendment notoriously includes a loophole. The first section states that slavery and involuntary servitude are prohibited “except as a punishment for crime whereof the party shall have been duly convicted.”[185] But courts have already found that this Exceptions Clause is entirely inapplicable to detained workers.[186] Immigration detention is not a punishment for a criminal conviction. Detention centers are instead “merely waiting rooms for immigrants seeking to have their day in court.”[187] Under long-standing Supreme Court precedent, both immigration detention and deportation are civil in nature, not criminal.[188] In Wong Wing v. United States, the Court found the government could not impose hard labor on an immigrant for unauthorized presence because that constituted a criminal punishment.[189] In fact, because they are theoretically not facing punishment, immigration respondents are deprived of various safeguards afforded criminal defendants, such as the right to an attorney and a jury trial.[190]

While detained workers defeat the criminal punishment exception, they must still meet the high and often convoluted bar established by Thirteenth Amendment case law. As Lauren Kares writes in her article “Unlucky Thirteenth,” “[t]he Thirteenth Amendment is notable for its lack of coherent jurisprudence.”[191] One of the few consistent themes in the doctrine is the overall restraint of courts in defining “slavery” and “involuntary servitude.”[192] This judicial deference “creates a considerable barrier to plaintiffs seeking to vindicate Thirteenth Amendment-based rights.”[193] Despite this, the Fifth Circuit suggested in dicta that a detained immigrant coerced into working for minor wages may meet the high and amorphous judge-made standard for a Thirteenth Amendment violation.[194]

Nonetheless, the Fifth Circuit ultimately found that the detained worker’s Thirteenth Amendment claim failed due to the “civic duty exception.”[195] In the scattered evolution of Thirteenth Amendment jurisprudence, courts have discerned various exceptions beyond the textual Exceptions Clause.[196] The civic duty exception traces its roots back to an early twentieth-century case finding the Thirteenth Amendment was “certainly . . . not intended to interdict enforcement of those duties which individuals owe to the State, such as services in the army, militia, on the jury, etc.”[197] Some courts have expanded the civic duty exception to include housekeeping tasks while in civil detention.[198]

Both GEO and CoreCivic raised this civic duty exception as an affirmative defense against statutory claims by detained workers.[199] But the Supreme Court has only applied the civic duty exception to government actors.[200] Two district courts held that the civic duty exception does not apply to private, for-profit corporations like GEO and CoreCivic.[201]

While the Thirteenth Amendment may appear to be a close match for detained workers, judges remain reticent to find implied rights of action[202] and the varied, narrowly construed judge-made jurisprudence is risky.[203] But two federal statutes are well suited to tackle coercive working conditions and subminimum wages: the Trafficking Victims Protection Act and the Fair Labor Standards Act.

B. Trafficking Victims Protection Act (TVPA)

Congress passed the TVPA in 2000.[204] The statute alludes both to Congress’s authority to regulate interstate commerce as well as its power to enforce the Thirteenth Amendment.[205] The omnibus bill includes a wide array of provisions to punish traffickers, provide assistance to victims, and implement mechanisms to prevent trafficking. The bill also contains a provision to protect victims from forced labor.[206]

The TVPA directly resolves many of the challenges arising out of Section One Thirteenth Amendment claims.[207] First, plaintiffs do not face the major procedural hurdle of convincing a judge to find an implied cause of action directly under the Thirteenth Amendment.[208] Second, the TVPA relies on the broader and more expansive Section Two of the Thirteenth Amendment.[209] While Section One prohibits the “narrow but nebulous territory of slavery or involuntary servitude,”[210] Section Two grants broad authority to Congress to prohibit the “badges and incidents” of slavery.[211]

Many are surprised to learn that arguably the most effective cause of action so far in challenging work conditions for detainees is within a human trafficking law. CoreCivic and GEO leverage this reaction in their purposivist defenses against these lawsuits. Both argued that the TVPA only applies to human trafficking and therefore is inapplicable to immigration detention.[212] In some cases, they argued it would be “absurd” to apply the TVPA, as plaintiffs do not allege human trafficking.[213] But courts have consistently rejected these arguments and employed traditional textualist statutory interpretation.[214] The Eleventh Circuit concluded that “the plain language of the TVPA brings within its scope for-profit government contractors operating work programs in federal immigration detention facilities, and such entities are not categorically excluded or shielded from liability under the TVPA.”[215] Following this same textualist approach, the notoriously conservative Fifth Circuit found that “on its face [the TVPA] unambiguously protects labor performed in work programs in federal immigration detention facilities . . . .”[216]

At least five active lawsuits alleging TVPA violations in immigration detention remain ongoing.[217] Most recently, Colorado detained workers' TVPA claims survived a motion for summary judgment and overcame an attempted interlocutory appeal at the Supreme Court.[218]

C. Fair Labor Standards Act (FLSA)

The FLSA establishes nationwide standards for employees, including a minimum wage.[219] The FLSA appears a fitting remedy for detained workers, whose one-dollar-per-day wages fall well below the federal minimum.[220] And yet, two circuit courts spanning four decades concluded the FLSA does not apply to workers held in immigration detention.[221] Scholars and litigators have identified opportunities to challenge the reasoning underlying these decisions.[222]

First, neither the Fourth Circuit in Alvarado Guevara v. I.N.S. nor the Fifth Circuit in Ndambi v. CoreCivic applied the prescribed legal test. Under Supreme Court precedent, courts should look to the “economic reality” of a labor relationship to determine if a worker qualifies as an employee protected by the FLSA.[223] When applied, this test tends to show that detained workers are, in fact, employees.[224] But both the Fourth and Fifth Circuits instead held that the FLSA was never meant to apply to workers in custody.[225]

Second, these decisions overly relied on prison labor cases, treating them as analogous to the civil immigration detention context. In Alvarado Guevara, the Fifth Circuit relied exclusively on prison labor case law to conclude that the FLSA did not apply to workers held in immigration detention.[226] In Ndambi,the Fourth Circuit relied, at least in part, on the Fifth Circuit’s reasoning in Alvarado Guevara.[227] But as discussed above, the analogy between incarcerated workers and workers held in immigration detention is inherently flawed.[228] Case law denying FLSA protections for prisoners often hinges directly or indirectly on the Exceptions Clause of the Thirteenth Amendment, which does not apply to immigration detainees who are not serving time for criminal convictions.[229]

Lastly, the Ndambi and Alvarado Guevara decisions both relied on their interpretations of the purpose of the FLSA, concluding that FLSA protections were not intended for detained workers because their well-being is already provided for by the facility and they work outside of the traditional labor market.[230] This departure from the text contravenes traditional principles of statutory interpretation.[231] The FLSA expressly enumerates specific and limited exceptions to the otherwise broad definition of “employee.”[232] There is no exemption for those in immigration detention and various cases acknowledge that no categorical exemption exists for prison laborers either.[233]

The Ndambi court cited precedent reasoning that no “express exception” exists in the FLSA for prisoners because it was “too outlandish” for Congress to consider at the time.[234] This conclusion is based on the logic that the FLSA’s goal of ensuring a “minimum standard of living” and “general well-being of workers” does not apply to custodial labor.[235] According to this reasoning, the FLSA was not intended to protect detained workers because their basic needs are already provided for by the facility.[236] But the Supreme Court already rejected the premise that FLSA coverage ceases when a worker’s food, shelter, and clothing are provided.[237] While workers may receive room and board as part of wages in kind,[238] acceptance of the arrangement must be “voluntary and uncoerced.”[239] Because detainees are required to live and eat in the detention center, their cost of detention cannot conceivably be considered part of their pay.[240] Additionally, the Ndambi court accepted as sufficient that CoreCivic is “morally and legally bound to meet detainees’ basic needs,” despite acknowledging plaintiffs’ claims that CoreCivic is failing to provide those necessities.[241] This disputed fact, not resolvable at the motion to dismiss stage, undermines the premise that detained workers do not need FLSA protections to ensure a “minimum standard of living.”[242]

Both the Ndambi and Alvarado Guevara decisions reasoned that custodial labor falls outside the “free labor market” and “American industry” contemplated by the FLSA.[243] But slight variations in detained worker wages according to the demand for labor suggest a “rudimentary labor market.”[244] Additionally, this custodial labor serves private, for-profit corporations operating in the market. By slashing labor costs, these companies undercut competitors for similar federal contracts.[245] This dynamic pressures other private detention companies to engage in similar labor violations to remain competitive, encouraging a “race to the bottom” that the FLSA sought to prevent.[246]

Despite embracing purposivist arguments, the Ndambi and Alvarado Guevara courts did not reckon with purpose-driven precedent that encourages an expansive reading of FLSA coverage.[247] The Fourth Circuit reasoned that detained workers do not fit the “traditional employment paradigm” the FLSA was enacted to protect.[248] But the Supreme Court held that the FLSA’s expansive coverage “stretches the meaning of ‘employee’” to cover parties not traditionally contemplated.[249] Nonetheless, the Ndambi and Alvarado Guevara precedent will likely limit this capacious application to detained workers.

III. State Causes of Action—Emerging Opportunities and Challenges

The only lawsuits alleging workers’ rights violations in immigration detention that have proceeded to trial (and won) are the parallel public-private lawsuits at NWDC. In Nwauzor and Washington v. GEO, the plaintiffs brought no constitutional or federal statutory claims. Instead, both lawsuits exclusively contemplated state law claims—specifically under Washington’s minimum wage statute and common law unjust enrichment.[250] These were not the first lawsuits to allege that private detention centers violated state laws, and others have followed suit.[251] These lawsuits highlight the challenges but also the power of leveraging state laws to address the violations of workers’ rights in immigration detention.

State causes of action may not appear ideal to address the national problem of workers’ rights violations in federal immigration detention. State law claims are, by definition, geographically bound. Moreover, empowering states to address immigration-related matters may backfire. States across the country are leveraging their local laws to increase detentions, deportations, and human rights abuses against immigrants.[252] State lawsuits may simply drive ICE and its contractors out of states enforcing their labor laws and into these more “hospitable jurisdictions.”[253] This could result in detainees being transferred further from their families and into jurisdictions with fewer protections.

However, these valid concerns should not necessarily stop advocates from considering state law claims. As organizer Maru Mora Villalpando counters, “ICE transfers people all the time, every single day. [This is] not just about being able to visit loved ones in horrible conditions.”[254] And state law offers a “reservoir of possibility to hold contractors accountable in ways that the federal government does not.”[255] Profit-driven detention corporations like GEO and CoreCivic rely on their near-free, captive workforce,[256] and requiring companies to comply with labor laws could disrupt the massive private immigration-detention industry. With more minimum wage and unjust enrichment class actions, detained workers may have their day in court, receive compensation, and financially impact institutions profiting from their detention.

A. Minimum Wage Statutes

Like all state laws, minimum wage statutes vary across the country. Some states have no minimum wage provisions at all.[257] While thirty-four states have a minimum wage higher than the federal $7.25 rate, many state statutory schemes closely emulate the FLSA.[258] In these jurisdictions, courts turn to FLSA precedent, including the economic realities test, to determine who is protected under state minimum wage laws. For example, the Ndambi plaintiffs raised minimum wage claims under both the FLSA and the New Mexico Minimum Wage Act (NMMWA).[259] But both parties stipulated that the NMMWA was interpreted in accordance with the FLSA.[260] So when the court threw out the plaintiffs’ FLSA claims, their NMMWA claims went with them.[261]

Some states apply their minimum wage statutes more broadly than the FLSA, potentially providing a more effective remedy for detained workers in these jurisdictions. For example, the California Minimum Wage Law (CMWL)[262] is expressly separate from and more expansive than the FLSA.[263] As a result, two federal courts in California rejected defendants’ attempts to import unfavorable FLSA case law and ruled in favor of detained workers on their CMWL claims.[264]

California is one of a handful of states embracing a more expansive alternative to the FLSA’s economic realities test.[265] Known as the ABC test, it presumes a worker is a covered employee unless the employer can prove three distinct elements.[266] The test has been lauded as worker friendly.[267] But the test is generally applied to distinguish employees from independent contractors. The Novoa court found there was no question of whether detained workers were independent contractors, and as such held the ABC test was inapplicable in determining whether detained workers are employees under the CMWL.[268] While the proliferation of the ABC test offers an opportunity to protect more workers, it may not reach detained workers.

In fact, Washington state is one of many states that has not adopted the ABC test,[269] and yet a jury and the state’s Supreme Court found detained workers are covered under state minimum wage law.[270] Like the California cases, the claims under the Washington Minimum Wage Act (MWA) survived in part because they were distinguished from the FLSA.[271] The MWA’s unique structure also supported the conclusion that workers held in immigration detention are covered employees. The MWA defines who is protected under the act through a series of narrowly construed statutory carve-outs.[272] As the Washington Supreme Court held in Nwauzor, “[b]y its very definition, ‘employee’ is defined by its exemptions.”[273] The court found workers detained in private immigration detention did not fall under any of the applicable exemptions and were therefore covered employees protected by the MWA.[274]

Unique statutory schemes defined by circumscribed categories have not always served detained workers bringing state minimum wage claims. In 2015, the court in Menocal v. GEO Group dismissed plaintiffs’ claims under the Colorado Minimum Wage Order (CMWO).[275] At the time, the CMWO only applied to employers and employees in four specific industries.[276] While the Nwauzor plaintiffs only needed to prove the Washington law carve-outs were inapplicable, the Menocal plaintiffs had to argue the industry of immigration detention fit into one of the statutory categories. The Menocal court also embraced the reasoning in the FLSA case Alvarado Guevara to hold that the plaintiffs, like prisoners, could not be employees.[277]

Until the Ndambi and Alvarado Guevara precedent is successfully challenged, detainees making one dollar per day will struggle to vindicate their rights under local minimum wage laws that follow the FLSA. But state minimum wage statutes that do not strictly follow the FLSA, like in California and Washington, may offer an effective remedy for workers in immigration detention.

B. Unjust Enrichment

Unjust enrichment is a “broad and relatively under-theorized” legal doctrine.[278] The concept reaches back to the Roman Empire and even earlier to ancient China.[279] Rooted at the intersection of common law and equity, the doctrine has been significantly shaped by academics.[280] Unjust enrichment can be understood as the transfer of value “without adequate legal basis.”[281] A defendant is generally unjustly enriched if they (1) receive a benefit (2) at the plaintiff’s expense and (3) where retaining the benefit without compensation would be unjust.[282]

Almost all major lawsuits challenging work conditions in private immigration detention raise state unjust enrichment claims.[283] State jurisprudence varies but the common law action exists in some form across all fifty states.[284] Unlike minimum wage claims, detainees do not need to prove they are “employees” to show defendants unjustly benefited from their near-free labor. While damages for statutory wage claims are capped at the minimum wage, “restitution for unjust enrichment is based on the fair market value of the labor performed.”[285]

The broad and flexible claim appears to map neatly onto the reality of detained workers. Private detention centers receive a benefit at the expense of detainees. Conservative estimates suggest that from 2018 to 2022, detainee labor saved GEO and CoreCivic $265 million in labor costs.[286] Defendants argue that detainees cannot raise unjust enrichment claims because they did not have a “reasonable expectation” of market value payment for their labor. But courts in Georgia, Washington, and Colorado dismissed this argument, finding that a reasonable expectation of payment was not necessary to claim unjust enrichment.[287] Alternatively, defendants claim they do not benefit from detainee labor but instead lose out on profits. GEO and CoreCivic contracts structure their payment as a certain profit margin percentage of their costs.[288] They argue that replacing detainee labor with fully compensated employees would actually increase their profit margins by increasing their total expenses.[289] For example, if their contracts grant them a 30 percent profit margin over all their expenditures, increased labor costs would actually increase their overall profits. But courts have rejected this argument, noting that the companies’ low costs, thanks to detained labor, grant them a competitive advantage in securing those ICE contracts in the first place.[290]

The core issue is the third element: whether it would be “unjust” for the defendants to retain this benefit. This element often hinges on other alleged violations, either minimum wage or forced labor as discussed above. For example, in Barrientos, the district and circuit courts agreed that plaintiffs’ unjust enrichment claim depended on a finding that detainees were coerced into working, in alignment with the raised TVPA claims.[291] Meanwhile, the district court in Menocal held that the unjust enrichment claim was “largely based” on the plaintiffs’ minimum wage allegations, and not the TVPA claims also raised.[292]

But the derivative nature of the claim can lead to a catch-22 readily exploited by defendants. Defendants argue unjust enrichment claims must be dismissed if the underlying claim fails. But where the underlying claim survives, defendants then argue unjust enrichment claims must be dismissed as duplicative.

When unjust enrichment hinges on another meritorious claim, it may ride the coattails of that underlying claim.[293] But defendants can also successfully dismiss an entire case by poking holes in the underlying claim, as in Ndambi. There, the district court first dismissed plaintiffs’ state and federal minimum wage claims.[294] The court then concluded that the plaintiffs’ unjust enrichment claim was “entirely dependent” on these minimum wage violations and therefore must be dismissed as well.[295] On review, the Fourth Circuit threw out plaintiffs’ unjust enrichment claims in one short footnote, as the plaintiffs conceded their unjust enrichment claim failed if they lost on their minimum wage claims.[296] But Menocal offers an alternative resolution. There, the district court permitted plaintiffs to proceed with their unjust enrichment claims, despite acknowledging they were “largely based” on dismissed minimum wage claims.[297]

Where the underlying claim survives, defendants argue unjust enrichment claims should be dismissed as duplicative. Courts often recognize unjust enrichment as an equitable remedy that is unavailable when an alternative adequate remedy exists.[298] For example, unjust enrichment is generally unavailable if an enforceable contract governs the dispute, and therefore contractual remedies are available.[299] Generally, plaintiffs can still proceed with unjust enrichment claims as pleadings in the alternative.[300] But where unjust enrichment claims overlap with other claims, plaintiffs face allegations of “recover[ing] twice for the same injury.”[301]

Despite the overlap, it is still possible to recover under both unjust enrichment and underlying violations. The Washington Attorney General sued GEO for injunctive relief under the state minimum wage statute and also disgorgement for unjust enrichment.[302] In parallel, a private class action sought damages for minimum wage violations.[303] In 2023, after both suits won on all claims at trial, the Washington Supreme Court held that the damages award to the private class did not foreclose equitable relief to the state.[304] Additionally, lawsuits raising both unjust enrichment and TVPA claims have survived motions to dismiss[305] and even motions for summary judgment.[306]

Interestingly, both GEO and CoreCivic have raised an unjust enrichment counterclaim, often overlapping with an offset affirmative defense.[307] The companies argued detained workers received the benefit of food, lodging, and other necessities and that payment of damages would unfairly burden the companies.[308] Therefore, defendants argued that if the plaintiffs prevail, the companies should be able to recover all costs associated with detaining the plaintiffs.[309] Plaintiffs moved to dismiss this claim, pointing out that ICE already fully reimbursed the companies for these expenses.[310] While the Washington district court dismissed the unjust enrichment counterclaim against Washington state,[311] the court rejected plaintiffs’ pretrial motions in the private class action to dismiss the counterclaim, which went all the way to trial.[312] By contrast, a Georgia district court quickly rejected this counterclaim, concluding: “Quite frankly, CoreCivic’s audacity in claiming that it should be compensated for providing basic necessities to detainees who it contracted to detain is astonishing.”[313]

Unjust enrichment is a promising, although under-analyzed, cause of action for detained workers. But its success often hinges on the success of the underlying claim for forced labor or minimum wage violations and then convincing the court to permit recovery under both theories.

C. Supremacy Clause Defenses: The Main Obstacle to State Causes of Action

The Nwauzor and Washington lawsuits won a combined verdict of $23.2 million against GEO for violating Washington state’s minimum wage statute and unjustly enriching itself at the NWDC.[314] Despite this major victory, the decisions remain disputed on appeal.[315] The Biden administration’s DOJ filed an amicus brief in support of GEO.[316] The DOJ argued that the Constitution’s Supremacy Clause prevented the application of Washington state law to detained workers in private immigration detention under the doctrines of preemption and intergovernmental immunity.[317] These Supremacy Clause defenses are the main obstacle to detained workers’ state law actions.

The Supremacy Clause of the U.S. Constitution states that federal law “shall be the supreme Law of the Land.”[318] State law may be unconstitutional under the Supremacy Clause in two ways.[319] First, contrary state law may be preempted by federal law.[320] Second, state law cannot directly regulate or discriminate against the federal government.[321] Both preemption and intergovernmental immunity limit states’ ability to regulate the federal government.

Private federal government contractors are increasingly appealing to these two federalism doctrines to insulate themselves from state law liability.[322] But as Professor David Rubenstein discusses in his piece Supremacy, Inc., these doctrines do not fit neatly onto increasingly outsourced private contractors.[323] Private contractors, unlike the federal government, are already unimpeded by constitutional and statutory public law.[324] By expanding these Supremacy Clause doctrines to private contractors, the federal government effectively shields itself and avoids liability.[325] As detailed below, detained workers have successfully pushed back on attempts to expand these doctrines to private contractors like GEO and CoreCivic.

1. Preemption

Under the Supremacy Clause, federal law may preempt “contrary” state law.[326] State law may be contrary to federal law in three different circumstances. First, where the federal government expressly states its intent to preempt state law.[327] Second, where the federal government establishes a comprehensive regulatory scheme that “occup[ies] the field” in that area.[328] And third, where federal and state law conflict such that it is a “physical impossibility”[329] to comply with both or state law “stands as an obstacle” to congressional objectives.[330] Both GEO and CoreCivic have raised each of these preemption arguments in response to detained worker state law actions.

Preemption analysis begins with the “basic assumption that Congress did not intend to displace state law.”[331] And where the state acts within its historic police powers, courts presume no preemption.[332] Courts have consistently found that labor law standards, such as minimum wage statutes, “fall[] within the traditional police power of the State.”[333] The Ninth Circuit held that “the power to regulate the employment of unauthorized aliens remains within the states’ historic police powers.”[334]

But “at first blush,” preemption seems to favor immigration detention contractors “given the long history of federal legislation and agency action in the area of immigration detention generally.”[335] The federal government maintains broad and extensive authority over immigration.[336] Accordingly, the U.S. Supreme Court has regularly struck down state laws seeking to regulate in the field of immigration.[337] Private immigration detention companies appeal to this principle in defending against state labor claims. They point to extensive federal regulation in the areas of immigration detention,[338] registration,[339] and employment.[340]

These arguments appear persuasive “at a high level of generality.”[341] But “[a]fter peeling back the rhetoric,” courts have consistently rejected these broad, sweeping preemption arguments.[342] Instead, courts have narrowed the “pertinent regulatory field” to immigration detainee wages.[343] In doing so, courts have rejected appeals to “a hodgepodge of federal statutes”[344] that regulate immigration generally, and instead have “examin[ed] the actual statutes and regulations, on the issue of detainee wages.”[345] Courts generally look to two particular statutes as most relevant for preemption analysis in this area: the IRCA and the statute covering immigration expenses known as section 1555(d).

IRCA imposes civil and criminal penalties on employers who knowingly hire immigrants without legal work authorizations.[346] One statutory reading understands these penalties as demonstrating congressional intent to keep noncitizens out of the U.S. workforce.[347] But an alternate statutory reading indicates that IRCA is actually meant to protect undocumented workers from discrimination, abuse, and exploitation.[348] Congress sought to “remove the economic incentive for employers to exploit and use” undocumented workers.[349] The question of the correct underlying purpose of IRCA is relevant here because the “ultimate touchstone in every preemption case” is congressional intent behind the regulation at issue.[350] Following the latter reading, enforcement of state workplace protections does not conflict with but rather furthers Congress’s objectives in passing IRCA.

Private immigration detention companies argue IRCA expressly preempts state law, occupies the field of detainee wages, and conflicts with state labor laws. IRCA expressly states that it “preempt[s] any State or local law imposing civil or criminal sanctions . . . upon those who employ” unauthorized workers.[351] GEO and CoreCivic argued that damages and restitution sought under state law claims for labor violations constitute “sanctions” for hiring unauthorized workers.[352] Courts have consistently rejected this argument.[353] Remedies such as backpay for violating labor laws do not appear to constitute “sanctions.”[354] Even if they did, none of the alleged state law claims seek damages or restitution for hiring unauthorized workers but rather for labor violations applicable regardless of work status.[355] Courts similarly rejected defendants’ arguments that IRCA “preempts the entire field” of immigration employment[356] or immigration detainee wages.[357]

Courts have more seriously considered whether state minimum wage laws conflict with IRCA. In Novoa, the Central District of California cited California precedent finding state wage laws do not conflict with or create an obstacle to IRCA.[358] The Novoa court recognized the “inherent tension” in allowing undocumented workers to raise claims arising from illegal employment.[359] But the court ultimately concluded that the alternative would only incentivize employers to violate both federal and state laws.[360] By contrast, the Western District of Washington in Washington v. GEO and Nwauzor v. GEO seemed to implicitly acknowledge a potential conflict between IRCA and Washington’s minimum wage law.[361] GEO argued it would be impossible to comply with IRCA, which prohibits hiring of unauthorized immigrants, and the state minimum wage provision, which would require GEO to treat detained workers as employees.[362] The court denied GEO’s motion to dismiss because “it relie[d] on factual determinations about the status of detainees” but acknowledged the issue “may become ripe at summary judgment or at trial.”[363] Notably, GEO did not raise any IRCA preemption arguments on appeal at the Ninth Circuit.[364] The DOJ’s amicus brief similarly made no IRCA preemption argument.[365] Instead, both parties argued that section 1555(d) preempted state law claims.

Section 1555(d) authorizes the “payment of allowances (at such rate as may be specified from time to time in the appropriation Act involved) to aliens, while held in custody under the immigration laws, for work performed.”[366] Congress set the allowance rate at “not in excess of” one dollar per day up until 1979, when the INS removed the program entirely from the appropriations budget.[367] Section 1555(d) originally appeared in a 1950 bill authorizing various recurring administrative expenses for the DOJ, including the federal immigration agency housed under it at the time.[368] Professor Anita Sinha identifies dual purposes underlying this authorization for the payment of detained workers: security and humanitarian concerns.[369] First, McCarthyism was increasingly reliant on immigration enforcement to detain and remove suspected communists, and nominally paid labor allegedly maintained order in detention.[370] Second, the provision was modeled after the newly ratified Geneva Convention Relative to the Treatment of Prisoners of War.[371] Compliance with state minimum wage and unjust enrichment laws does not seem to undermine either identified purpose.

Both GEO and CoreCivic have argued section 1555(d) occupies the field of detainee wages and that state laws conflict with or undermine the statute’s objectives. Courts dismissed GEO and CoreCivic’s assertions that this “one subsection” creates a “‘pervasive’ or ‘exhaustive’ set of regulations on detainee wages” sufficient to occupy the field.[372] Both the Central District of California and the Western District of Washington held that Congress “abandon[ed]” appropriations for this purpose after 1979.[373] This disregard “refutes any reasonable inference that Congress left no room for states to supplement the field.”[374]

In Owino, the Southern District of California found this same fact prevented a finding of conflict preemption.[375] The Owino court found there was a conflict between California’s minimum wage law, which required payment well over $1 per day, and the 1978 Appropriations Act, which “require[ed] no more than $1/day.”[376] But the Owino court ultimately found the appropriations bill did not “still control[] to this day.”[377] The court acknowledged this was a “close question” and did not “foreclose the possibility” that Congress did in fact intend to bind future parties.[378] Nonetheless, Owino found that Congress failed to demonstrate “clear and manifest purpose” and therefore rejected CoreCivic’s motion to dismiss under this theory of conflict preemption.[379]

The Biden administration filed an amicus brief on behalf of GEO in their Ninth Circuit appeal of Nwauzor and Washington.[380] The DOJ conceded that section 1555(d) does not conflict with Washington’s minimum wage: The one-dollar-per-day rate operates as a floor, which does not prevent contractors from paying greater amounts.[381] Instead, the federal government argued obstacle preemption, claiming state laws “will undermine the Program Congress has established.”[382] The government argued that the application of state minimum wage laws would “imperil the Program’s ongoing viability,” pointing to the increased unreimbursed costs and the suspension of the work program at NWDC.[383] The Ninth Circuit rejected this argument, finding section 1555(d) did not obstacle preempt Washington’s minimum wage law.[384] Much like the district court cases analyzing field preemption, the Ninth Circuit found the appropriations act setting the one dollar rate was “no longer in force.”[385] But the court went on, finding that even if it were in force, it still would not preempt state minimum wage law.[386] The court identified that the appropriations act only established a cap for reimbursement by the government—not for payment by a contractor.[387] More powerfully, the court found the appropriations act only contemplated detainees held in government-run facilities, not private, for-profit contractors.[388] GEO is appealing the decision to the Supreme Court.[389]

Private immigration detention corporations consistently invoke preemption to shield themselves from state law claims. These companies largely cite to IRCA and section 1555(d) but also to the FLSA,[390] the Immigration and Nationality Act (INA),[391] and even their own contracts with the government.[392] Despite the persistence of these detention companies, detained workers have so far largely prevailed against all claims of express, field, and obstacle preemption.

2. Intergovernmental Immunity

In the seminal case McCulloch v. Maryland, Chief Justice Marshall interpreted the Supremacy Clause as preventing states from “retard[ing], imped[ing], burden[ing], or in any manner control[ing]” the federal government.[393] Considering McCulloch v. Maryland ultimately found a state tax on a federal bank unconstitutional, the concept is often called “intergovernmental tax immunity.”[394] But courts have extended the doctrine beyond taxation.[395] Intergovernmental immunity shields the federal government from both direct regulation and discrimination by the states.[396]

To what extent does intergovernmental immunity apply to private companies like GEO and CoreCivic?[397] A long line of precedent applies intergovernmental immunity not just to the federal government but also to direct regulation of its “instrumentalities”[398] and discrimination against “those with whom it deals.”[399] An “instrumentality” must be “so closely connected to the Government that the two cannot realistically be viewed as separate entities.”[400] Most federal contractors do not reach this high bar.[401] But the Supreme Court has also concluded, in dicta, that “a federally owned facility performing a federal function is shielded from direct state regulation, even though the federal function is carried out by a private contractor.”[402] Drawing on this precedent, the Ninth Circuit concluded in a footnote that “[f]or purposes of intergovernmental immunity, federal contractors are treated the same as the federal government itself.”[403]

And yet, in reviewing detained workers’ Washington state law claims against GEO, a split Ninth Circuit majority recently held that “[f]or purposes of intergovernmental immunity, federal contractors are not equivalent to the federal government.”[404] The court detailed the many “significant differences” between government entities and private contractors.[405] Citing precedent from the early nineteenth century, the majority concluded that “states may impose some regulations on [private] federal contractors that they would not be able to impose on the federal government itself.”[406]

Detained workers have successfully overcome GEO and CoreCivic’s arguments that state law claims constitute either direct regulation or discrimination against the federal government.[407] A state regulation that merely increases costs to the federal government generally does not constitute direct regulation in violation of intergovernmental immunity.[408] So while the federal government might ultimately absorb the increased labor costs to run immigration detention facilities, courts have found state minimum wage laws do not constitute unlawful direct regulation.[409] Additionally, the Central District of California found GEO was not an “instrumentality” of the federal government and therefore could not appeal to intergovernmental immunity to shield it from direct state regulation.[410]

The issue of discrimination is arguably more complex. State laws are discriminatory if they “single out” the federal government or its contractors or regulate them unfavorably due to their governmental status.[411] To successfully invoke immunity from this discriminatory treatment, the government or its contractor must identify a “comparable reference group” and prove inferior relative treatment.[412] This is particularly difficult in the context of immigration detention, which “has no clear analogs.”[413] Immigration detention is an “ostensibly civil system that notoriously mimics criminal incarceration in key respects.”[414] Because immigration detention is a civil system, the Thirteenth Amendment Exceptions Clause permitting forced labor of people convicted of crimes does not apply. This makes jails and prisons an improper comparable reference. But what about other forms of civil detention? Most states do not apply minimum wage laws to pretrial detainees,[415] civilly committed sex offenders,[416] or involuntarily committed psychiatric patients.[417] But the Ninth Circuit rejected these comparisons in the combined Washington cases.[418] Unlike most criminal incarceration and other forms of civil detention, immigration detention is almost entirely privately operated.[419] Accordingly, the Ninth Circuit held that the comparable reference group is not other state facilities—but rather privately operated state facilities.[420] While there are no such facilities in Washington, the divided Ninth Circuit embraced the Washington Supreme Court’s ruling that the Minimum Wage Act would apply equally to such a facility.[421] In Novoa, the Central District of California denied cross motions for summary judgment on intergovernmental immunity due to “material disputes.”[422] But the court similarly embraced other privately run state facilities as the appropriate comparison group.[423]

Intergovernmental immunity remains one of the greatest obstacles to bringing state claims on behalf of detained workers. But recent case law has limited the defense to claims of discrimination, dismissing direct regulation claims. Detainees, litigators, and advocates are likely to have the most success defending against these claims by emphasizing the unique nature of the privatized civil facilities and directing courts to continue down their line of reasoning requiring close comparators.

Conclusion

Success to me is to put a stop something that’s unjust. To me, that’s success. Well, that’s a small victory. Success would be like these systems don’t exist. That would be victory, having this exploitation cease . . . . We know this is an ongoing battle, as long as these places exist there are gonna be violations.

—Jose Ruben Hernandez Gomez.[424]

In 2023, two years after their historic jury trial win, detainees at the Northwest Detention Center went on seven hunger strikes.[425] In February of 2024, armored guards deployed tear gas against hunger strikers and held them at gunpoint.[426] And even after a jury found detained workers were owed a legal wage, GEO refused and instead ended the work program entirely at NWDC.[427] The fallout since the lawsuit at NWDC exposes the reality of immigration detention: Workers’ rights violations are only one of the many human rights violations permeating the privatized system.

The federal and state law claims outlined in this Note offer one path to challenge this dehumanization entrenched in the immigration detention system and disrupt the industry that profits from it. While the Trafficking Victims Prevention Act remains arguably the most successful federal cause of action, advocates and litigators still have room to argue for the application of the Thirteenth Amendment and the Fair Labor Standards Act to detained workers. Meanwhile, state claims remain an underappreciated path with great potential. Recent victories in Washington, Colorado, and California demonstrate that state minimum wage and unjust enrichment claims provide a potent legal tool to challenge the for-profit detention industry, successfully overcoming complex Supremacy Clause defenses.

Multidisciplinary and trans-substantive advocacy, inside and outside the courtroom, offers the best chance of successfully challenging private immigration detention through workers’ rights. Policy advocates can further expand state minimum wage laws to include detained workers. Litigators can continue to push back on private federal contractors shielding themselves behind Supremacy Clause defenses. Academics can demystify the undertheorized claim of unjust enrichment. There is also a need for greater research and partnership between detained workers and advocates, litigators, and academics to understand organizing efforts within and outside detention centers as well as opportunities and pitfalls of state and federal administrative remedies. Ultimately, by targeting the corporate greed that sustains the system, these legal challenges seek to disrupt the system, restore human dignity, and truly champion workers’ rights for all.

Copyright © 2026 Rebecca Cooley, J.D. 2025, University of California, Berkeley, School of Law. This Note is dedicated to immigration detainees facing and fighting injustice. This Note was only possible thanks to the advocates championing detained workers’ rights, especially those who took time away from their efforts to share their knowledge with me. A very special thank you in particular to Jose Ruben Hernandez Gomez, Maru Mora Villalpando, Andrew Free, Mariel Villarrel, Priya Patel, and Serafin Andrade. Thank you to my professors Diana Reddy, Alexa Koenig, Eric Stover, David Hausman, and Urmila Taylor for your insights and support. And thank you to the excellent team at the California Law Review.

       [1]. J. Cipriano Ríos Alegria, Angelica Cházaro, Tomás Madrigal, Murphy Mitchel, Josefina Mora, Maru Mora Villalpando, Wendy Pantoja, Naomi Strand & Megan Ybarra, A Hunger Strikers Handbook 11 (2017) [hereinafter Hunger Strikers Handbook] (emphasis added).

       [2]. ICE rebranded and renamed NWDC as the Northwest ICE Processing Center (NWIPC), but the facility remains operated by the same for-profit corporation, The GEO Group, Inc. (GEO). See Cascade PBS Staff, ICE Rebranded the Tacoma Detention Center. It’s Still Inhumane, Cascade PBS (Sep. 24, 2019), https://www.cascadepbs.org/opinion/2019/09/ice-rebranded-tacoma-detention-center-its-still-inhumane/ [https://perma.cc/53Y7-SVGC].

[3]. See Alex Altman, Prison Hunger Strike Puts Spotlight on Immigration Detention, TIME (Mar. 17, 2014), https://time.com/27663/prison-hunger-strike-spotlights-on-immigration-detention/ [https://perma.cc/PQ5W-2NCW]; Hunger Strikers Handbook, supra note 1, at 16.

       [4]. After 56 Days, Northwest Detention Center Hunger Strike Concludes, Not1More (May 5, 2014), https://web.archive.org/web/20190715221547/http://www.notonemoredeportation.com/2014/05/05/after-56-days-northwest-detention-center-hunger-strike-concludes-newly-formed-collective-of-detainees-releases-statement/ [https://perma.cc/G3P3-C4M2] (sourced via Internet Archive).

       [5]. Angélica Cházaro & Dan Berger, Guest: What’s Behind the Hunger Strike at Northwest Detention Center, Seattle Times (Mar. 19, 2014), https://www.seattletimes.com/opinion/guest-whatrsquos-behind-the-hunger-strike-at-northwest-detention-center/ [https://perma.cc/WT39-R4VM]; see also Hunger Strikers Handbook, supra note 1, at 19.

       [6]. NWDC is adjacent to the Tacoma Tideflats, “where a coal gasification plant leeched toxic sludge into the soil for more than thirty years.” David Pellow & Jasmine Vazin, The Intersection of Race, Immigration Status, and Environmental Justice, 11 Sustainability 1, 10 (2019); see also Rico Moore, In Toxic Detention, Nation (May 15, 2025), https://www.thenation.com/article/environment/margin-tacoma-toxic-detention/ [https://perma.cc/7QZF-TE3H] (providing maps showing the sources of contamination in the area by the detention center).

[7]. About the Northwest Detention Center, Nw. Immigrant Rts. Project (2021), https://www.nwirp.org/resources/nwdc/ [https://perma.cc/PZ5K-EXN3].

       [8]. Lewis Kamb, A Rare Look Inside Tacoma’s Northwest Detention Center, News Trib. (Jan. 31, 2017), https://www.thenewstribune.com/news/special-reports/article25860412.html [https://perma.cc/H76B-SA2J] (“[T]he Tacoma City Council . . . anticipated the new facility ‘would bring hundreds of family-wage job opportunities’ . . . . As it turned out, only about 45 jobs initially were created.”).

       [9]. The company that secured the original contract for the detention center was the Correctional Services Corporation (CSC). See Contract Between Corr. Servs. Corp. and Immigr. and Customs Enf’t (2002), https://www.documentcloud.org/documents/2631226-Northwest-Detention-Center-CDF-Contract-2002 [https://perma.cc/8M4Y-SE2L]. GEO purchased CSC and took over the facility and its contract with ICE in 2005. GEO Group Buys Out Correctional Services Corporation, Prison Legal News (Jan. 15, 2006), https://www.prisonlegalnews.org/news/2006/jan/15/geo-group-buys-out-correctional-services-corporation/ [https://perma.cc/E9PH-SKE4]; The GEO Group, Annual Report (Form 10-K), at 8 (Mar. 2, 2007), https://www.sec.gov/Archives/edgar/data/923796/000095014407001829/g05716e10vk.htm [https://perma.cc/4E34-GU4R].

     [10]. Complaint ¶ 4.4, Nwauzor v. GEO Grp., Inc., No. 17-cv-05769-RJB, 2017 WL 11646687 (W.D. Wash. Sep. 26, 2017) (“Rather than hire from the local workforce, GEO relies upon captive detainee workers to clean, maintain, and operate NWDC.”).

[11].Nwauzor v. GEO Grp., Inc., 127 F.4th 750, 758 (9th Cir. 2025).

     [12]. Complaint ¶¶ 4.5, 4.7, Nwauzor, 2017 WL 11646687.

     [13]. Martin Kaste, Detainees Who Earned Just $1 a Day Are Owed $17 Million in Back Pay, a Jury Says, NPR (Oct. 29, 2021), https://www.npr.org/2021/10/29/1050520220/detainees-who-earned-just-1-a-day-are-owed-17-million-in-back-pay-a-jury-orders [https://perma.cc/WZ3T-E4FS].

[14]. See Washington v. GEO Grp., Inc., 283 F. Supp. 3d 967, 973 (W.D. Wash. 2017).

     [15]. This Note will generally use the term “detained workers” to refer to all those detained who labor in immigration detention in any capacity. While imperfect, this term highlights that those laboring are workers who merit protections, in spite or arguably because of their involuntary residence.

[16].Chen v. GEO Grp., Inc., 287 F. Supp. 3d 1158, 1161 (W.D. Wash. 2017). The lead named plaintiff in this class action, Chao Chen, was later replaced by Ugochukwu Goodluck Nwauzor, and the case name changed accordingly. See Nwauzor v. GEO Grp., Inc., No. 17-cv-05769-RJB, 2021 WL 5824384 (W.D. Wash. Dec. 8, 2021).

     [17]. Verdict and Settlement Summary, Nwauzor, 2021 WL 9374966; Verdict, Agreement and Settlement, Washington v. GEO Grp., Inc., No. 17-cv-05769-RJB, 2021 WL 5571111 (W.D. Wash. Oct. 27, 2021).

[18].Nwauzor v. GEO Grp., Inc., 540 P.3d 93, 93–94 (Wash. 2023); Nwauzor v. GEO Grp., Inc., 127 F.4th 750, 756 (9th Cir. 2025).

[19].Nwauzor v. GEO Grp., Inc., 146 F.4th 1280, 1281 (9th Cir. 2025).

     [20]. Petition for Writ of Certiorari, GEO Grp., Inc., v. Nwauzor, No. 25A464 (U.S. Jan. 9, 2026). The Supreme Court recently ruled unanimously against GEO in a similar case by detained workers. Geo Grp., Inc. v. Menocal, No. 24-758, 2026 WL 513536 (U.S. Feb. 25, 2026).

     [21]. Gene Johnson, GEO Group Halts Work Program at Tacoma Jail Instead of Upping Detainee Pay, KNKX Pub. Radio (Nov. 5, 2021), https://www.knkx.org/law/2021-11-05/geo-group-halts-work-program-at-tacoma-jail-instead-of-upping-detainee-pay [https://perma.cc/8ET9-DFMN].

     [22]. Brief of La Resistencia, Fair Work Center & Prof. Angelina Snodgrass Godoy as Amici Curiae in Support of Appellee at 11, 16, 25, Nwauzor v. GEO Grp., Inc.,Nos. 21-36024 & 21-36025 (9th Cir. May 27, 2022).

[23]. Id. at 15–21 (detailing worsening conditions after ending the work program at NWDC, including dirty floors and bathrooms, insufficient food service, and even increased commissary prices).

[24]. See Dep’t of Lab. & Indus. v. GEO Secure Servs., LLC, No. C24-5095-BHS, 2024 WL 3250426, at *12 (W.D. Wash. July 1, 2024); Bill Lucia, State Workplace Inspectors Will Have Access to Immigration Detention Center in Tacoma, Wash. State Standard (Oct. 2, 2024), https://washingtonstatestandard.com/2024/10/02/state-workplace-inspectors-will-have-access-to-immigration-detention-center-in-tacoma/ [https://perma.cc/ND9S-JSSU].

[25]. Seeinfra Part I.B.

[26]. See Jacqueline Stevens, One Dollar Per Day: The Slaving Wages of Immigration Jail, From 1943 to Present, 29 Geo. Immigr. L.J. 391, 402 (2015); see alsoinfra Part I.B (describing the for-profit immigration detention model); infra note 290 and accompanying text(countering claims that private detention centers would benefit economically from paying full labor costs).

[27]. See infra notes 135–37 and accompanying text.

     [28]. For further discussion on the role of litigation and non-litigation avenues in remedying violations of detained workers’ rights, see infra Part I.E.

[29]. See e.g., Owino v. CoreCivic, Inc., 60 F.4th 437, 450 (9th Cir. 2022); Barrientos v. CoreCivic, Inc., 951 F.3d 1269, 1280 (11th Cir. 2020); Menocal v. GEO Grp., Inc., 635 F. Supp. 3d 1151, 1199 (D. Colo. 2022), appeal dismissed, No. 22-1409, 2024 WL 4544184, at 4 (10th Cir. Oct. 22, 2024), aff’d, No. 24-758, 2026 WL 513536 (U.S. Feb. 25, 2026); Novoa v. GEO Grp., Inc., No. 5:17-cv-02514-JGB-SHK, 2022 WL 2189626, at 27 (C.D. Cal. Jan. 25, 2022).

[30]. See Nwauzor v. GEO Grp., Inc., 127 F.4th 750 (9th Cir. 2025); Menocal v. GEO Grp., Inc., No. 22-1409, 2024 WL 4544184 (10th Cir. Oct. 22, 2024), aff’d, No. 24-758, 2026 WL 513536 (U.S. Feb. 25, 2026).

     [31]. U.S. Immigr. & Customs Enf’t, Annual Report 22 (2024), https://www.ice.gov/doclib/eoy/iceAnnualReportFY2024.pdf [https://perma.cc/UC88-WLLK].

     [32]. Zoom Interview with Jose Ruben Hernandez Gomez (Mar. 20, 2024).

[33]. United States, Glob. Det. Project, https://www.globaldetentionproject.org/countries/americas/united-states [https://perma.cc/UPJ5-A4P2].

     [34]. U.S. Immigr. & Customs Enf’t, supra note 31, at 23; see alsoICE Enforcement and Removal Operation Statistics, U.S. Immigr. & Customs Enf’t, https://www.ice.gov/statistics [https://perma.cc/TN7L-W55R].

     [35]. Suzanne Gamboa, Julia Ainsley, Gabe Gutierrez & Laura Strickler, Trump’s Stepped-Up Immigration Arrests Escalate Need for More Detention Space, NBC News (Jan. 31, 2025), https://www.nbcnews.com/news/latino/trumps-stepped-immigration-arrests-escalate-need-detention-space-rcna190217 [https://perma.cc/Q7BN-5LC2].

[36]. See Human Rights First, U.S. Detention of Asylum Seekers: Seeking Protection, Finding Prison 2 (2009); My Khanh Ngo & Judy Rabinovitz, Asylum Seeker’s Wrongful 6-Year Detention is Emblematic of a Broken System, ACLU (Oct. 10, 2023), https://www.aclu.org/news/immigrants-rights/asylum-seekers-wrongful-6-year-detention-is-emblematic-of-a-broken-system [https://perma.cc/CU7M-CNPK].

[37]. See Immigr. Pol’y Ctr., Am. Immigr. Council, The Ones They Leave Behind: Deportation of Lawful Permanent Residents Harms U.S. Citizen Children 1 (2010), https://www.americanimmigrationcouncil.org/wp-content/uploads/2025/01/Childs_Best_Interest_Fact_Sheet_042610.pdf [https://perma.cc/9SBM-WSTS] (noting that 10 percent of all people deported each year are legal permanent residents).

[38]. See U.S. Gov’t Accountability Off., GAO-21-487, Immigration Enforcement: Actions Needed to Better Track Cases Involving U.S. Citizenship Investigations 21–24 (2021), https://www.gao.gov/assets/gao-21-487.pdf (finding that since 2015, ICE has arrested 674, detained 121, and deported 70 potential U.S. citizens).

     [39]. U.S. Immigr. & Customs Enf’t, supra note 31, at 22.

[40]. See Juliet P. Stumpf, Civil Detention and Other Oxymorons, 40 Queen’s L.J. 55, 63–65 (2014) (citing The Chinese Exclusion Case, 130 U.S. 58, 582 (1889); Korematsu v. United States, 323 U.S. 214, 216 (1944)).

[41]. See Anita Sinha, Slavery by Another Name: Voluntary Immigrant Detainee Labor and the Thirteenth Amendment, 11 Stan. J. C.R. & C.L. 1, 8 (2015) (quoting William D. Owen, Report Of The Select Committee On Immigration And Naturalization, H.R. Rep. No. 51-3472, at II–III (1891)); see also César Cuauhtémoc García Hernández, The Perverse Logic of Immigration Detention: Unraveling the Rationality of Imprisoning Immigrants Based on Markers of Race and Class Otherness, 1 Colum. J. Race & L. 353, 357–58 (2012).

     [42]. Emily Kassie, Detained: How the U.S. Created the Largest Immigrant Detention System in the World, Marshall Project (Sep. 24, 2019), https://www.themarshallproject.org/2019/09/24/detained [https://perma.cc/X8DW-EP9L].

[43]. See Sinha, supra note 41, at 13–16.

[44]. See 8 U.S.C. § 1225(b)(1)(B)(iii)(IV) (requiring mandatory detention of asylum-seekers pending a final determination of credible fear of persecution).

     [45]. Id. § 1225(b)(1)(C); see also Omnibus Consolidated Appropriations Act, 1997, Pub. L. No. 104–208, 110 Stat. 3009 (1996).

[46]. See Sinha, supra note 41, at 14–16.

     [47]. Laken Riley Act, Pub. L. No. 119-1, 139 Stat. 3 (2025).

     [48]. García Hernández, supra note 41, at 360.

[49]. See Stumpf, supra note 40, at 58 (“[M]ass immigration detention grew up in the same time and space as mass incarceration, shar[es] the same facilities and actors to achieve a nearly identical restraint on liberty.”).

     [50]. Sinha, supra note 41, at 20.

[51]. See Immigration and Nationality Act § 287(g), 8 U.S.C. § 1357(g); Delegation of Immigration Authority Section 287(g) Immigration and Nationality Act, Immigr. & Customs Enf’t (2025), https://www.ice.gov/identify-and-arrest/287g [https://perma.cc/XF56-JWLC]; ImmigrationDetainers: An Overview, Am. Immigr. Council (Mar. 21, 2017), https://www.americanimmigrationcouncil.org/research/immigration-detainers-overview [https://perma.cc/C5ED-53UA]; Arrest to Deportation Pipeline, Cal. Immigr. Pol’y Ctr., https://caimmigrant.org/what-we-do/policy/arrest-to-deportation-pipeline/ [https://perma.cc/HXN6-MHFM].

     [52]. Jennifer Safstrom, Thirteenth Amendment Litigation in the Immigration Detention Context, 26 Mich. J. Race & L. 205, 216 (2020).

     [53]. Seth H. Garfinkel, The Voluntary Work Program: Expanding Labor Laws to Protect Detained Immigrant Workers, 67 Case W. Rsrv. L. Rev. 1287, 1312 (2017); see alsoinfra Part II.A (discussing the similarities and differences between civil immigration detention and criminal incarceration).

     [54]. Some immigration offenses, such as unlawful re-entry, are criminal offenses with criminal penalties. See 8 U.S.C. §§ 1325(a), 1326. But presence in the United States in violation of immigration laws, such as overstaying a visa, is a civil rather than criminal offense. See 8 U.S.C. § 1227; Safstrom, supra note 52, at 216.

[55]. See Fong Yue Ting v. United States, 149 U.S. 698, 730–31 (1893).

     [56]. See Wong Wing v. United States, 163 U.S. 228, 235 (1896); U.S. Immigr. & Customs Enf’t, Performance-Based National Detention Standards 2011, at i (2016) [hereinafter PBNDS], https://www.ice.gov/doclib/detention-standards/2011/pbnds2011r2016.pdf [https://perma.cc/LM2W-XFE2] (“ICE detains people for no purpose other than to secure their presence both for immigration proceedings and their removal . . . .”).

[57]. See Fong Yue Ting, 149 U.S. at 730–31 (reasoning that because deportation is “not a punishment for a crime,” it does not constitute a deprivation “of life, liberty, or property without due process of law” that would entitle the procedural protections of the criminal justice system); Sinha, supra note 41, at 5, 8, 11; Garfinkel, supra note 53, at 1312 n.158 (detailing the various procedural protections available in the criminal justice system but not in the immigration context including Miranda warnings, appointed counsel, right to a bail hearing, right to a speedy trial, prohibition of illegally obtained evidence, and right to trial before jury).

[58]. See ACLU & Physicians for Hum. Rts., Behind Closed Doors: Abuse and Retaliation Against Hunger Strikers in U.S. Immigration Detention (2021), https://www.aclu.org/report/report-behind-closed-doors-abuse-retaliation-against-hunger-strikers-us-immigration-detention [https://perma.cc/5SA6-S6TD]. Plaintiffs in Menocal v. GEO Group and Hernandez Gomez v. GEO Group were deported, disrupting proceedings against GEO. See e.g.,Writ of Habeas Corpus Ad Testificandum, Menocal v. GEO Grp., Inc., No. 1:14-cv-02887-JLK-MEH (D. Colo. June 11, 2020). In Alvarado Guevara v. INS, two named plaintiffs were summoned back to the United States for depositions after being deported. See Plaintiff’s Proposed Order to Show Cause for Preliminary and Permanent Injunction, No. B-86-106 (S.D. Tex. Mar. 16, 1987). After their depositions, Border Patrol agents detained them at the airport and allegedly assaulted them. Id. In Hernandez Gomez v. GEO, one of the named plaintiffs alleged sexually invasive pat downs and permanently disabling refeeding tactics after hunger striking against labor conditions. Justo Robles, Californian Who Joined Hunger Strike in ICE Detention Seeks $1 Million in Complaint, CalMatters (Dec. 22, 2023), https://calmatters.org/california-divide/2023/12/ice-detention-california/ [https://perma.cc/X75G-73R2].

     [59]. For example, the Trafficking Victims Protection Act (TVPA) is both a criminal statute and a trafficking statute. Successful TVPA claims theoretically open up the possibility of pursuing U visas (available for victims of crime) or T visas (available for victims of trafficking) on behalf of at least named plaintiffs and potentially unnamed class members. Recently, one of the named plaintiffs in a TVPA class action in Colorado obtained certification for a U visa. See Emergency Motion for U Visa Certification,Menocal v. GEO Grp., Inc., No. 1:14-cv-2887-JLK-CYC (D. Colo. Sep. 15, 2025). But most federal or state subminimum wage claims would not provide any immigration relief as these labor violations generally do not relate to trafficking or criminal liability. Seeinfra Parts II.B, III.A.

     [60]. For example, in Hernandez Gomez v. GEO Group, lead counsel Rosen Bien Galvan and Grunfeld bring expertise suing prisons and litigating workers’ rights. Their co-counsel, the California Collaborative for Immigrant Justice, brings expertise from serving residents in immigration detention through direct legal services and advocacy. See Second Amended Complaint, Hernandez Gomez v. GEO Grp., Inc., No. 1:22 Civ. 00868-ADA-CDB, 2022 WL 19266228 (E.D. Cal. Dec. 23, 2022).

     [61]. Wesley Carrasco & Megan Ybarra, Unjust Enrichment: The Struggle for Fair Wages and an End to All Immigrant Cages, ISSU (June 20, 2021), https://web.archive.org/web/20240814052024if_/https:/issuu.com/muybarra/docs/unjust_enrichment_updated_14jun2021 [https://perma.cc/5S7H-E8FC].

[62]. See Eunice Cho, Unchecked Growth: Private Prison Corporations and Immigration Detention, Three Years Into the Biden Administration, ACLU (Aug. 7, 2023), https://www.aclu.org/news/immigrants-rights/unchecked-growth-private-prison-corporations-and-immigration-detention-three-years-into-the-biden-administration [https://perma.cc/BW37-V7CH].

     [63]. Jonathon Booth, Ending Forced Labor In ICE Detention Centers: A New Approach, 34 Geo. Immigr. L.J. 573, 579 (2020). GEO and CoreCivic are not the only private detention center operators, and similar labor lawsuits are pending against these smaller companies as well. See Yeend v. Akima Glob. Servs., LLC, 347 F.R.D. 405 (N.D.N.Y. 2024); Barahona v. LaSalle Mgmt. Co., LLC, No. 7:23-CV-24-WLS, 2025 WL 961437 (M.D. Ga. Mar. 31, 2025).

     [64]. Cho, supra note 62.

[65]. Id.

[66]. See Emily Stewart, Trump’s Secret Weapon on Immigration, Bus. Insider (Feb. 3, 2025), https://www.businessinsider.com/private-prisons-profit-trump-executive-order-immigration-deportation-geo-corecivic-2025-2 [https://perma.cc/Z72A-YVTY]; Lauren-Brooke Eisen, Private Prison Companies’ Enormous Windfall: Who Stands to Gain as ICE Expands, Just Sec. (Sep. 24, 2025), https://www.justsecurity.org/121226/private-prison-companies-gain-ice-expands/ [https://perma.cc/74N7-BKXV].

[67].Stewart, supra note 66.

     [68]. Booth, supra note 63, at 580.

[69]. Id.

[70]. Id.

     [71]. The GEO Group, Annual Report (Form 10-K), at 70 (Feb. 29, 2024), https://investors.geogroup.com/static-files/b6d0f4ba-9763-408a-a707-7cb552b65b80 [https://perma.cc/A9NC-MRGJ].

     [72]. CoreCivic, Annual Report (Form 10-K), at 74 (Feb. 20, 2024), https://ir.corecivic.com/static-files/94995c0f-831c-40e6-858c-298a6a570bc6 [https://perma.cc/L9MM-JBEW].

     [73]. Stevens, supra note 26, at 402.

     [74]. Zoom Interview with Jose Ruben Hernandez Gomez (Mar. 20, 2024).

     [75]. Declaration of Demetrio A. Valerga ¶ 4, Menocal v. GEO Grp., Inc., No. 1:14-cv-02887-JLK-CYC (D. Colo. May 6, 2016).

[76]. Id.

     [77]. After conducting extensive Freedom of Information Act litigation, Professor Stevens estimated that detention centers generally pay detained workers one to three dollars per day. Stevens, supra note 26, at 415–17.

     [78]. Amended Complaint for Declaratory and Injunctive Relief and Damages ¶ 9, Barrientos v. CoreCivic, Inc., No. 4:18-cv-00070-CDL, 2020 WL 7021904 (M.D. Ga. Apr. 17, 2018).

[79]. See 29 U.S.C. § 206(a)(1)(C).

     [80]. Camille Pendley, Immigrants Seeking Asylum Face Long Odds in Stewart County Detention Center, Creative Loafing (Nov. 2, 2016), https://creativeloafing.com/content-267459-immigrants-seeking-asylum-face-long-odds-in-stewart-county-detention [https://perma.cc/GQU2-WYMY].

     [81]. Detention Standards, U.S. Immigr. & Customs Enf’t, https://www.ice.gov/detain/detention-management [https://perma.cc/4TRT-X6WU].

     [82]. PBNDS, supra note 56, at 405–09.

[83]. Id. at 405.

[84]. Id.

[85]. Id. at 407.

[86]. Id.

     [87]. Stevens, supra note 26, at 409.

     [88]. See An Act to Authorize Certain Administrative Expense for the Department of Justice, and Other Purposes, Pub. L. No. 81-503, § 6, 64 Stat. 380 (1950).

     [89]. 8 U.S.C. § 1555(d).

[90]. See Department of Justice Appropriation Act of 1978, Pub L. No. 95-86, 91 Stat. 419; An Act to Authorize Certain Administrative Expenses for the Department of Justice and for Other Purposes, Pub L. No. 81-626, 64 Stat. 380 (1950) (codified at Pub. L. No. 89-554, 80 Stat. 378, 656 (1966)).

[91]. Consumer Price Index Inflation Calculator, U.S. Bureau of Lab. Stat., https://data.bls.gov/cgi-bin/cpicalc.pl?cost1=1.00&year1=195001&year2=202506 [https://perma.cc/R8UB-YMXG] (based on September 2025 data).

[92]. See Department of Justice Appropriation Act of 1978, Pub L. No. 95–86, 91 Stat. 426 (final year where dollar-per-day program is included in the appropriations bill);Stevens, supra note 26, at 465–66.

[93]. See U.S. Dep’t of Homeland Sec., Off. of Inspector Gen., OIG-18-32, Concerns about ICE Detainee Treatment and Care at Detention Facilities 8 (2019) [hereinafter DHS OIG Report], https://www.oig.dhs.gov/sites/default/files/assets/2019-06/OIG-19-47-Jun19.pdf [https://perma.cc/Q5QY-9X2X] (reporting on potentially unsafe food at four different facilities, including “spoiled, wilted, and moldy produce . . . as well as food past its expiration date”); see also Cal. Collaborative for Immigr. Just., Starving for Justice: the Denial of proper nutrition in Immigration Detention 5–10 (2022), https://www.ccijustice.org/_files/ugd/733055_c43b1cbbdda341b894045940622a6dc3.pdf [https://perma.cc/GZG4-KL6N]; Univ. of Wash., Ctr. for Hum. Rts., Conditions at NWDC: Sanitation of Food & Laundry (2020), https://jsis.washington.edu/humanrights/2020/03/27/nwdc-sanitation-of-food-laundry/ [https://perma.cc/LA8B-XQD8].

[94]. See Jack Herrera, In ICE Detention, Forced to Pay for Soap, Nation (Apr. 30, 2020), https://www.thenation.com/article/politics/coronavirus-ice-detention-soap/ [https://perma.cc/R7DL-EVDR]; DHS OIG Report, supra note 93, at 7 (reporting on the lack of basic hygiene supplies such as toilet paper, shampoo, soap, lotion, and toothpaste in at least two detention centers).

[95]. See Shannon Najmabadi, Detained Migrant Parents Have to Pay to Call Their Family Members. Some Can’t Afford To, Texas Trib. (July 3, 2018), https://www.texastribune.org/2018/07/03/separated-migrant-families-charged-phone-calls-ice/ [https://perma.cc/RJ3C-W8J3]; Alexandra Starr, At Low Pay, Government Hires Immigrants Held At Detention Centers, NPR (July 23, 2015), https://www.npr.org/2015/07/23/425511981/at-low-pay-government-hires-immigrants-held-at-detention-centers [https://perma.cc/MTF9-MW4H] (quoting Professor Nancy Hiemestra, remarking that “[s]ome people will work for two weeks just to make one phone call”).

     [96]. Michelle Conlin & Kristina Cooke, $11 Toothpaste: Immigrants Pay Big for Basics at Private ICE Lock-Ups, Reuters (Jan. 18, 2019), https://www.reuters.com/article/idUSKCN1PC0DJ/ [https://perma.cc/ATV3-Y9KU].

     [97]. Second Amended Complaint ¶ 89, Hernandez Gomez v. GEO Grp., Inc., No. 1:22 Civ. 00868-ADA-CDB, 2022 WL 19266228 (E.D. Cal. Dec. 23, 2022).

     [98]. Complaint for Declaratory and Injunctive Relief and Damages ¶¶ 1–2, Barrientos v. CoreCivic, Inc., No. 4:18-cv-00070-CDL, 2020 WL 7021904 (M.D. Ga. Apr. 17, 2018); see also Second Amended Complaint ¶ 89, Hernandez Gomez, 2022 WL 19266228; Samantha Sherman, Defining Forced Labor: The Legal Battle to Protect Detained Immigrants from Private Exploitation, 88 U. Chi. L. Rev. 1201, 1214 (2021); Jamila S. Cambridge, Land of the Free? An Examination of the Constitutionality of Forced Labor in U.S. Immigrant Detention Centers, 63 How. L.J. 405, 410 (2020).

[99]. See Stevens, supra note 26, at 416–17.

[100]. See, e.g.,Third Amended Complaint ¶¶ 139–40, Novoa v. GEO Grp., Inc., No. 5:17-cv-02514-JGB-SHK, 2019 WL 8329599 (C.D. Cal. Sep. 16, 2019) (alleging some detainees worked without pay for months before being hired into the work program); see also Mia Steinle, Slave Labor Widespread at ICE Detention Centers, Lawyers Say, Project on Gov’t Oversight (Sep. 7, 2017), https://www.pogo.org/investigations/slave-labor-widespread-at-ice-detention-centers-lawyers-say [https://perma.cc/PT7A-NNAK] (reporting on waitlists for job placements at NWDC).

[101]. See Yana Kunichoff, “Voluntary” Work Program in Private Detention Centers Pays Detained Immigrants $1 a Day, Prison Legal News (Aug. 15, 2012), https://www.prisonlegalnews.org/news/2012/aug/15/voluntary-work-program-in-private-detention-centers-pays-detained-immigrants-1-a-day/ [https://perma.cc/M7VV-LQPN]; Sherman, supra note 98, at 1210–11; Sinha, supra note 41, at 33.

[102]. See Steinle, supra note 100 (“They [GEO] make you feel like you are the one requesting the job.”).

   [103]. U.S. Const. amends. VI–VII.

[104]. See Sinha, supra note 41, at 4–5 (noting that the timing in “immigration detention is inherently indeterminate” (citations omitted)).

[105]. See Christie Thompson & Andrew Rodriguez Calderón, More Immigrants Are Giving Up Court Fights and Leaving the U.S., Marshall Project (May 8, 2019), https://www.themarshallproject.org/2019/05/08/more-detained-immigrants-are-giving-up-court-fights-and-leaving-the-u-s [https://perma.cc/GM7H-ESZL] (“‘Voluntary departure’ applications surge as immigrants decide it’s better to return to their native countries than languish in a detention facility.”).

   [106]. Interview with Maru Mora Villalpando (Apr. 8, 2024).

[107]. See Brief for the United States as Amicus Curiae in Support of Neither Party, Barrientos v. CoreCivic, Inc., 951 F.3d 1269 (11th Cir. 2020) (No. 18-15081).

[108]. See, e.g.,Todd Schulte, Tearing Families Apart: The Impact of Trump’s Immigration Agenda, FWD.us (Sep. 29, 2020), https://www.fwd.us/news/the-impact-of-trumps-immigration-agenda/ [https://perma.cc/7QQX-BQ6F]; A Timeline of The Trump Administration’s Efforts to End Asylum, Nat’l Immigr. Just. Ctr., https://immigrantjustice.org/timeline-trump-administrations-efforts-end-asylum [https://perma.cc/QQT8-7LTN].

[109]. See Brief for the United States as Amicus Curiae in Support of Appellant, Washington v. GEO Grp., Inc., 127 F.4th 783 (9th Cir. 2025) (Nos. 21-36024, 21-36025).

[110]. See, e.g., Alvarado Guevara v. INS, 902 F.2d 394 (5th Cir. 1990); see also Channer v. Hall, 112 F.3d 214 (5th Cir. 1997).

[111]. See generally Erwin Chemerinsky, Closing the Courthouse Door: How Your Constitutional Rights Became Unenforceable 20–92 (2017).

[112]. SeeAlvarado Guevara, 902 F.2d at 396 n.2 (“Further evidence that alien detainees are not government ‘employees’ lie in critical features of government employment.”). But see Brief for Hearing En Banc at 8–10, Alvarado Guevara, 902 F.2d 394 (“Even if Defendants were forbidden to employ aliens, the Defendants must pay the minimum wage if they nevertheless do, in fact, employ aliens.”).

[113]. See, e.g.,Menocal v. GEO Grp., Inc., 635 F. Supp. 3d 1151, 1181 (D. Colo. 2022) (“ICE is aware of the claims in this case and has not affirmatively asserted any interest.”), appeal dismissed, No. 22-1409, 2024 WL 4544184, at 4 (10th Cir. Oct. 22, 2024), aff’d, No. 24-758, 2026 WL 513536 (U.S. Feb. 25, 2026); Chen v. GEO Grp. Inc., No. 3:17-CV-05769-RJB, 2018 WL 1963669, at 3 (W.D. Wash. Apr. 26, 2018); Washington v. GEO Grp. Inc., No. 3:17-CV-05806-RJB, 2018 WL 1963792, at *3 (W.D. Wash. Apr. 26, 2018)(finding that ICE is aware of the pending litigation but has not appeared in any similar case or asserted any interest). But see Statement of Interest of the United States, Washington v. GEO Grp., Inc., No. 3:17-cv-05806-RJB (W.D. Wash. Aug. 20, 2019) (claiming that the lawsuit by the state of Washington interferes with federal immigration enforcement).

[114]. See U.S. Const. art. VI, cl. 2.

[115]. See generally Kate Sablosky Elengold & Jonathan D. Glater, The Sovereign Shield, 73 Stan. L. Rev. 969 (2021); David S. Rubenstein & Pratheepan Gulasekaram, Privatized Detention & Immigration Federalism, 71 Stan. L. Rev. Online 224 (2019).

   [116]. For example, California’s Occupational Safety and Health Administration (OSHA) fined GEO for violating workplace safety measures for detained workers, thereby defining them as employees protected by the agency. OSHA, Inspection Detail – The Geo Group, Inc. Dba Golden State Annex (Report ID: 0950647), (2022), https://www.osha.gov/ords/imis/establishment.inspection_detail?id=1609228.015 [https://perma.cc/FQ6V-G9VU].

[117]. See, e.g., GEO’s Notice of Removal to Federal Court, Washington v. GEO Grp., Inc., No. 3:17-cv-05806, 2017 WL 11646686 (W.D. Wash. Aug. 14, 2017). For further discussion, see infra Part III.

[118]. See Lydia Saad, Americans Showing Increased Concern About Immigration, Gallup (Feb. 13, 2023), https://news.gallup.com/poll/470426/americans-showing-increased-concern-immigration.aspx [https://perma.cc/9P5R-ZAGM]; Fidel Martinez, Is America Turning Anti-Immigrant?, L.A. Times (Jan. 17, 2024), https://www.latimes.com/delos/newsletter/2024-01-17/migrants-drowning-texas-greg-abbott-immigration-operation-lone-star-latinx-files-latinx-files [https://perma.cc/26EE-84AA].

   [119]. See infra note 140 and accompanying text.

   [120]. For example, the Washington Attorney General brought suit against GEO after years of organizing and political pressure. See Washington v. GEO Grp., Inc., 283 F. Supp. 3d 967 (W.D. Wash. 2017).

[121]. Seeinfra note 139 and accompanying text.

   [122]. Third Amended Complaint ¶ 96, Novoa v. GEO Grp., Inc., No. 5:17-cv-02514-JGB-SHK, 2019 WL 8329599 (C.D. Cal. Sep. 16, 2019).

[123]. Id. ¶ 97.

[124]. Id. ¶ 98.

[125]. Id. ¶ 109.

[126]. Id. ¶ 108.

[127]. Id. ¶ 11.

   [128]. See Novoa v. GEO Grp., Inc., No. 5:17-cv-02514-JGB-SHK, 2019 WL 7195331, at 20 (C.D. Cal. Nov. 26, 2019). But see Novoa v. GEO Grp., Inc., No. 5:17-cv-02514-JGB-SHK, 2021 WL 4913286, at 7 (C.D. Cal. Sep. 30, 2021) (dismissing GEO’s motion to decertify the facility-specific class but decertifying the nationwide class). Deeper analysis of the obstacles and opportunities for class certification is largely outside of the scope of this Note but merits further study.

   [129]. Booth, supra note 63, at 606.

   [130]. Betsy Swan, Private Prison Bosses Beg Taxpayers to Pay Human-Trafficking Lawsuit Bills, Daily Beast (July 17, 2019), https://www.thedailybeast.com/private-prison-bosses-beg-taxpayers-to-pay-human-trafficking-lawsuit-bills [https://perma.cc/JXJ5-8877].

[131]. Id.

[132]. See Stevens, supra note 26, at 402.

[133]. See generally Steven P. Croley, Civil Justice Reconsidered: Toward a Less Costly, More Accessible Litigation System (2017); Yotam Kaplan & Ittai Paldor, Social Justice and the Structure of the Litigation System, 101 N.C. L. Rev. 469 (2023).

   [134]. For a recent creative, contract-based solution to addressing violations of detainees’ labor rights in private immigration facilities, seeJames Sabia, Protecting Non-Citizens in Private Detention Facilities from Reprisal for Exercising Their Labor Rights: A Two-Step Procurement-Based, Pre-Award Solution, 53 Pub. Cont. L.J. 855 (2024).

   [135]. Since the facility-wide hunger strike at the Northwest Detention Center, detainees at facilities across the country have protested conditions, including work conditions. Freedom for Immigrants documented at least 1,600 detainees on hunger strike across twenty detention facilities between May 2015 and early 2020. SeeHunger Strikes, Freedom for Immigrants, https://www.freedomforimmigrants.org/hunger-strikes [https://perma.cc/9GSM-RVKA]. For an overview of various creative forms of resistance that “occur both within and beyond the walls of confinement,” see Savannah Kumar, Compelling Labor and Chilling Dissent: Creative Resistance to Coercive Uses of Solitary Confinement in Prisons and Immigration Detention Centers, 36 Harv. Blackletter L.J. 93, 112–16 (2020).

[136]. See, e.g., OSHA, supra note 116 (fining GEO for violating COVID-19 workplace safety regulations as applied to detained workers); National Labor Relations Board Files Complaint Against GEO Group for Retaliation Against Workers at Mesa Verde Detention Facility, Cal. Collaborative for Immigr. Just. (CCIJ) (Jan. 14, 2025), https://www.ccijustice.org/post/nlrb-complaint-against-geo [https://perma.cc/6X9W-KB64].

[137]. See Justice is Not for Sale Act, S. 2054, 114th Cong. (2015) (introducing a federal bill to end private immigration detention). State legislatures in New Jersey, Oregon, Illinois, Maryland, California, and Washington successfully passed legislation to end or limit immigration detention or private immigration detention. See N.J. Rev. Stat. §§ 30:4–8.15-16; Or. Rev. Stat. Ann. § 181A.829; 5 Ill. Comp. Stat. 805/15(g)(1); Md. Code Ann. Corr. Servs. § 1-102; Cal. Penal Code § 9501; Wash. Rev. Code Ann. § 70.395.030. But seeinfra note 140 and accompanying text. States like California and Washington have also passed laws empowering state agencies to regulate and investigate conditions of immigration detention. See Wash. Rev. Code Ann. § 70.395.040; Cal. Gov’t Code § 12532. But seeinfra note 140 and accompanying text.

[138]. See ACLU & Physicians for Hum. Rts., Behind Closed Doors: Abuse and Retaliation Against Hunger Strikers in U.S. Immigration Detention (2021), https://www.aclu.org/report/report-behind-closed-doors-abuse-retaliation-against-hunger-strikers-us-immigration-detention [https://perma.cc/5SA6-S6TD] (reporting on retaliation measures in response to hunger strikes, including force-feeding, solitary confinement, and retaliatory transfers); Robles, supra note 58 (reporting on retaliation against one of the lead detained worker organizers).

   [139]. The Cal/OSHA finding, imposing a $100,000 fine on the multibillion-dollar corporation for COVID-19-related violations from 2022, remains on appeal three years later. See OSHA, supra note 116. It took three years from the original filing until the National Labor Relations Board (NLRB) issued their official complaint. See Order Consolidating Cases, Consolidated Complaint, and Notice of Hearing, GEO Grp., Inc., Nos. 31-CA-309785, 31-CA-313988, 31-CA-322777 (NLRB Jan. 6, 2025). Months later, the newly inaugurated Trump administration dropped the case. Josh Eidelson, Trump NLRB Drops Case Over GEO Group Treatment of Detainees, Bloomberg News (Mar. 19, 2025), https://news.bloomberglaw.com/us-law-week/trump-nlrb-drops-case-over-geo-group-treatment-of-detainees [https://perma.cc/ZWD5-6VL7]. Courts have also struck down laws authorizing administrative actions against private immigration detention centers. See, e.g., United States v. California, 921 F.3d 865, 895 (9th Cir. 2019) (finding California’s law regulating immigration detention conditions unconstitutional).

   [140]. Both GEO and CoreCivic spend millions of dollars annually in political contributions and lobbying. See Thomas Ferraro, Raising a $1-a-Day Wage Seems Like a No-Brainer. Not to Congress, In These Times (July 21, 2022), https://inthesetimes.com/article/dollar-a-day-wages-immigration-detention-geo-corecivic-congress [https://perma.cc/WLP7-5UGB]; CoreCivic, Inc., OpenSecrets, https://www.opensecrets.org/orgs/corecivic-inc/summary?id=D000021940 [https://perma.cc/TU7U-3MB9]); see also Garfinkel, supra note 53, at 1303–05; Booth, supra note 63, at 585–87. Where states effectively passed legislation limiting immigration detention, GEO and CoreCivic have successfully challenged their constitutionality in court. See, e.g.,GEO Grp., Inc. v. Newsom, 50 F.4th 745, 763 (9th Cir. 2022) (finding California’s bill unconstitutional); GEO Grp., Inc. v. Inslee, 702 F. Supp. 3d 1043, 1051 (W.D. Wash. 2023) (finding Washington’s bill unenforceable); CoreCivic, Inc. v. Murphy, 690 F. Supp. 3d 467, 493 (D.N.J. 2023), aff’d sub nom.,CoreCivic, Inc. v. Governor of N.J., 145 F.4th 315 (3d Cir. 2025) (finding New Jersey’s bill unconstitutional). But see McHenry County v. Raoul, 44 F.4th 581, 594 (7th Cir. 2022) (upholding the constitutionality of Illinois’s bill).

   [141]. “For us—it’s not about winning a case. It’s just one of the many tools that we use. None of them by themselves are the solution.” Interview with Maru Mora Villalpando (Apr. 8, 2024).

[142]. See Barrientos v. CoreCivic, Inc., No. 4:18-cv-00070-CDL (M.D. Ga. Aug. 28, 2023) (ECF No. 342).

[143]. See infra Part II.A.

[144]. See infra Part III.

   [145]. 8 U.S.C. § 1324(a)(1)(A).

[146]. See Owino v. CoreCivic, Inc., No. 17-CV-1112 JLS (NLS), 2018 WL 2193644, at 20 (S.D. Cal. May 14, 2018); Novoa v. GEO Grp., Inc., No. 5:17-cv-02514-JGB-SHK, 2022 WL 2189626, at 26 (C.D. Cal. Jan. 25, 2022).

[147]. Owino, 2018 WL 2193644, at 20; see Novoa, 2022 WL 2189626, at 26.

[148]. See U.S. Gov’t Accountability Off., supra note 38; Kathryn Krawczyk, ICE is Rounding Up Green Card Holders, Week (June 28, 2018), https://theweek.com/speedreads/781984/ice-rounding-green-card-holders [https://perma.cc/SE52-M8XC].

[149]. See 12 U.S. Citizenship and Immigr. Servs., Lawful Permanent Resident Admission for Naturalization, pt. D, ch. 2 (2018), https://www.uscis.gov/policy-manual/volume-12-part-d-chapter-2#footnotelink-32 [https://perma.cc/E2D3-7PXX].

   [150]. Stevens, supra note 26, at 450.

[151]. See id.; 8 U.S.C. § 1324(a)(1)(A) (making it unlawful to hire a noncitizen “knowing” they are unauthorized).

[152].Washington v. GEO Grp., Inc., 283 F. Supp. 3d 967, 978 (W.D. Wash. 2017); Chen v. GEO Grp., Inc., 287 F. Supp. 3d 1158, 1167 (W.D. Wash. 2017).

[153]. See, e.g.,Patel v. Quality Inn S., 846 F.2d 700, 704 (11th Cir. 1988) (finding IRCA did not prevent awarding of damages to undocumented workers under FLSA); Madeira v. Affordable Hous. Found., Inc., 469 F.3d 219, 254 (2d Cir. 2006) (finding IRCA did not prevent awarding of damages to an undocumented worker for personal injury under state labor law). But see Hoffman Plastic Compounds, Inc. v. NLRB, 535 U.S. 137, 151 (2002) (finding IRCA prevented the NLRB from awarding backpay to an undocumented worker).

[154]. SeeMadeira, 469 F.3d at 245; Patel, 846 F.2d at 704 (explaining that if FLSA did not cover undocumented individuals, employers would have an incentive to hire and underpay undocumented workers).

[155]. See Owino v. CoreCivic, Inc., No. 17-CV-1112 JLS (NLS), 2018 WL 2193644, at *23 (S.D. Cal. May 14, 2018) (rejecting Defendant’s motion to dismiss arguing IRCA prohibits employing “unauthorized aliens” but acknowledging that the argument that IRCA could bar recovery may be “potentially relevant in future motions”).

[156].Novoa v. GEO Grp., Inc., No. 5:17-cv-2514-JGB-SHK, 2022 WL 2189626, at *26 (C.D. Cal. Jan. 25, 2022).

[157]. See infra Part III.C.1.

[158]. See U.S. Const. amend. XI;Cohens v. Virginia, 19 U.S. (6 Wheat.) 264, 411–12 (1821); United States v. Lee, 106 U.S. 196, 204 (1882); United States v. N.Y. Rayon Imp. Co., 329 U.S. 654, 663 (1947).

[159]. See Yearsley v. W.A. Ross Const. Co., 309 U.S. 18, 20–21 (1940); Campbell-Ewald Co. v. Gomez, 577 U.S. 153, 166–69 (2016).

   [160]. SeeCampbell-Ewald, 577 U.S. at 156, 160, 166.

   [161]. Geo Grp., Inc. v. Menocal, No. 24-758, 2026 WL 513536 (U.S. Feb. 25, 2026). The Court therefore concluded that GEO could not seek interlocutory appeal of the district court’s rejection of its “derivative sovereign immunity” defense. Id.

   [162]. See, e.g., Washington v. GEO Grp., Inc., No. 17-5806 RJB, 2019 WL 3565105, at 5 (W.D. Wash. Aug. 6, 2019); Nwauzor v. GEO Grp., Inc., No. 17-cv-05769-RJB, 2018 WL 4150909, at 1 (W.D. Wash. Aug. 6, 2018); Nwauzor v. GEO Grp., Inc., 127 F.4th 750, 770–71 (9th Cir. 2025); Novoa v. GEO Grp., Inc., No. 5:17-cv-02514-JGB-SHK, 2022 WL 2189626, at 20–21 (C.D. Cal. Jan. 25, 2022); Menocal v. GEO Grp., Inc., 635 F. Supp. 3d 1151, 1173–77 (D. Colo. 2022); Menocal v. GEO Grp., Inc., No. 22-1409, 2024 WL 4544184, at 1 (10th Cir. Oct. 22, 2024), aff’d, No. 24-758, 2026 WL 513536 (U.S. Feb. 25, 2026).

   [163]. See supra note 162 (citing cases).

[164]. See Brief for the United States as Amicus Curiae in Support of Neither Party at 8–11, Barrientos v. CoreCivic, Inc., 951 F.3d 1269 (11th Cir. 2020) (No. 18-15081).

   [165]. PBNDS, supra note 56, at 407. GEO admitted there was no “prohibition on paying more” and in fact did so at various facilities. See Plaintiffs-Appellees’ Answer Brief at 9, Menocal, 2024 WL 4544184.

   [166]. Plaintiffs-Appellees’ Answer Brief at 9, Menocal, 2024 WL 4544184; see also Nwauzor, 127 F.4th at 771 (“The contract sets a minimum compensation of $1 per day, but it does not forbid payments in excess of that amount.”); Menocal, 635 F. Supp. 3d at 1176 (“[T]here is no evidence that ICE prohibited GEO from compensating its workers more than $1.00 per day. Instead, the evidence suggests the participants’ compensation was left to GEO’s discretion.”).

[167].Boyle v. United Techs. Corp., 487 U.S. 500 (1988); see also Cunningham v. Gen. Dynamics Info. Tech., Inc., 888 F.3d 640, 646 n.4 (4th Cir. 2018) (“Boyle is inapposite to determining the applicability of derivative sovereign immunity.” (citing Boyle, 487 U.S. 500)); Elengold & Glater, supra note 115, at 987 n.79 (discussing their debatable decision to discuss Boyle alongside Yearsley as reinforcing their premise that these defenses are “interwoven, borrowed, and exploited to expand the scope of the sovereign shield, without those asserting the defense worrying too much about the original intent and meaning behind any one doctrine”).

   [168]. For further discussion on preemption, see infra Part III.C.1.

[169]. Boyle, 487 U.S. at 500 (internal quotations omitted).

[170]. See Menocal v. GEO Grp., Inc., 113 F. Supp. 3d 1125, 1134–35 (D. Colo. 2015); see also Menocal v. GEO Grp., Inc., 635 F. Supp. 3d 1151, 1172 (D. Colo. 2022), appeal dismissed, No. 22-1409, 2024 WL 4544184, at *11 n.3 (10th Cir. Oct. 22, 2024), aff’d, No. 24-758, 2026 WL 513536 (U.S. Feb. 25, 2026).

[171]. See Sinha, supra note 41 (Thirteenth Amendment); Booth, supra note 63 (Trafficking Victims Prevention Act); Garfinkel, supra note 53 (Fair Labor Standards Act); Stevens, supra note 26, at 398–409 (Fair Labor Standards Act).

[172]. See David S. Rubenstein, Supremacy, Inc., 67 UCLA L. Rev. 1130, 1142 (2020) (explaining that federal contractors are generally immune from constitutional strictures under the state action doctrine (citing Gillian E. Metzger, Privatization as Delegation, 103 Colum. L. Rev. 1367, 1369–70, 1403–06 (2003))); see also Lillian BeVier & John Harrison, The State Action Principle and Its Critics, 96 Va. L. Rev. 1767, 1786 (2010) (“Constitutional rules are almost all addressed to the government.”).

[173].Richardson v. McKnight, 521 U.S. 399 (1997) (citing 42 U.S.C. § 1983 and holding that private prison operators faced with § 1983 claims could not claim qualified immunity).

   [174]. Federal agents may be held liable for constitutional violations under Bivens v. Six Unknown Named Agents of Federal Bureau of Narcotics, 403 U.S. 388, 389 (1971). But neither § 1983 claims nor Bivens claims have been successful against private federal actors. See, e.g.,Doe v. United States, 831 F.3d 309, 316 (5th Cir. 2016) (finding a detainee could not bring a § 1983 claim against CoreCivic because it was performing a federal function); Minneci v. Pollard, 565 U.S. 118 (2012) (finding a prisoner could not bring a Bivens action against the private operator of a federal halfway house).

   [175]. U.S. Const. amend. XIII, § 1.

   [176]. U.S. Const. amend. XIII, § 2.

   [177]. George Rutherglen, State Action, Private Action, and the Thirteenth Amendment, 94 Va. L. Rev. 1367, 1371 (2008).

[178].The Slaughter-House Cases, 83 U.S. 36, 50 (1872).

[179]. Id.; see also Bailey v. Alabama, 219 U.S. 219 (1911) (finding that compelling a person to work to pay off a debt is unconstitutional under the Thirteenth Amendment).

[180]. The Slaughter-House Cases, 83 U.S. at 37 (reasoning that the Thirteenth Amendment’s protection extends to labor by Mexican and Chinese immigrants); see also Sinha, supra note 41, at 41–42; Safstrom, supra note 52, at 216.

[181].Channer v. Hall, 112 F.3d 214, 217 n.5 (5th Cir. 1997) (quoting The Civil Rights Cases, 109 U.S. 3, 20 (1883)). Contrast Turner v. Unification Church, 473 F. Supp. 367, 371 (D.R.I. 1978), aff’d, 602 F.2d 458 (1st Cir. 1979) (refusing to recognize an implied right of action for private conduct under the Thirteenth Amendment), with Martinez v. Calimlim, 651 F. Supp. 2d 852, 864 (E.D. Wis. 2009) (recognizing a plaintiff’s direct cause of action against a private actor under the Thirteenth Amendment).

   [182]. Rutherglen, supra note 177, at 1367 (citing The Civil Rights Cases, 109 U.S. 3, 20 (1883)).

[183]. See generally Sinha, supra note 41; Safstrom, supra note 52; Cambridge, supra note 98. See also Lisa Knox, Hamid Yazdan Panah & Serafin Andrade Lopez, Strategic Lessons From Abolitionist Labor Struggle in Immigration Detention, LPE Project (May 30, 2023), https://lpeproject.org/blog/strategic-lessons-from-abolitionist-labor-struggle-in-immigration-detention/ [https://perma.cc/BW2Z-U28F] (blog post by an immigration attorney, legal advocate, and formerly detained organizer pointing out that the Exceptions Clause in the Thirteenth Amendment does not apply in immigration detention).

[184]. See, e.g., Channer, 112 F.3d. 214.

   [185]. U.S. Const. amend. XIII, § 1 (emphasis added).

[186]. See Channer, 112 F.3d at 215–16, 217 n.4 (finding the Thirteenth Amendment’s Exception Clause did not apply to Channer while he was in immigration detention, despite Channer being transferred there from prison where he served a sentence for an armed robbery conviction).

   [187]. Garfinkel, supra note 53, at 1314.

[188]. See Fong Yue Ting v. United States, 149 U.S. 698, 730 (1893);Wong Wing v. United States, 163 U.S. 228, 235–38 (1896).

   [189]. 163 U.S. at 238.

[190]. See supra note 57 and accompanying text.

   [191]. Lauren Kares, Unlucky Thirteenth: A Constitutional Amendment in Search of a Doctrine, 80 Corn. L. Rev. 372, 372 (1995) (citing Steirer v. Bethlehem Area Sch. Dist., 789 F. Supp. 1337 (E.D. Pa. 1992), aff’d,987 F.2d 989 (3d Cir. 1992), cert. denied, 510 U.S. 824 (1993)).

[192]. See Kares, supra note 191, at 374–75.

[193]. Id. at 380.

[194]. See Channer v. Hall, 112 F.3d 214, 218 (5th Cir. 1997) (“Assuming without deciding that segregated detention is a form of legal punishment, we find that there would be at least some evidence that Channer’s services were compelled by the use of legal coercion.”); id. at 218 n.7 (“Channer’s services were not necessarily voluntary merely because he was paid for his labors. While receiving payment for services is relevant to determining voluntariness, we cannot resolve that factual question against Channer on this record.”).

[195]. Id. at 218–19.

[196]. See Brianne Power, When Is Forced Labor Consistent with the General Basic System of Free Labor? An Analysis of Historical Judge-Made Exceptions to the Thirteenth Amendment, onlabor (Apr. 26, 2018), https://onlabor.org/when-is-forced-labor-consistent-with-the-general-basic-system-of-free-labor-an-analysis-of-historical-judge-made-exceptions-to-the-thirteenth-amendment [https://perma.cc/C7VY-6E2K].

[197].Butler v. Perry, 240 U.S. 328, 333 (1916); see also United States v. Kozminski, 487 U.S. 931, 943–44 (1988) (reasoning the Thirteenth Amendment “does not prevent the State or Federal Governments from compelling their citizens, by threat of criminal sanction, to perform certain civic duties”).

[198]. See Bayh v. Sonnenburg, 573 N.E.2d 398, 411 (Ind. 1991) (finding patients who cooked, cleaned, and did laundry while in a mental hospital fell within the civic duty exception to the Thirteenth Amendment). But see McGarry v. Pallito, 687 F.3d 505, 514 (2d Cir. 2012) (finding that compelling a pretrial detainee to work in a laundry up to fourteen hours a day for three days a week cannot be a personally related housekeeping chore exempt from the Thirteenth Amendment).

   [199]. Both companies raised the civic duty exception as an affirmative defense against claims from detained workers under the TVPA, a federal statute promulgated by Congress under both Section Two of the Thirteenth Amendment and the Commerce Clause. See Sherman, supra note 98, at 1217;discussion infra Part II.B.

[200]. See Sherman, supra note 98, at 1217.

[201]. See Menocal v. GEO Grp., Inc., 113 F. Supp. 3d 1125, 1133 (D. Colo. 2015); Owino v. CoreCivic, Inc., No. 17-CV-1112 JLS (NLS), 2018 WL 2193644, at 15–20, 15 n.5. (S.D. Cal. May 14, 2018).

[202]. See Turner v. Unification Church, 473 F. Supp. 367, 374 (D.R.I. 1978), aff’d, 602 F.2d 458 (1st Cir. 1979). But see Martinez v. Calimlim, 651 F. Supp. 2d 852, 864 (E.D. Wis. 2009).

[203]. See generally Kares, supra note 191.

   [204]. Victims of Trafficking and Violence Protection Act of 2000, Pub. L. No. 106-386, 114 Stat. 1464.

[205]. See, e.g., id. § 102(b)(12) (“Trafficking in persons substantially affects interstate and foreign commerce.”); id. § 112(a)(2) (“Whoever knowingly . . . in or affecting interstate commerce . . . .”); id. § 102(b)(1) (“Trafficking in persons is a modern form of slavery, and it is the largest manifestation of slavery today.”); see also Sherman, supra note 98, at 1215; Heather Odell, Accountable to None? Challenging Sovereign Immunity Through the Trafficking Victims Protection Act, 63 B.C. L. Rev. 1517, 1549–50 (2022).

   [206]. Victims of Trafficking and Violence Protection Act of 2000, Pub. L. No. 106-386, § 112(a)(2), 114 Stat. 1464.

[207]. See Booth, supra note 63, at 588 (noting that a statutory strategy is more promising for detained workers than a constitutional strategy in light of Thirteenth Amendment precedent); Sherman, supra note 98, at 1221.

   [208]. In 2003, Congress amended the TVPA to provide a civil remedy for victims. Trafficking Victims Protection Reauthorization Act of 2003, Pub. L. No. 108-193, § 4(a)(4)(A), 117 Stat. 2875, 2878 (codified at 18 U.S.C. § 1595).

   [209]. U.S. Const. amend. XIII, § 2 (“Congress shall have power to enforce this article by appropriate legislation.”); Safstrom, supra note 52, at 218–19; Sherman, supra note 98, at 1218.

   [210]. Kares, supra note 191, at 376.

[211].Jones v. Alfred Mayer Co., 392 U.S. 409, 439 (1968) (“Whether or not the Amendment itself did any more than [abolish slavery]—a question not involved in this case—it is at least clear that the Enabling Clause of that Amendment empowered Congress to do much more. For that clause clothed ‘Congress with power to pass all laws necessary and proper for abolishing all badges and incidents of slavery in the United States.’” (quoting The Civil Rights Cases, 109 U.S. 3, 20 (1883))); see also Kares, supra note 191, at 377–79.

[212]. See, e.g.,Menocal v. GEO Grp., Inc., 113 F. Supp. 3d 1125, 1132 (D. Colo. 2015) (“[GEO] argue[s] that the TVPA is inapplicable because its purpose was to prevent human trafficking, and cases exclusively apply the TVPA to trafficking persons for labor and/or sex.”); Novoa v. GEO Grp., Inc., No. 5:17-cv-02514-JGB-SHK, 2018 WL 3343494, at *12 (C.D. Cal. June 21, 2018) (“[GEO] . . . argues applying the TVPA here would go beyond the intent and purpose of the statute, which was to prosecute and deter the trafficking of persons over geographic spaces.”); Barrientos v. CoreCivic, Inc., 332 F. Supp. 3d 1305, 1310 (M.D. Ga. 2018), aff’d, 951 F.3d 1269 (11th Cir. 2020) (“CoreCivic contends that Plaintiffs fail to state a claim under the TVPA because Congress did not intend the statute to apply to lawfully held detainees.”).

[213]. See, e.g., Barrientos, 332 F. Supp. 3d at 1310(“CoreCivic argues that the TVPA is intended to apply narrowly to forced labor in the human trafficking context and that applying it to detainee work programs is ‘absurd’ and contrary to the intentions of Congress.”); see Owino v. CoreCivic, Inc., No. 17-CV-1112 JLS (NLS), 2018 WL 2193644, at *3 (S.D. Cal. May 14, 2018) (“[CoreCivic] contends that applying a forced labor statute to lawfully-detained civil immigration detainees would be both extreme and absurd.”). The “absurdity doctrine” is a canon of statutory interpretation that allows for the rare departure from the ordinary meaning of the plain text if no reasonable person could intend that result. See Antonin Scalia & Bryan A. Garner, Reading Law: The Interpretation of Legal Texts (2012).

   [214]. Various academics have extensively detailed how both the plain text and the legislative intent behind 18 U.S.C. § 1589, the TVPA’s forced labor statute, confirm its application to detained workers. See, e.g.,Booth, supra note 63, at 593–97; Safstrom, supra note 52, at 220–28.

[215].Barrientos v. CoreCivic, Inc., 951 F.3d 1269, 1278 (11th Cir. 2020).

[216].Gonzalez v. CoreCivic, Inc., 986 F.3d 536, 539 (5th Cir. 2021).

[217]. See supra Part I.E.1 (Table 1); see also Hernandez Gomez v. GEO Grp., Inc., No. 1:22 Civ. 00868-ADA-CDB, 2022 WL 19266228 (E.D. Cal. Dec. 23, 2022); Ruderman v. McHenry County, No. 3:22-CV-50115, 2023 WL 130496 (N.D. Ill. Jan. 9, 2023); Yeend v. Akima Glob. Servs., LLC, 347 F.R.D. 405 (N.D.N.Y. 2024).

[218].Menocal v. GEO Grp., Inc., 635 F. Supp. 3d 1151, 1198 (D. Colo. 2022), appeal dismissed, No. 22-1409, 2024 WL 4544184, at *4 (10th Cir. Oct. 22, 2024), aff’d, No. 24-758, 2026 WL 513536 (U.S. Feb. 25, 2026).

   [219]. 29 U.S.C. §§ 201–219.

   [220]. Id. § 206(a)(1)(C). Notably, the FLSA applies to immigrants regardless of work authorization status. See Jerusalem Cafe, LLC v. Lucas, 721 F.3d 927, 937 (8th Cir. 2013); Lamonica v. Safe Hurricane Shutters, Inc., 711 F.3d 1299, 1308 (11th Cir. 2013).

[221].Alvarado Guevara v. INS, 902 F.2d 394 (5th Cir. 1990); Ndambi v. CoreCivic, Inc., 990 F.3d 369 (4th Cir. 2021).

[222]. See generally Petition for Rehearing En Banc at 2, 4–9, Ndambi, 990 F.3d 369 (No. 19-2207); Garfinkel, supra note 53; Stevens, supra note 26; Comment, Employment Law—Fair Labor Standards Act—Fourth Circuit Holds That Detained Immigrant Workers Are Not “Employees” Under the Fair Labor Standards Act, 135 Harv. L. Rev. 1512 (2022) [hereinafter Harvard Ndambi Comment].

[223]. See Rutherford Food Corp. v. McComb, 331 U.S. 722, 728 (1947); Goldberg v. Whitaker House Co-op., Inc., 366 U.S. 28, 33 (1961).

[224]. See Garfinkel, supra note 53, at 1324–25 (applying the economic realities test to detained workers and finding all but one of the factors suggest an employment relationship).

   [225]. This aligns with a uniform approach by courts to reject the application of the economic realities test to prison workers. Seeid. at 1322 (citing Matthew J. Lang, The Search for a Workable Standard for When Fair Labor Standards Act Coverage Should Be Extended to Prison Workers, 5 U. Pa. J. Lab. & Emp. L. 191, 204 (2003)).

[226]. Alvarado Guevara, 902 F.2d at 396 (finding the detained immigrant laborers similar to prison laborers “in that they have been incarcerated and are under the direct supervision and control of a governmental entity” and therefore “should not be protected under the FLSA”).

[227]. Ndambi, 990 F.3d at 371 (“The economic reality of the Plaintiffs’ situation . . . is almost identical to that of a prison inmate and does not share commonality with that of a traditional employer-employee relationship.” (quoting Alvarado Guevara, 902 F.2d at 396)).

[228]. Seesupra Part II.A; Garfinkel, supra note 53, at 1311, 1315 (noting that decisions like Alvarado Guevara, that deny FLSA protections in immigration detention, rely on an “unsubstantiated comparison” with criminal detention and therefore “should not have determinative authority over cases arising from the Voluntary Work Program”).

[229]. See, e.g.,Gilbreath v. Cutter Biological, Inc., 931 F.2d 1320, 1325 (9th Cir. 1991) (denying FLSA protections to prison workers in light of the “exceptions carved out by the courts” from involuntary servitude for those duly tried and imprisoned for a crime); see also Owino v. CoreCivic, Inc., No. 17-CV-1112 JLS (NLS), 2018 WL 2193644, at *23–25 (S.D. Cal. May 14, 2018) (rejecting analogies between civil immigrant detainees and prisoners for the purposes of minimum wage protections).

[230]. Ndambi, 990 F.3d at 372; Alvarado Guevara, 902 F.2d at 396.

[231]. See Petition for Rehearing En Banc at 2, 4–9, Ndambi, 990 F.3d 369 (No. 19-2207) (arguing the panel’s opinion violated “well-established principles of statutory interpretation”).

[232]. See, e.g., 29 U.S.C. § 203(e)(3)–(4).

[233]. See, e.g., Hale v. Arizona, 993 F.2d 1387, 1392 (9th Cir. 1993) (“[W]e cannot agree that the FLSA categorically excludes all labor of any inmate.”); Watson v. Graves, 909 F.2d 1549, 1556 (5th Cir. 1990) (finding FLSA coverage applied to prisoners performing work for a third party).

[234]. Ndambi, 990 F.3d at 372 (citing Bennett v. Frank, 395 F.3d 409, 410 (7th Cir. 2005)).

   [235]. 29 U.S.C. § 202(a); seeBennett, 395 F.3d at 410 (“People are not imprisoned for the purpose of enabling them to earn a living.”).

[236]. Ndambi, 990 F.3d at 372–73.

[237].Tony & Susan Alamo Found. v. Sec’y of Lab., 471 U.S. 290, 293–94 (1985) (finding that volunteers who received food, shelter, and clothing from a nonprofit foundation were covered under the FLSA).

   [238]. Garfinkel, supra note 53, at 1319 (citing Tony & Susan Alamo Found., 471 U.S. at 293–94).

   [239]. 29 C.F.R. § 531.30 (2012).

   [240]. Garfinkel, supra note 53, at 1320–21; see also Barrientos v. CoreCivic, Inc., No. 4:18-cv-00070-CDL, 2021 WL 6931810, at *4 (M.D. Ga. Mar. 10, 2021) (dismissing CoreCivic’s unjust enrichment counterclaim against plaintiffs and finding that CoreCivic’s “audacity in claiming that it should be compensated for providing basic necessities to detainees who it contracted to detain is astonishing”).

[241].Ndambi v. CoreCivic, Inc., 990 F.3d 369, 373 (4th Cir. 2021).

[242]. See 29 U.S.C. § 202(a);Petition for Rehearing En Banc at *13, Ndambi, 990 F.3d 369 (No. 19-2207) (arguing that “[w]hether CoreCivic met Appellants’ basic needs is a question of fact not suitable for disposition on a Rule 12(b)(6) motion”).

[243]. Ndambi, 990 F.3d at 374 (concluding that “the custodial detention context is inconsistent with the free labor market envisioned by the FLSA”); Alvarado Guevara v. INS, 902 F.2d 394, 396 (5th Cir. 1990) (reasoning that detained workers are “removed from American industry” and therefore “not within the group that Congress sought to protect in enacting the FLSA”).

   [244]. Stevens, supra note 26, at 441–42; see also Complaint for Declaratory and Injunctive Relief and Damages ¶ 31, Barrientos, v. CoreCivic, Inc., No. 4:18-cv-00070-CDL, 2020 WL 7021904 (M.D. Ga. Apr. 17, 2018) (“CoreCivic occasionally increases the wage rate it pays to kitchen workers to up to $8 per day when it needs workers to work twelve hours or more per day.”).

   [245]. Harvard Ndambi Comment, supra note 222, at 1517.

[246]. Id.; see 29 U.S.C. § 202(a)(3) (“The Congress finds that . . . labor conditions detrimental to the maintenance of the minimum standard of living . . . constitutes an unfair method of competition in commerce.”).

   [247]. Harvard Ndambi Comment, supra note 222, at 1516 (“[The Ndambi court’s] parsimonious reading of the statute stands at odds with the long-standing judicial credo that a broad reading of the FLSA’s provisions is necessary to effectuate the Act’s principal aims” (citing Mitchell v. Lublin, McGaughy & Assocs., 358 U.S. 207, 211 (1959)).).

[248].Ndambi v. CoreCivic, Inc., 990 F.3d 369, 372 (4th Cir. 2021) (quoting Harker v. State Use Indus., 990 F.2d 131, 133 (4th Cir. 1993)).

[249].Nationwide Mut. Ins. Co. v. Darden, 503 U.S. 318, 326 (1992).

   [250]. Complaint, Nwauzor v. GEO Grp., Inc., No. 17-cv-05769-RJB, 2017 WL 11646687 (W.D. Wash. Sep. 26, 2017); Complaint, Washington v. GEO Grp., Inc., 283 F. Supp. 3d 967 (W.D. Wash. 2017).

   [251]. See, e.g., Class Action Complaint for Unpaid Wages and Forced Labor, Menocal, 2014 WL 5389925; Second Amended Complaint, Hernandez Gomez, 2022 WL 19266228; Third Amended Complaint, Novoa, 2019 WL 8329599.

   [252]. Operation Lone Star in Texas offers a stark example of the dangers of state infringement in federal immigration. See Acacia Coronado & Paul J. Weber, Justice Department Sues Texas Over Law That Would Let Police Arrest Migrants Who Enter US Illegally, AP News (Jan. 3, 2024), https://apnews.com/article/texas-immigration-border-justice-department-ad3814ef6c6160d62fd899d7570ffc0b [https://perma.cc/56SM-HH24].

   [253]. Rubenstein & Gulasekaram, supra note 115, at 227.

   [254]. Interview with Maru Mora Villalpando (Apr. 8, 2024).

   [255]. Rubenstein, supra note 172, at 1136.

[256]. See Stevens, supra note 26, at 402; see alsosupra Part I.B (describing the for-profit immigration detention model); infra note 290 and accompanying text (countering claims that private detention centers would benefit economically from paying full labor costs).

   [257]. Alabama, Louisiana, Mississippi, South Carolina, and Tennessee. State Minimum Wages, Nat’l Conf. of State Legislatures (Jan. 1, 2025), https://www.ncsl.org/labor-and-employment/state-minimum-wages [https://perma.cc/HVV2-WBST].

[258]. Id.

   [259]. N.M. Stat. Ann. §§ 50-4-21–50-4-27(2020).

[260].Ndambi v. CoreCivic, Inc., 990 F.3d 369, 371 n.1 (4th Cir. 2021).

[261]. Id. at 371.

   [262]. Cal. Lab. Code §§ 1194, 1197, 1197.1.

[263]. See Martinez v. Combs, 231 P.3d 259, 279–81 (Cal. 2010).

[264].Owino v. CoreCivic, Inc., No. 17-CV-1112 JLS (NLS), 2018 WL 2193644, at 25–26 (S.D. Cal. May 14, 2018) (rejecting the FLSA economic realities test and denying CoreCivic’s motion to dismiss detained workers’ California minimum wage claims); Novoa v. GEO Grp., Inc., No. 5:17-cv-02514-JGB-SHK, 2018 WL 3343494, at 6–7, 9 (C.D. Cal. June 21, 2018) (rejecting interpreting CMWL in accordance with the FLSA and denying GEO’s motion to dismiss detained workers’ California minimum wage claims); id. at 15 (granting detained workers’ motion for summary judgment on CMWL claims).

[265]. See Cal. Lab. Code § 2775(b)(1); see also Lynn Rhinehart, Celine McNicholas, Margaret Poydock & Ihna Mangundayao, Econ. Pol’y Inst., Misclassification, the ABC Test, and Employee Status 8 (2021),https://www.epi.org/publication/misclassification-the-abc-test-and-employee-status-the-california-experience-and-its-relevance-to-current-policy-debates/ [https://perma.cc/72UP-PDHU] (identifying the following states as adopting the ABC test for wage and hour laws without industry limits: California, Connecticut, Massachusetts, Nebraska, New Jersey, and Vermont).

[266]. See Guy Davidov & Pinna Alon-Shenker, The ABC Test: A New Model for Employment Status Determination?, 51 Indus. L.J. 235, 235 (2022); Hanoch Dagan & Catherine Fisk, Independent Contractors and the ABCs of Contract Law, 66 Ariz. L. Rev. 607, 609 (2024).

[267]. See Eric Markovits, Easy as ABC: Why the ABC Test Should Be Adopted as the Sole Test of Employee–Independent Contractor Status, Cardozo L. Rev. de•novo, 224, 254 (2020) (noting that “the ABC test could lead to many more workers receiving the benefits and protections they lawfully deserve”); Michael T. Alario, “We Want All Workers To Have The Right To Bargain Collectively”: How The ABCs Can Equalize The Gig Economy, 64 B.C. L. Rev. 1203, 1236–39 (2023).

[268]. Novoa, 2022 WL 2189626, at *10–11.

   [269]. Washington considered adopting the ABC test across most of the State’s labor laws in 2019 through HR 1515 and SB 5513 but neither passed. See Rebecca Smith, Washington State Considers ABC Test for Employee Status, Nat’l Emp. L. Project (Jan. 28, 2019), https://www.nelp.org/washington-state-considers-abc-test-employee-status/ [https://perma.cc/47GL-QPTQ].

[270].Nwauzor v. GEO Grp., Inc., 540 P.3d 93, 97, 104 (Wash. 2023).

[271]. Id. at 103–04 (“While we have stated that federal authority under the FLSA may provide helpful guidance in interpreting the MWA, we have also recognized that these two statutory schemes ‘are not identical and we are not bound by such authority.’” (citation omitted)).

   [272]. Wash. Rev. Code § 49.46.010(4) (defining employee as “any individual employed by an employer but shall not include” those individuals described in sections (a)–(p)); see also Rocha v. King County, 460 P.3d 624, 630 (Wash. 2020) (en banc) (“Instead of being primarily defined by employments included, the MWA carves out from the definition of ‘employee’ more narrow provisions that operate as exemptions.”).

[273]. Nwauzor, 540 P.3d at 101.

[274]. Id. at 98–102, 105.

[275].Menocal v. GEO Grp., Inc., 113 F. Supp. 3d 1125, 1129 (D. Colo. 2015).

[276]. Id. at 1129 (“The CMWO applies to employers and employees in four industries: (1) Retail and Service; (2) Commercial Support Service; (3) Food and Beverage; and (4) Health and Medical.” (citing Colo. Code Regs. § 1103-1:1)).

[277]. Menocal, 113 F. Supp. 3d at 1129; seesupra Part II.C.

   [278]. Annie Hollister, Litigating ICE’s “Voluntary Work Program,” onlabor (Apr. 10, 2022), https://onlabor.org/litigating-ices-voluntary-work-program/ [https://perma.cc/4KV5-WU7Z]; see also Peter Birks, Unjust Enrichment 1 (2005) (“[E]ven at the beginning of the 21st century unjust enrichment is still unfamiliar to most common lawyers.”).

[279]. See The Intellectual History of Unjust Enrichment, 133 Harv. L. Rev. 2077, 2078–79 (2020); Siyi Lin, A Tale of Transplantation: The Historical Evolution of the Law of Unjust Enrichment in China, in Rethinking Unjust Enrichment: History, Sociology, Doctrine, and Theory (Warren Swain & Sagi Peari eds., 2023).

[280]. See The Intellectual History of Unjust Enrichment, supra note 279, at 2086–89.

   [281]. Restatement (Third) of Restitution and Unjust Enrichment § 1 cmt. b (A.L.I. 2011).

   [282]. Hollister, supra note 278.

[283]. See, e.g., Complaint ¶¶ 8.1–8.9, Gonzalez v.CoreCivic, Inc., No. 1:18 Civ.169, 2018 WL 1043718 (W.D. Tex. Feb. 22, 2018); Second Amended Complaint ¶¶ 280–86, Hernandez Gomez v. GEO Grp., Inc.,No. 1:22 Civ. 00868-ADA-CDB, 2022 WL 19266228 (E.D. Cal. Dec. 23, 2022); Amended Complaint for Declaratory and Injunctive Relief and Damages ¶¶ 128–33, Barrientos v. CoreCivic, Inc., No. 4:18-cv-00070-CDL, 2020 WL 7021904 (M.D. Ga. Apr. 17, 2018); Third Amended Complaint ¶¶ 203–11, Novoa v. GEO Grp., Inc., No. 5:17-cv-02514-JGB-SHK, 2019 WL 8329599 (C.D. Cal. Sep. 16, 2019); GEO’s Notice of Removal to Federal Court at 7, Washington v. GEO Grp., Inc., No. 3:17-cv-05806, 2017 WL 11646686 (W.D. Wash. Oct. 9, 2017); see alsosupra Part I.E.1 (Table 1).

[284]. Cause of Action for Unjust Enrichment in California, Thomson Reuters (Dec. 21, 2022), https://legal.thomsonreuters.com/blog/cause-of-action-for-unjust-enrichment-in-california/#source-1[https://perma.cc/74RY-TQFJ] (“Unjust enrichment is acknowledged and put into practice in 49 other states and by half of the Californian courts.”).

   [285]. Hollister, supra note 278.

   [286]. Ferraro, supra note 140.

[287].Menocal v. GEO Grp., Inc., 882 F.3d 905, 923 (10th Cir. 2018); Washington v. GEO Grp., Inc., 283 F. Supp. 3d 967, 980–81 (W.D. Wash. 2017); Barrientos v. CoreCivic, Inc., 332 F. Supp. 3d 1305, 1313 (M.D. Ga. 2018).

   [288]. Defendant’s Motion for Summary Judgment at 16, Menocal v. GEO Grp., Inc., 635 F. Supp. 3d 1151 (D. Colo. 2015) (No. 1:14 Civ. 02887-JLK-MEH) (citing evidence that GEO receives a 15 percent markup fee on their total costs, including labor); Novoa v. GEO Grp., Inc., No. 5:17-cv-02514-JGB-SHK, 2022 WL 2189626, at *16 (C.D. Cal. Jan. 25, 2022) (quoting GEO’s CEO stating that increased labor would increase their costs and therefore their profits, which are calculated as a percentage of their overhead).

   [289]. Defendant’s Motion for Summary Judgment at 36, Menocal, 635 F. Supp. 3d 1151 (No. 1:14 Civ. 02887-JLK-MEH); Defendant The GEO Group, Inc.’s Notice of Motion and Motion for Summary Judgment at 45, Novoa v. GEO Grp., Inc., No. 5:17-cv-02514-JGB-SHK, 2022 WL 2189626 (C.D. Cal. Jan. 25, 2022).

[290]. Menocal, 635 F. Supp. at 1197; Novoa, 2022 WL 2189626, at *16.

[291]. Barrientos, 332 F. Supp. at 1313, aff’d, 951 F.3d 1269 (11th Cir. 2020).

[292].Menocal v. GEO Grp., Inc., 113 F. Supp. 3d 1125, 1133 (D. Colo. 2015)

[293]. See, e.g.,Gonzalez v. CoreCivic, Inc., No. 1:18-CV-169-LY, 2019 WL 2572540, at *3 (W.D. Tex. Mar. 1, 2019), aff’d, 986 F.3d 536 (5th Cir. 2021) (“Having determined that Gonzalez’s claim under the TVPA survives the motion to dismiss, the court concludes that the derivative claims [of negligence and unjust enrichment] also survive at this time.”).

[294].Ndambi v. CoreCivic, Inc., No. CV RDB-18-3521, 2019 WL 4735428, at *2 (D. Md. Sep. 27, 2019), aff’d, 990 F.3d 369 (4th Cir. 2021).

[295]. Id.

[296].Ndambi v. CoreCivic, Inc., 990 F.3d 369, 371 n.1 (4th Cir. 2021) (“The parties agree that the NMMWA should be interpreted in accordance with the FLSA and that the appellants’ unjust enrichment claim depends on the success of their FLSA claim. Thus, our discussion is limited to interpreting the FLSA.”).

[297]. Menocal, 113 F. Supp. 3d at 1133.

[298]. See, e.g., id.; Novoa v. GEO Grp., Inc., No. 5:17-cv-02514-JGB-SHK, 2018 WL 3343494, at 14 (C.D. Cal. June 21, 2018); Owino v. CoreCivic, Inc., No. 17-CV-1112 JLS (NLS), 2018 WL 2193644, at 27–28 (S.D. Cal. May 14, 2018).

[299]. See Prac. L. Com. Litig., Asserting Quasi-Contract Claims, Practical Law Practice Note w-014-9227 (Westlaw Practical Law). While some detention centers have detained workers sign agreements, courts have dismissed these agreements as unconscionable and unenforceable. Menocal v. GEO Grp., Inc., 635 F. Supp. 3d 1151, 1198 (D. Colo. 2022), appeal dismissed, No. 22-1409, 2024 WL 4544184 (10th Cir. Oct. 22, 2024), aff’d, No. 24-758, 2026 WL 513536 (U.S. Feb. 25, 2026).

[300]. See Menocal, 113 F. Supp. 3d at 1133; Owino, 2018 WL 2193644, at *27 (citing Fed. R. Civ. P. 8).

[301]. See Owino, 2018 WL 2193644, at *28.

   [302]. Washington v. GEO Grp., Inc., No. 3:17-cv-05806, 2017 WL 11646686, at *1 (W.D. Wash. Oct. 9, 2017).

   [303]. Nwauzor v. GEO Grp., Inc., No. 17-cv-05769-RJB,2017 WL 11646687, ¶ 1.2 (W.D. Wash. Aug. 15, 2017).

[304].Nwauzor v. GEO Grp., Inc., 540 P.3d 93, 104 (Wash. 2023).

[305]. See, e.g.,Gonzalez v. CoreCivic, Inc., No. 1:18-CV-169-LY, 2019 WL 2572540, at *3 (W.D. Tex. Mar. 1, 2019), aff’d, 986 F.3d 536 (5th Cir. 2021).

[306].Menocal v. GEO Grp., Inc., 635 F. Supp. 3d 1151, 1198 (D. Colo. 2022), appeal dismissed, No. 22-1409, 2024 WL 4544184 (10th Cir. Oct. 22, 2024), aff’d, No. 24-758, 2026 WL 513536 (U.S. Feb. 25, 2026).

[307].Washington v. GEO Grp., Inc., No. 3:17-cv-05806-RJB, 2018 WL 1083862, at *2 (W.D. Wash. Feb. 28, 2018); Chen v. GEO Grp., Inc., 297 F. Supp. 3d 1130, 1132–33 (W.D. Wash. 2018).

   [308]. Chen, 297 F. Supp. at 1133.

[309].Barrientos v. CoreCivic, Inc., No. 4:18-CV-00070-CDL, 2021 WL 6931810, at *1 (M.D. Ga. Mar. 10, 2021); Chen, 297 F. Supp. 3d at 1132.

[310]. Chen, 297 F. Supp. 3d at 1133; Nwauzor v. GEO Grp., Inc., No. 17-cv-05769-RJB, 2020 WL 1689728, at *9 (W.D. Wash. Apr. 7, 2020).

[311]. Washington, 2018 WL 1083862, at *2.

[312]. Nwauzor, 2020 WL 1689728, at *9.

[313]. Barrientos, 2021 WL 6931810, at *4.

[314]. The GEO Group Comments on Unfavorable Jury Verdict and Judgments in the U.S. District Court for the Western District of Washington, GEO (Nov. 4, 2021), https://investors.geogroup.com/news-releases/news-release-details/geo-group-comments-unfavorable-jury-verdict-and-judgments-us [https://perma.cc/46UQ-9PWE].

   [315]. The Ninth Circuit recently ruled in favor of detained workers. See Nwauzor v. GEO Grp., Inc., 127 F.4th 750 (9th Cir. 2025). GEO is now seeking to appeal to the Supreme Court. Petition for Writ of Certiorari, GEO Grp., Inc., v. Nwauzor, No. 25A464 (U.S. Jan. 9, 2026).

[316]. See Brief for the United States as Amicus Curiae in Support of Appellant, Washington v. GEO Grp., Inc., 127 F.4th 783 (9th Cir. 2025) (Nos. 21-36024, 21-36025). Notably, the Trump DOJ also filed a brief in support of GEO’s petition for rehearing. Brief for the United States as Amicus Curiae in Support of Petition for Rehearing, Nwauzor, 127 F.4th 750 (Nos. 21-36024, 21-36025).

[317].Brief for the United States as Amicus Curiae in Support of Appellant at 11, Washington, 127 F.4th 783 (Nos. 21-36024, 21-36025).

   [318]. U.S. Const. art. VI, cl. 2.

   [319]. Rubenstein, supra note 172, at 1137.

[320]. Id. at 1137, 1137 n.28 (citing Va. Uranium, Inc. v. Warren, 587 U.S. 761, 767 (2019) (explaining “[t]he Supremacy Clause supplies a rule of priority” as between federal and state law)); Hughes v. Talen Energy Mktg., LLC, 578 U.S. 150, 162 (2016) (“Put simply, federal law preempts contrary state law.”).

[321].Rubenstein, supra note 172, at 1137.

[322]. See generallyid.; Rubenstein & Gulasekaram, supra note 115; Elengold & Glater, supra note 115.

   [323]. Rubenstein, supra note 172, at 1137; see also Elengold & Glater, supra note 115, at 971.

[324]. See Rubenstein, supra note 172, at 1133 (noting that private contractors are exempt from both constitutional protections and from statutory constraints on the government, such as the Administrative Procedure Act and Freedom of Information Act).

   [325]. Seeid. at 1133, 1139, 1156; see also Elengold & Glater, supra note 115, at 971–73 (arguing that the “expansion to cover nongovernmental actors” through preemption and intergovernmental immunity “enables and fosters a mutually beneficial alliance between corporate contractors and the executive branch,” which “operates to strip legal remedies from private individuals and state governments and to impede oversight from state and federal regulators”).

   [326]. U.S. Const. art. VI, cl. 2 (“Laws of the United States . . . shall be the supreme Law of the Land . . . Laws of any State to the Contrary notwithstanding.”); see, e.g., Hughes v. Talen Energy Mktg., LLC, 578 U.S. 150, 162 (2016).

[327].Cipollone v. Liggett Grp., Inc., 505 U.S. 504, 516 (1992).

[328].Crosby v. Nat’l Foreign Trade Council, 530 U.S. 363, 372 (2000) (citations omitted).

[329].Fla. Lime & Avocado Growers, Inc. v. Paul, 373 U.S. 132, 142–43 (1963).

[330].Hines v. Davidowitz, 312 U.S. 52, 67 (1941).

[331].Maryland v. Louisiana, 451 U.S. 725, 746 (1981).

[332].Owino v. CoreCivic, Inc., No. 17-CV-1112 JLS (NLS), 2018 WL 2193644, at *16 (S.D. Cal. May 14, 2018) (“[C]ourts should assume that ‘the historic police powers of the States’ are not superseded ‘unless that was the clear and manifest purpose of Congress.’” (quoting Arizona v. United States, 567 U.S. 387, 400 (2012))); see also Wyeth v. Levine, 555 U.S. 555, 565 (2009).

[333].Fort Halifax Packing Co. v. Coyne, 482 U.S. 1, 21 (1987); see also Nwauzor v. GEO Grp., Inc., 127 F.4th 750, 768 (9th Cir. 2025) (“The MWA falls squarely within the states’ historic police powers to establish and require payment of a minimum wage . . . . We therefore apply the presumption against preemption.”).

[334].Chicanos Por La Causa, Inc. v. Napolitano, 558 F.3d 856, 865 (9th Cir. 2009).

[335].Washington v. GEO Grp., Inc., 283 F. Supp. 3d 967, 978 (W.D. Wash. 2017);Chen v. GEO Grp., Inc., 287 F. Supp. 3d 1158, 1167 (W.D. Wash. 2017).

[336]. See De Canas v. Bica, 424 U.S. 351, 354 (1976) (“Power to regulate immigration is unquestionably exclusively a federal power.”); Plyler v. Doe, 457 U.S. 202, 225 (1982) (“The States enjoy no power with respect to the classification of aliens. This power is ‘committed to the political branches of the Federal Government.’” (quoting Matthews v. Diaz, 426 U.S. 67, 81 (1976))); see also U.S. Const. art. I, § 8, cl. 4 (granting Congress the power “[t]o establish an uniform Rule of Naturalization”).

[337]. See, e.g.,Hines v. Davidowitz, 312 U.S. 52, 61–62 (1941); Arizona, 567 U.S. at 394, 401–02.

[338]. See Owino v. CoreCivic, Inc., No. 17-CV-1112 JLS (NLS), 2018 WL 2193644, at 15 n.9, 17 (S.D. Cal. May 14, 2018); Washington, 283 F. Supp. 3d at 977; Chen, 287 F. Supp. 3d at 1165; Novoa v. GEO Grp., Inc., No. 5:17-cv-02514-JGB-SHK, 2022 WL 2189626, at *25 (C.D. Cal. Jan. 25, 2022).

[339].Novoa v. GEO Grp., Inc., No. 5:17-cv-02514-JGB-SHK, 2018 WL 3343494, at *4 (C.D. Cal. June 21, 2018).

[340]. Id.

[341]. Owino, 2018 WL 2193644, at *17.

[342]. Washington, 283 F. Supp. 3d at 978; Chen, 287 F. Supp. 3d at 1167.

[343]. Novoa, 2018 WL 3343494, at 4; Washington, 283 F. Supp. 3d at 977; Chen, 287 F. Supp. 3d at 1165; Owino, 2018 WL 2193644, at 17.

[344]. Chen, 287 F. Supp. 3d at 1165 (citing 8 U.S.C. §§ 1103, 1225, 1226, 1226a, 1231, 1324a, 1555(d)).

[345]. Washington, 283 F. Supp. 3d at 978; Chen, 287 F. Supp. 3d at 1167.

[346]. See 8 U.S.C. § 1324a(a)(1)(A) (making it illegal to knowingly hire undocumented immigrants); id. § 1324a(e)(4) (imposing civil penalties); id. § 1324a(f) (imposing criminal penalties).

[347]. See Garfinkel, supra note 53, at 1309.

[348]. See id. (analyzing the congressional purpose underlying IRCA (citing Kati L. Griffith, When Federal Immigration Exclusion Meets Subfederal Workplace Inclusion: A Forensic Approach to Legislative History, 17 N.Y.U. J. Legis. & Pub. Pol’y 881, 887–88 (2014))).

   [349]. Immigration Reform and Control Act of 1986, Pub. L. No. 99-603, § 111(d), 100 Stat. 3359, 3381.

[350].Wyeth v. Levine, 555 U.S. 555, 565 (2009) (citations omitted).

   [351]. 8 U.S.C. § 1324a(h)(2).

[352].Washington v. GEO Grp., Inc., 283 F. Supp. 3d 967, 975 (W.D. Wash. 2017); Chen v. GEO Grp., Inc., 287 F. Supp. 3d 1158, 1164 (W.D. Wash. 2017).

   [353]. See Novoa v. GEO Grp., Inc., No. 5:17-cv-02514-JGB-SHK, 2018 WL 3343494, at *3 (C.D. Cal. June 21, 2018); Washington, 283 F. Supp. 3d at 975; Chen, 287 F. Supp. 3d at 1164.

[354]. See Chen, 287 F. Supp. 3d at 1164; Washington, 283 F. Supp. 3d at 975.

[355]. Novoa, 2018 WL 3343494, at *3 (concluding that California Labor Code “does not impose penalties on employers for hiring unauthorized aliens”); Washington, 283 F. Supp. 3d at 975–76 (“Even if, as Defendant argues, the provisions of the Washington Minimum Wage Act are construed as ‘sanctions,’ they would not be imposed on account of employers hiring unauthorized aliens, but rather because of the failure to pay the prevailing minimum wage.”); Chen, 287 F. Supp. 3d at 1164 (same).

[356].Owino v. CoreCivic, Inc., No. 17-CV-1112 JLS (NLS), 2018 WL 2193644, at *18 (S.D. Cal. May 14, 2018) (citing Salas v. Sierra Chem. Co., 327 P.3d 797, 806 (Cal. 2014) (“[F]ederal regulation imposed by the Immigration Reform and Control Act of 1986 is not so pervasive as to leave no room for any state law on the same subject.”)).

[357]. Novoa, 2018 WL 3343494, at *4.

[358]. Id. at *5–6 (citing Reyes v. Van Elk, Ltd., 148 Cal. App. 4th 604, 617 (2007), cert. denied sub nom., Van Elk, Ltd. v. Reyes, 552 U.S. 1180 (2008)).

[359]. Novoa, 2018 WL 3343494, at *6.

[360]. Id.

[361].Washington v. GEO Grp., Inc., 283 F. Supp. 3d 967, 978 (W.D. Wash. 2017); Chen v. GEO Grp., Inc., 287 F. Supp. 3d 1158, 1167 (W.D. Wash. 2017) (later Nwauzor v. GEO when the lead plaintiff was substituted).

[362]. Washington, 283 F. Supp. 3d at 978; Chen, 287 F. Supp. 3d at 1167.

[363]. Washington, 283 F. Supp. 3d at 978; Chen, 287 F. Supp. 3d at 1167.

   [364]. Reply Brief of Defendant-Appellant The GEO Group at 20–25, Nwauzor v. GEO Grp., Inc., 127 F.4th 783 (9th Cir. 2025) (Nos. 21-36024, 21-36025) (making a section 1555(d) preemption argument only).

   [365]. Brief for the United States as Amicus Curiae in Support of Appellant at 12–20, Washington v. GEO Grp., Inc., 127 F.4th 783 (9th Cir. 2025) (Nos. 21-36024, 21-36025) (making a section 1555(d) obstacle preemption argument only).

   [366]. An Act to Authorize Certain Administrative Expense for the Department of Justice, and Other Purposes, Pub. L. No. 81-503, § 6, 64 Stat. 380, 381 (1950) (codified as 8 U.S.C. § 1555(d)).

[367]. See Department of Justice Appropriation Act of 1978, Pub. L. No. 95-86, 91 Stat. 424, 426; Stevens, supra note 26, at 465–66.

   [368]. An Act to Authorize Certain Administrative Expense for the Department of Justice, and Other Purposes, Pub. L. No. 81-503, 64 Stat. 380, 381 (1950); see Reorganization Plan Number V of 1940, 5 Fed. Reg. 2223, 54 Stat. 1238 (1940).

   [369]. Sinha, supra note 41, at 25–29.

[370]. Id. at 27–29.

[371]. Id. at 29; Geneva Convention Relative to the Treatment of Prisoners of War art. 62, Aug. 12, 1949, 6 U.S.T. 3316, 75 U.N.T.S. 135.

[372].Novoa v. GEO Grp., Inc., No. 5:17-cv-02514-JGB-SHK, 2018 WL 3343494, at *4 (C.D. Cal. June 21, 2018) (citing Bernstein v. Virgin Am., Inc., 227 F. Supp. 3d 1049, 1071 (N.D. Cal. 2017)); Salas v. Sierra Chem. Co., 327 P.3d 797, 806 (Cal. 2014)); see also Washington v. GEO Grp., Inc., 283 F. Supp. 3d 967, 977 (W.D. Wash. 2017) (“From the text of § 1555(d), Defendant has not shown that Congress intended to preempt state law regarding detainee wages.”); Chen v. GEO Grp., Inc., 287 F. Supp. 3d 1158, 1166 (W.D. Wash. 2017) (same).

[373]. Novoa, 2018 WL 3343494, at *4; Washington, 283 F. Supp. 3d at 977; see also Chen, 287 F. Supp. 3d at 1166.

[374]. Novoa, 2018 WL 3343494, at *4.

[375].Owino v. CoreCivic, Inc., No. 17-CV-1112 JLS (NLS), 2018 WL 2193644, at *20 (S.D. Cal. May 14, 2018).

[376]. Id. at *19.

[377]. Id.

[378]. Id. at *20.

[379]. Id.

   [380]. Brief for the United States as Amicus Curiae in Support of Appellant, Washington v. GEO Grp., Inc., 127 F.4th 783 (9th Cir. 2025) (Nos. 21-36024, 21-36025).

[381]. Id. at 16–17.

[382]. Id. at 14.

[383]. Id. at 16.

[384].Nwauzor v. GEO Grp., Inc., 127 F.4th 750, 767–70 (9th Cir. 2025).

[385]. Id. at 769.

[386]. Id.

[387]. Id.

[388]. Id.

   [389]. Petition for Writ of Certiorari, GEO Grp., Inc., v. Nwauzor, No. 25A464 (U.S. Jan. 9, 2026).

[390].Novoa v. GEO Grp., Inc., No. 5:17-cv-02514-JGB-SHK, 2018 WL 3343494, at *5 (C.D. Cal. June 21, 2018) (citing and rejecting GEO’s argument that FLSA field preempts California’s state minimum wage law). Notably, the federal government conceded in an amicus brief that “[p]reemption here does not arise from the FLSA,” which establishes a floor, not a ceiling, for state wage regulations. Brief for the United States as Amicus Curiae in Support of Appellant at 13 n.2, Washington v. GEO Grp., Inc., 127 F.4th 783 (9th Cir. 2025) (Nos. 21-36024, 21-36025) (citing Bruns v. Mun. of Anchorage, 182 F.3d 924 (9th Cir. 1999)); see 29 U.S.C. § 218(a).

[391].Owino v. CoreCivic, Inc., No. 17-CV-1112 JLS (NLS), 2018 WL 2193644, at *18 (S.D. Cal. May 14, 2018) (citing and rejecting CoreCivic’s argument that the “INA preempts any state regulation in the field of immigration detention”).

   [392]. GEO and CoreCivic have also argued their contractual relationship with ICE itself preempts detained workers’ state law claims. Professor Rubenstein critiques this type of argument as “preemption by contract.” Rubenstein, supra note 172, at 1166–90. District courts have readily dismissed these claims in detained worker cases, emphasizing that these contracts, and the PBNDS incorporating the VWP into them, do not have force of law and therefore cannot have any preemptive effect. See Washington v. GEO Grp., Inc., 283 F. Supp. 3d 967, 977 (W.D. Wash. 2017); Chen v. GEO Grp., Inc., 287 F. Supp. 3d 1158, 1166 (W.D. Wash. 2017); Novoa, 2018 WL 3343494, at *5; see also GEO Grp., Inc. v. Inslee, 720 F. Supp. 3d 1029, 1060 (W.D. Wash. 2024), reconsideration denied, No. C23-5626 BHS, 2024 WL 1621078 (W.D. Wash. Apr. 15, 2024) (finding PBNDS “does not indicate the intent of Congress” and therefore has no preemptive effect against state legislation regulating private immigration detention centers).

[393].McCulloch v. Maryland, 17 U.S. 316, 436 (1819).

   [394]. Id.; see Davis v. Mich. Dep’t of Treasury, 489 U.S. 803, 810 (1989).

   [395]. Elengold & Glater, supra note 115, at 993; see also Novoa v. GEO Grp., Inc., No. 5:17-cv-02514-JGB-SHK, 2022 WL 2189626, at *22 (C.D. Cal. Jan. 25, 2022) (acknowledging some “ambiguity” but finding “that sufficient support exists for application of tax cases to state regulation cases.”). But courts have occasionally distinguished intergovernmental tax immunity as compared to other types of regulations. See Rubenstein, supra note 172, at 1192 n.361.

[396].United States v. Washington, 596 U.S. 832, 835 (2022); North Dakota v. United States, 495 U.S. 423, 435 (1990).

[397]. See generally Rubenstein, supra note 172, at 1190–99 (analyzing the application of intergovernmental immunity to private federal contractors as “privatized immunity”).

[398]. See United States v. New Mexico, 455 U.S. 720, 735 (1982); see also supra Part I.E.2.b.

[399]. Washington, 596 U.S. at 838 (citing North Dakota, 495 U.S. at 435; South Carolina v. Baker, 485 U.S. 505, 523 (1988)); see also United States v. County of Fresno, 429 U.S. 452, 462–63 (1977)).

[400]. New Mexico, 455 U.S. at 735.

[401]. See United States v. Muskegon Township, 355 U.S. 484, 486–87 (1958) (finding that contractor was not a federal instrumentality); Rubenstein, supra note 172, at 1193 n.366 (citing William J. Rich, Modern Constitutional Law § 34:54 (3d ed. 2011) (“Independent contractors or employees are not considered federal instrumentalities.”)).

[402].Goodyear Atomic Corp. v. Miller, 486 U.S. 174, 181 (1988) (citing Hancock v. Train, 426 U.S. 167, 174 n.23, 179 (1976)).

[403].United States v. California, 921 F.3d 865, 882 n.7 (9th Cir. 2019) (citing Goodyear Atomic, 486 U.S. at 181).

   [404]. Nwauzor v. GEO Grp., Inc., 127 F.4th 750, 760 (9th Cir. 2025).

[405]. Id. at 760, 765.

[406]. Id. at 760 (quoting GEO Grp., Inc. v. Newsom, 50 F.4th 745, 760 n.10 (9th Cir. 2022) (en banc)); see also United States v. New Mexico, 455 U.S. 720, 735 n.11 (1982).

[407]. Nwauzor,127 F.4th at 759–67 (dismissing GEO’s post-trial appeal on intergovernmental immunity); Novoa v. GEO Grp., Inc., No. 5:17-cv-02514-JGB-SHK, 2022 WL 2189626, at *21–23 (C.D. Cal. Jan. 25, 2022) (denying GEO’s motion for summary judgment but also denying plaintiff’s motion for summary judgment to strike the affirmative defense).

[408]. See Newsom, 50 F.4th at 755; see also United States v. Washington, 596 U.S. 832, 839 (2022) (“[A] state law is thus no longer unconstitutional just because it indirectly increases costs for the Federal Government, so long as the law imposes those costs in a neutral, nondiscriminatory way.”).

[409]. See Washington, 596 U.S. at 839; Nwauzor, 127 F.4th at761 (citing Newsom, 50 F.4th at 755); Novoa, 2022 WL 2189626, at *22 (citing United States v. Boyd, 378 U.S. 39, 44 (1964)).

[410]. Novoa, 2022 WL 2189626, at *22 (“That ICE outsources work to private detention facilities like GEO does not make federal contractors instrumentalities, particularly as the latter pursue their ‘own private ends – in connection with commercial activities carried on for profit.’” (quoting Boyd, 378 U.S. at 44)).

[411]. Washington, 596 U.S. at 839 (citing Washington v. United States, 460 U.S. 536, 546 (1983); United States v. North Dakota, 495 U.S. 423, 438 (1990)).

   [412]. Rubenstein & Gulasekaram, supra note 115, at 230.

[413]. Id. at 234–35 (citing César Cuauhtémoc García Hernández, Immigration Detention as Punishment, 61 UCLA L. Rev. 1346, 1382–92 (2014)).

[414]. Id. at 235.

[415]. See, e.g., Ruelas v. County of Alameda, 108 F.4th 1208 (9th Cir. 2024) (California).

[416]. See, e.g., Miller v. Dukakis, 961 F.2d 7 (1st Cir. 1992) (Massachusetts); Sanders v. Hayden, 544 F.3d 812 (7th Cir. 2008) (Wisconsin); Strauss v. Dreyfus, No. C10-5646BHS, 2011 WL 809797, at *1 (W.D. Wash. Mar. 2, 2011) (Washington).

   [417]. Rubenstein, supra note 172, at 1198.

[418]. See Washington v. GEO Grp., Inc., No. 3:17-cv-05806-RJB, 2021 WL 3675011, at *4 (W.D. Wash. Aug. 19, 2021); Nwauzor v. GEO Grp., Inc., 127 F.4th 750, 763–67 (9th Cir. 2025).

[419]. See supra Part I.B.

[420]. Nwauzor, 127 F.4th at 765–66.

[421]. Id. at 766 (citing Nwauzor v. GEO Grp., Inc., 540 P.3d 93, 99–100 (Wash. 2023)).

[422].Novoa v. GEO Grp., Inc., No. 5:17-cv-02514-JGB-SHK, 2022 WL 2189626, at *23 (C.D. Cal. Jan. 25, 2022).

[423]. Id.

   [424]. Interview with Jose Ruben Hernandez Gomez (Mar. 20, 2024).

   [425]. Andrew Benoit, Combating Human Rights Violations at Tacoma’s Northwest Detention Center, Trail (Feb. 29, 2024), https://trail.pugetsound.edu/?p=18636 [https://perma.cc/49EB-NH5D].

   [426]. Gustavo Sagrero Álvarez, Video Shows Heavily Armed Guards Detaining Hunger Strikers at Tacoma ICE Center After Using Tear Gas, KUOW (Feb. 1, 2024), https://www.kuow.org/stories/video-shows-heavily-armed-guards-detaining-hunger-strikers-at-tacoma-ice-center-after-using-tear-gas [https://perma.cc/8E7S-W38G].

   [427]. Johnson, supra note 21.

This Article exposes privacy law’s complicity in this phenomenon of “brokered abuse.” Today, victims seeking obscurity can ask data brokers to remove their online information. But a web of privacy laws props up a fragmented and opaque system that forces victims to navigate potentially hundreds of distinct opt-out processes, wait months for their information to be removed, and then repeat this process continuously to ensure their information doesn’t resurface. The status quo compels victims to manage their own privacy, placing the burden of maintaining obscurity on already overburdened shoulders.

In response, this Article proposes a new regulatory regime premised on a transformative reallocation of responsibility. In short, it proposes a techno-legal system that would enable victims to obscure their information across all data brokers with a single request, redistributing the burden away from victims and onto brokers. Such a system is justified, feasible, and constitutional. The data broker industry is eager to assert that it has a First Amendment right to exploit people’s data, but this Article develops a trio of arguments to confront this controversial claim of corporate power. By blending theory, policy, and technical design, this Article charts a path toward meaningful privacy protections for victims and, ultimately, a more empathetic legal landscape for those most at risk.

Table of Contents Show

Introduction

Ella was in college when the abuse began.[1] She dated Nick for a while before trying to end their relationship. At that point, he “turned rather obsessive, showing up at my school, then showing up at my work.” Before long, he came to her home to threaten her. She went to the police but wasn’t taken seriously. Then, things escalated. Nick showed up with a weapon. “I was almost killed.” After the incident, what stood between her and almost getting killed again? A few dollars and a few clicks on the internet. This is the plight of brokered abuse—the phenomenon of how data brokers enable and exacerbate stalking, harassment, and violence.[2]

Ella had left her abusive relationship, sought help, and fought for a restraining order. But none of that protected her when her abuser could still track her. “I’m not sure how, but [he] found information for my parents and made threatening calls to [them] as well. . . . [W]e knew it was him, but we were never able to do anything about it.” With police unwilling or unable to intervene, Ella tried to erase herself—a defense mechanism that victims of brokered abuse know all too well. She abandoned the internet, moved cities, and changed her name, number, and career. And yet every time she tried to rebuild her life, her stalker found her again. The terror of knowing that digital breadcrumbs could lead him back to her consumed her. “If I Google my name and I’m showing up on Whitepages, People Finder, Spokeo, TrueIdentity—the list goes on and on and on—it’s scary.” Scrubbing her data became a constant, exhausting necessity. “This act of shielding myself became part of my everyday life.”

Unfortunately, Ella’s story is not unique. Countless others are trapped in cycles of fear and vigilance, their safety undermined by the ruthless machinery of the data broker economy.[3] Data brokers are entities that collect personal information from public records, such as voter registrations and court filings, as well as private sources, including online purchases, social-media activity, and GPS location data.[4] They use this information to create comprehensive dossiers, detailing intimate aspects of individuals’ lives—addresses, phone numbers, financial histories, family relationships, and more.[5] Brokers then sell these dossiers to businesses, government agencies, and even individuals.[6] The entire industry thrives on eliminating obscurity by systematically dismantling the practical difficulty of accessing and compiling personal information.[7] Victims like Ella are not seeking obscurity for the sake of secrecy but rather as a matter of safety and wellbeing. Yet the data-broker economy makes personal information widely accessible, fueling cycles of interpersonal abuse.[8]

The harms inflicted by data brokers are twofold: primary and secondary.[9] Primary harms arise from the immediate danger victims face when their personal information is exposed.[10] For victims like Ella, the knowledge that their abuser can easily track their movements, find their home, or access their contact details creates a constant state of fear and vulnerability. Despite the many steps Ella took to disappear, she could not escape the reach of data brokers that repeatedly exposed her location. This persistent threat can force victims to withdraw from social, professional, and community life, sacrificing opportunities and relationships in an effort to stay safe.[11]

Beyond these primary harms, victims endure secondary harms as they attempt to protect themselves in a broken system. The fragmented and opaque processes required to remove information from broker databases are arduous and retraumatizing.[12] Victims must locate and contact hundreds of brokers, verify compliance, and continuously monitor whether their data resurfaces.[13] Each step exacts an emotional and financial toll, forcing victims to revisit their trauma and confront the very systems that profit from their exposure.[14] These secondary harms compound victims’ suffering.

The law is complicit in these harms. The American legal tradition of individual rights has left an indelible mark on U.S. privacy law.[15] Today’s privacy landscape defaults to a system of privacy self-management that Daniel Solove has long critiqued as both unrealistic and inequitable.[16] Privacy self-management assumes that individuals can and should navigate complex systems to manage their own privacy, making informed choices about how their data is collected, stored, and shared.[17] While this concept aligns with a legal tradition rooted in individual rights and personal autonomy, it falls woefully short in the face of today’s sprawling data ecosystems. Building on Solove’s work, Ella Corren has effectively shown that privacy self-management places an impossible burden on individuals; her work offers an empirical rebuttal to the presumption that people generally have the resources, expertise, and bandwidth to make meaningful decisions about their privacy.[18] For victims of brokered abuse, the framework of privacy self-management is especially harmful. It forces them to bear the weight of achieving obscurity, navigating labyrinthine systems, and negotiating with powerful data brokers—all while managing the immediate risks posed by their abusers.[19] By defaulting to privacy self-management, the legal system fails to protect the vulnerable and allows brokers to profit from the commodification of personal information.[20] For abuse victims, obscurity is not a luxury—it can be a necessity for survival. The failure of privacy self-management to deliver this obscurity underscores the urgent need for systemic reform.

At least in this narrow context, a fundamental shift in privacy law is needed to center victims and redistribute the burden of achieving obscurity. As Ella’s story demonstrates, even without the need to request the removal of personal information, victims might already be overwhelmed by the challenges of escaping and surviving abuse.[21] The law should instead shift the responsibility for achieving obscurity to the parties that create and profit from the risk: data brokers. These entities possess the resources, technology, and expertise to manage the logistical and technical challenges of obscuring sensitive information. Unlike victims, brokers are well positioned to systematically remove identifying data from their systems and prevent its reappearance.[22] Redistributing this burden would not only be fair but also represent a more effective and sustainable solution to the problem of brokered abuse. By compelling brokers to take responsibility for the risks they create, the law can begin to rectify the systemic injustices that leave victims like Ella fighting for their safety alone.[23]

Redistributing this burden requires a centralized regulatory and technical solution. A fragmented, decentralized approach has proven incapable of addressing the pervasive and evolving threats posed by the broker industry.[24] Victims should be able to invoke their right to obscurity with a single request and expect each broker to honor an ongoing responsibility to identify and remove all their relevant information across every database and ensure that information never resurfaces.[25]

This idea has already entered the regulatory imagination, at least in part. California recently passed the DELETE Act,[26] several other states have proposed their own versions,[27] and a federal DELETE Act has been proposed in Congress.[28] These efforts reflect a small but growing consensus that individuals should not have to navigate hundreds of opaque data broker opt-out processes on their own.

Yet these laws suffer from two fundamental flaws. First, they broadly regulate everyone’s data without tailoring protection to those most at risk, opening them up to viable First Amendment challenges.[29] Scholars like Robert Post, Frederick Schauer, and Amanda Shanor have documented how companies are increasingly wielding the First Amendment to serve a deregulatory agenda,[30] and data brokers have left little mystery as to their willingness to challenge privacy regulations on First Amendment grounds.[31] Accordingly, policymakers should expect fierce First Amendment opposition to privacy regulation and proactively address constitutional scrutiny. Second, the California and federal statutes delegate core technical questions of implementation to future rulemaking,[32] leaving open the risk of ineffective or even injurious compliance mechanisms, which entrench broker control rather than alleviate the burden on victims.[33] In contrast, this Article offers a detailed regulatory approach that would strengthen both the legal durability and practical efficacy of broker regulation.

To develop and justify our central proposal, this Article builds on scholarship framing obscurity as a discrete privacy interest,[34] highlighting the unique vulnerabilities of abuse victims,[35] critiquing the framework of privacy self-management,[36] and confronting the First Amendment’s deregulatory “Lochnerian” turn.[37] In so doing, this Article makes three main contributions. First, it demonstrates how the combination of decentralized broker opt-out systems and privacy self-management subjects victims to harms beyond those arising from the abuse itself: arduous and relentless vigilance to maintain their obscurity, retraumatization from repeatedly revisiting their abuse, and withdrawal from society for fear of generating more identifying data.[38]

Second, the Article proposes a novel regulatory framework that centers victims and harnesses obscurity to protect human safety. By blending theory, policy, and technical design, the Article presents and justifies a centralized system that would transfer the burden of obscurity onto the billion-dollar industry profiting from brokered abuse.[39] While lawmakers have begun flirting with this idea, the devil is in the details, and this Article answers complex questions about how regulators could implement such an approach—and why they should.[40]

Third, this Article tackles a doctrinal question often avoided by privacy and First Amendment scholars (and by many privacy laws): Can a data privacy law that covers information gathered from government records and other public sources be constitutional?[41] Through a trio of arguments, the Article anticipates and confronts any claim that the Constitution insulates the broker industry from a centralized obscurity system mandated by law. As an initial matter, the Article challenges the assumption that broker practices are covered by the First Amendment, building on emerging scholarship that questions whether the commodification of personal information serves First Amendment values.[42] It then contests arguments that data dossiers constitute noncommercial speech.[43] Finally, it details why legislating a centralized obscurity system for victims should survive strict scrutiny.[44] In doing so, it joins the work of scholars critiquing how expansive interpretations of the First Amendment undermine regulatory efforts and privilege corporate interests over human dignity.[45]

The Article proceeds in four parts. Part I outlines the mechanics of brokered abuse, detailing the inner workings of the broker industry and why its practices endanger people facing various forms of interpersonal harm. It argues that the harms of brokered abuse are not inevitable; rather, they are partly the result of a legally constructed and broken system that asks too much of the most vulnerable.

Part II then explains why a victim-centered solution demands a paradigm shift in privacy law—one that redistributes responsibility from individuals to the data brokers who profit from their exposure. This Part justifies such a redistribution by emphasizing principles of fairness and efficiency, arguing that brokers, as the least-cost avoiders, are uniquely equipped to shoulder the logistical and technical obligations of obscurity.[46] It concludes by drawing lessons from the private and public sectors to demonstrate the feasibility and efficacy of a centralized approach to obscurity.[47]

Part III provides the statutory and technical blueprint for a centralized system that would allow people like Ella to obscure their information with a single request. It begins by identifying critical gaps in the state and federal DELETE Acts that undermine their ability to deliver meaningful protection.[48] It then presents a detailed framework that integrates a centralized victim opt-out registry, rigorous compliance obligations, and advanced technical tools like cryptographic matching to ensure robust enforcement.[49]

Finally, Part IV confronts thorny First Amendment questions that such a regulatory regime would face. It casts doubt on brokers’ claims that a law like the one we propose would even trigger constitutional scrutiny. It then argues that, at most, our intervention should be assessed under intermediate scrutiny, the same standard reserved for regulations of commercial speech. Regardless, this Part concludes by detailing why a centralized obscurity system for abuse victims should nevertheless survive strict scrutiny as a narrowly tailored regulation to achieve a compelling government interest. This doctrinal analysis not only fortifies the proposal against constitutional attack but also contributes to broader debates on the role of the First Amendment in data privacy regulation.[50]

The status quo requires victims like Ella to manage their own privacy, placing the burden of maintaining obscurity on already overburdened shoulders. This Article offers a path forward that transforms obscurity from an unobtainable ideal into an enforceable reality.

I. Existing Law and Technology Puts the Obscurity Burden on Victims

I was spending hundreds of hours online just looking and searching and going through everything. . . . It’s like playing whack-a-mole . . . . And it’s frustrating because it’s such a huge waste of time as well—such a burden on your daily life.

—Ella

The responsibility of achieving personal safety through obscurity—grounded in the principle of protecting privacy by making personal information difficult to access—currently rests almost entirely on the shoulders of victims of abuse.[51] For these individuals, obscurity is not a theoretical concept; it can be their best, and often their only, defense against abusers who exploit the data-broker ecosystem to locate, surveil, and harm them.

Despite the potential life-or-death stakes, victims must navigate a fragmented and convoluted system of opt-out processes to secure this obscurity, shouldering the dual burden of primary and secondary harms.[52] Primary harms arise directly from the loss of obscurity—the stalking, harassment, and violence enabled by data brokers who make sensitive personal information easily accessible to abusers.[53] Secondary harms, by contrast, are inflicted by the legally constructed privacy system itself, which forces victims to undertake the grueling and often futile task of privacy self-management.[54]

This dual burden—the risk of exposure on one hand and the demands of self-management on the other—defines the plight of victims in the brokered-data economy. These burdens are not merely onerous; by failing to take seriously the experiences of many victims, they are devoid of empathy.[55] This regulatory failure is a systemic injustice that prioritizes corporate convenience over human safety. It is complicit in the harm perpetrated by abusers and data brokers.[56]

The following Part examines these intertwined harms in detail and critiques current privacy laws for their failure to prioritize the safety and dignity of victims.

A. The Harms of Brokered Abuse

The primary harms of brokered abuse—the stalking, harassment, and physical and psychological threats—are exacerbated by the secondary harms tied to privacy self-management. The privacy self-management system, ostensibly designed to protect privacy, is, for victims, a system that instead inflicts further injury. Obscurity is critical to victims’ safety, but achieving it has become tantamount to fighting a broker hydra while equipped with little more than desperate conviction.

1. The Primary Harms of Data Exposure

For victims of abuse, obscurity is not an abstract privacy ideal; it can be their best, and often only, defense against abusers.[57] The primary harms of brokered abuse are rooted in the destruction of obscurity.[58] By making sensitive personal information easily accessible, data brokers empower abusers to locate and target their victims.[59] Commodifying obscurity can manifest physical, psychological, financial, and social harms for those who are vulnerable or in life-threatening situations.

The broker industry is a sprawling, multibillion-dollar ecosystem that harvests and sells personal information.[60] There is little oversight or accountability.[61] These companies build their businesses by acquiring and repackaging information from both public records and private sources, often without explicit permission from the people whose lives they catalog.[62] At a minimum, brokers aggregate public documents accessible to anyone with the means to request them, such as property deeds, voter rolls, and marriage licenses.[63] Many exploit transparency tools, such as the Freedom of Information Act (FOIA), to obtain government-held information,[64] while some partner with third parties to collect data about online behaviors from apps, e-commerce, social media, and subscription services.[65] Advanced technologies, such as facial recognition and real-time geolocation tracking, carry the potential to supercharge datasets with unprecedented accuracy and granularity.[66] The proliferation of machine learning models enables brokers to infer new data points—including religion, sexual orientation, and even mental health—from existing datasets.[67] These curated dossiers form the core of the broker business model.[68] While there are arguable benefits to data brokerage, such as its use in journalism,[69] law enforcement,[70] and reconnecting with lost relatives,[71] the potential for harm is both pervasive and severe.[72]

The erosion of obscurity underlies many of these harms. In the context of interpersonal abuse, unquantifiable risk and injury stems from people exploiting brokered data and the data brokers enabling them.[73] The broker ecosystem arms malicious actors with the ability to reconstruct a target’s personal history, locate their current whereabouts, or predict their movements.[74] Through these services, abusers can gain access to information that would otherwise be difficult or impossible to acquire. For example, a victim may relocate, change phone numbers, enroll in address-confidentiality programs, and take other steps to disappear; yet all this will be in vain when a broker finds and sells their updated information.

Worse still, brokers make these dossiers cheaper and easier to access than ever,[75] paving the way for nefarious use.[76] Tracking someone once required significant financial and logistical effort—hiring private investigators, obtaining court orders, or waiting for public records to update.[77] Today, abusers can purchase detailed reports on a victim’s location, family connections, and employment history for as little as a few dollars.[78] This democratization of surveillance tools transforms victims’ lives into open books, indexed for convenience and accessible to anyone with a computer, an internet connection, and a credit card. Brokers can render even the most robust protective measures futile.[79]

The pervasive and persistent availability of brokered data undermines victims’ ability to rebuild their lives, forcing them into cycles of isolation and hypervigilance. Obscurity is not just about safety; it is often the prerequisite for healing and stability.[80] Without it, victims might live in constant fear of discovery, unable to feel secure in their surroundings or relationships. Advanced tools turn fleeting interactions into lasting vulnerabilities and loved ones into unwitting accomplices to abuse. For example, facial recognition databases can turn a single photo uploaded to social media by a friend into a surveillance data point. Similarly, the decision to kill time on Candy Crush can generate real-time location data for the taking.[81] Even attempts to adopt new routines can be thwarted by behavior-modeling tools that allow abusers to anticipate a victim’s actions or locations based on historical patterns.[82]

The erosion of obscurity forces victims to withdraw from social and professional life to avoid leaving traces that could expose them to harm. Fear of exposure can lead victims to delete social media accounts, avoid professional networking, decline opportunities that might publicly associate them with a new location, and even refrain from voting.[83] While these steps may reduce immediate risk, they can come at a steep cost, cutting victims off from the support systems and opportunities necessary to rebuild their lives.[84] This isolation might not only deepen the psychological wounds inflicted by abuse but also amplify the societal stigma that can accompany such experiences, leaving victims feeling abandoned and unsupported.[85]

Further, the economic impact of losing access to opportunities and resources exacerbates the damage inflicted by brokered abuse. Victims often face significant financial costs associated with escaping abuse, including relocation expenses, legal fees, and lost wages.[86] Forgoing opportunities to earn income or receive external help can make these costs insurmountable. For marginalized individuals, these economic challenges are even more pronounced, as systemic inequities compound the difficulties of navigating both abuse and the exploitative practices of data brokers.[87]

2. The Secondary Harms of Privacy Self-Management

The secondary harms of brokered abuse arise from the expectation that victims are wholly responsible for managing their own privacy to achieve safety, a framework referred to as “privacy self-management.”[88] This approach, rooted in the belief that individuals should control how their personal data is collected, shared, and accessed, assumes that individuals are best positioned to make decisions about their privacy and can protect themselves by asserting their rights.[89] In practice, however, self-management places an overwhelming and unfair burden on people who are poorly equipped to bear it, particularly in the context of brokered abuse.[90] The cumulative toll of this dynamic is profound, and it encompasses logistical, psychological, and financial costs that victimize individuals anew.[91]

Privacy self-management demands that individuals navigate a convoluted system of data brokers, each with unique and burdensome opt-out procedures.[92] First, victims must scour the internet to identify the brokers that hold their personal information.[93] Then, victims must generally contact each broker individually to request removal of their data.[94] The removal process often involves submitting detailed forms, verifying their identity, and in many cases, providing sensitive personal documentation, such as copies of government-issued IDs.[95] In a cruel irony, the opt-out processes can force victims to hand brokers more sensitive information than brokers had in the first place.[96]

From a technological perspective, brokers’ opt-out systems often lack uniformity or automation, preventing scalable privacy self-management.[97] Some brokers require physically mailed requests, while others mandate the use of proprietary online portals with arcane navigation.[98] Many broker review processes are manual, relying on individual contractors to process opt-out requests.[99] This lack of technical sophistication not only makes the process unpredictable but also ensures that it is both labor intensive and error prone.[100]

After victims successfully submit requests, there is still no guarantee of obscurity. Even when victims comply with every byzantine requirement, brokers may refuse to act on requests, citing legal exemptions or internal policies.[101] For victims lucky enough to succeed, the same data can still resurface, or new identifying information might emerge.[102]

For some individuals, third-party services offer a partial reprieve. These services, such as DeleteMe or Privacy Bee, act as intermediaries, navigating the complex web of brokers on behalf of their clients.[103] By consolidating the opt-out process, they reduce victims’ direct interaction with brokers, allowing victims to work through a single point of contact. However, these services come with significant shortcomings.[104] They can be prohibitively expensive.[105] Further, third-party removal processes, too, can require the disclosure of sensitive personal information, like proof of address. Additionally, even the best-intentioned services cannot guarantee permanent removal of data because brokers can exploit legal loopholes to retain as much information as possible and poorly enforce their own removal policies.[106]

The demands of privacy self-management force victims into an unrelenting cycle of labor and stress. Victims must monitor the internet, follow up on pending requests, submit new ones, and continually search for additional brokers.[107] These efforts consume significant time, money, and emotional bandwidth—resources many victims lack.[108] Opting out diligently often requires taking time off work, incurring costs such as postage fees, and reliving the trauma of abuse through repeated interactions with brokers.[109] Forcing victims to become full-time stewards of their own obscurity creates a two-tiered system where only those with the means to pay can access meaningful protection.[110]

This systemic burden is compounded by the inherent power imbalance between individuals and the broker industry. Brokers operate vast, interconnected networks that aggregate and sell personal information with minimal oversight or accountability.[111] Victims, by contrast, are left to navigate this labyrinthine system alone, often without guidance or assurance that their efforts will result in meaningful protection.[112] The asymmetry of information, resources, and power sets victims up to fail by leaving them exposed and disempowered—robbed of agency over their own safety.[113]

The retraumatization caused by engaging with these processes can compound the psychological harm victims endure.[114] To protect themselves, victims must repeatedly reopen old wounds for brokers, entities that are neither trauma-informed professionals nor concerned with victims’ dignity.[115] Some brokers even demand detailed documentation of abuse, such as restraining orders, police reports, or affidavits. Each form submitted, each identity verified, and each explanation of abuse can drag victims back into the shadows of their trauma.[116] The very act of putting these experiences together into words can be deeply triggering, confronting victims with the fear, pain, and humiliation they might hope to leave behind.[117] The demand to hand over personal information can feel eerily reminiscent of the privacy invasions they experienced at the hands of their abusers.[118]

By leaving victims with no option other than pursuing obscurity through privacy self-management, the law neglects the unique vulnerabilities and lived experiences of abuse victims. This framework prioritizes corporate convenience over human safety, creating systemic barriers that retraumatize and disadvantage victims while failing to provide meaningful or lasting protection. Addressing this injustice requires a paradigm shift away from placing the burden of safety on victims and toward holding brokers accountable for the risks their practices create.

B. The Inadequacy of Existing Laws

The patchwork of privacy laws in the United States fails to adequately address the primary and secondary harms of brokered abuse, often to the point of complicity.[119] These laws either curb brokered abuse narrowly and indirectly[120] or craft interventions that impose new burdens on victims.[121] Together, they amount to a system where victims must shoulder the overwhelming responsibility of managing their own privacy, while abusers and brokers exploit the gaps with impunity.[122]

Laws addressing abuse fall into two categories: those targeting abusers directly, by criminalizing stalking, harassment, or violence; and those aimed at brokers that knowingly facilitate abuse, such as through doxing.[123] While vital in theory, both types are deficient in practice. Laws targeting abusers directly require waiting until brokered data is used to perpetuate harm. On the other hand, while laws targeting broker activities are closer to addressing the root cause of the problem, they sometimes rely on scienter requirements, such as proving intent or knowledge of harm.[124] This effectively immunizes brokers that sell data dossiers at scale.[125] For instance, California’s antidoxing provisions prohibit sharing registered stalking victims’ data with intent to incite harm, but brokers can evade liability by disclosing data without vetting purchasers.[126]

Anti-abuse laws are impractical for more than just their substance. Their legal processes typically require victims to interact with police, prosecutors, lawyers, or judges, which could deter many from pursuing claims due to intimidation, financial barriers, or distrust of institutions. Moreover, proceedings are often too slow to address the immediate dangers of brokered abuse, and even successful cases fail to address the systemic issue of brokers continually replenishing their data stockpiles. Victims would need to file repeated claims against brokers, new and old, whenever their information reemerges online, creating an untenable cycle of litigation that does little to disrupt the larger ecosystem.

In contrast to anti-abuse laws, laws targeting data brokers often fall into three categories: (1) transparency laws, (2) laws restricting data collection, and (3) laws restricting data disclosure. Transparency laws aim to address harms caused by the broker industry by shedding light on broker practices to inform regulators (administrative transparency) or empower individuals (popular transparency).[127] For example, Vermont and California require brokers to register with state agencies and disclose information about brokers’ data sources and practices,[128] while California’s “right to know” laws allow individuals to access data brokers hold about them.[129] However, more information about the mechanisms of brokered abuse does little to protect victims from it. And, in addition to being ineffective, these laws might even sap political will from stronger proposals, allowing brokers to hide harmful practices under the veneer of compliance.[130]

Another approach involves stemming the tide of personal data at the source. Longstanding laws prohibit deceptive practices, hacking, and unauthorized scraping,[131] while newer laws, such as the California Consumer Privacy Act (CCPA), seek to limit nonconsensual data collection.[132] However, these measures are riddled with loopholes.[133] The CCPA, for example, exempts “publicly available information” and “lawfully obtained, truthful information that is a matter of public concern”[134]—categories encompassing “vast troves of brokered data.”[135] Brokers need not resort to illegal practices when a plethora of information is available legally.

While restricting data disclosure is perhaps the most promising approach, it remains fraught with challenges. Some regulations, such as tort liability for disclosing sensitive information or bans on selling location data, address the issue indirectly and incompletely.[136] More direct measures, like California’s right to opt out, allow individuals to prevent businesses from selling their data.[137] For abuse victims, California’s Safe at Home program provides more robust protections, requiring brokers to conceal registered victims’ home addresses and phone numbers for four years.[138] Victims can also seek damages for intentional violations.[139] However, victims must still approach brokers individually, submit forms repeatedly, and monitor compliance over time.[140]

Ultimately, by focusing narrowly on isolated aspects of data brokerage, the existing regulatory responses fail to disrupt the systemic features of brokered abuse.[141] Worse, they impose an untenable burden on victims, making the law complicit in the harm it purports to address.

II. Protecting Safety Through Obscurity Demands Redistributing Responsibilities

I felt like it was my responsibility to do the opting-out. . . . [I]t was this thought that if I left any kind of stone unturned, that would cause harm to me or my family. . . . Why should this be a responsibility that I need to bear?

—Ella

This Part contends that addressing brokered abuse requires a paradigm shift in privacy law that prioritizes meaningful protections for victims by redistributing responsibility to those who profit from exposing a victim’s information. The consequences of inaction are dire: If we wait for an elusive privacy panacea to cure all the ills of “informational capitalism,”[142] abusers and data brokers will continue to exploit the gaps in current law, exposing victims to primary harms like stalking and harassment as well as the retraumatizing secondary harms of navigating a fractured system to protect themselves. Recasting the pursuit of privacy as the pursuit of safety underscores the need for a centralized obscurity system for victims.[143] Drawing on models from the private and public sectors, this Part illustrates the feasibility and urgency of holding brokers accountable while relieving victims of unsustainable and unjust burdens.

A. Justifying Redistribution

To address the systemic failures of brokered abuse and privacy self-management, we propose reframing victim privacy as a shared responsibility to promote safety and redistributing the labor of achieving obscurity from victims to data brokers. The law’s overreliance on privacy self-management is, at least in this context, an indefensible abdication of policymaking responsibility.[144] This framework is poorly suited to serving anyone’s privacy, let alone the especially vulnerable among us—people fleeing violence, harassment, and exploitation.[145] A safety-focused lens for obscurity demands a paradigm shift.

The concept of privacy self-management assumes that victims can and should be responsible for interacting with thousands of data brokers, each with their own processes, policies, and pitfalls.[146] It demands vigilance, technical sophistication, and both time and resources that are rare even among the most privileged.[147] For victims of abuse, this model is not just burdensome; it is retraumatizing.[148] Many victims will need to play some role in their own protection, and asserting the right to obscurity might be a necessary initial step. However, this invocation of their right to obscurity should mark the end—not the beginning—of their involvement.

Principles of fairness and efficiency support this reallocation of responsibility. From a fairness perspective, data brokers are the most appropriate entities to bear this burden. These brokers profit directly from the dissemination of data dossiers that put vulnerable populations at risk. Holding brokers accountable for alleviating brokered abuse aligns with societal norms that require industries to mitigate the risks they create, much like environmental regulations that compel polluters to bear cleanup costs. Danielle Citron has made an analogous fairness argument in the context of justifying strict-liability regulations when companies’ private databases are hacked.[149] As she contends, “[w]hen an organization engages in reasonable risky behavior—that is, nonwrongful conduct where an injurer’s freedom to impose the risk is more valuable than a victim’s forgone security . . . —fairness requires that the injurer pay for the victim’s harm.”[150] While our proposal is concerned more with obscurity than compensation, the same point rings true when applied to the broker industry.

Moreover, brokers are the least-cost avoiders: the entities best positioned to implement systemic solutions.[151] With centralized databases, established processes for managing data, and advanced technological capabilities, brokers can integrate obscurity protections far more efficiently than individual victims. Citron again provides some theoretical scaffolding for our proposal, observing that operators of corporate databases “constitute the cheapest cost avoiders vis-à-vis individuals whose information sits in a private entity’s database,” in part because they “have distinct informational advantages about the vulnerabilities in their computer networks.”[152] While individuals lack the knowledge and ability to assess a company’s cybersecurity, Citron argues that “database operators can most efficiently spread the costs of data leaks” and are “best situated to make the optimal choice of either taking additional security precautions or insuring against security-breach losses.”[153] A similar efficiency argument can be applied to the broker industry. The cost of implementing a centralized obscurity process could be modest for an industry already thriving on the commodification of personal data, while the cost to victims of managing their own obscurity is immense. For victims, this redistribution could be a lifeline; for brokers, it would be little more than a reasonable cost of doing business.

B. Required Features of Redistribution

This Section argues that a safety-focused approach to obscurity also demands both the imposition of ongoing broker obligations to keep victim data offline and a centralized governance and enforcement mechanism to ensure brokers comply.

1. Ongoing Duties

Obscurity calls for more than a one-time response to an opt-out request. Redistributing responsibility to brokers requires imposing ongoing obligations on them to keep victim information perpetually offline. The reality of data brokerage is that information constantly flows through partnerships, secondary markets, and automated data scrapers.[154] Without robust mechanisms to prevent and deter the re-collection and redistribution of data, any initial removal will be rendered meaningless.[155]

Brokers should be compelled to implement systems that proactively guard against re-exposure. This could include closing loopholes that allow data to re-enter their networks, monitoring compliance through periodic audits, and collaborating to eliminate weak points in the broader ecosystem.[156] Treating obscurity as a one-time obligation ignores the nature of the threat; victims’ safety depends on sustained vigilance.

2. Centralized Governance and Oversight

Even if brokers were forced to bear greater responsibility for obscuring victims’ data, effective protection for victims requires centralized governance to coordinate and enforce compliance. Individual brokers cannot be trusted to regulate themselves in a decentralized system riddled with gaps and inconsistencies.[157] A centralized framework, overseen by government regulators, would provide the necessary structure to ensure that brokers fulfill their obligations.

This system would shift the burden of oversight away from victims, who are currently forced to monitor their own exposure and pursue opt-out processes individually. Instead, the government would take on the responsibility of systemic oversight, creating mechanisms for victims to report noncompliance and for regulators to periodically audit brokers. By centralizing these functions, the framework would provide victims with a single point of recourse, relieving them of the painful and painstaking task of managing their own privacy across a fragmented landscape.

Centralized governance could also ensure accountability at a systemic level, addressing gaps in enforcement that allow brokers to evade meaningful consequences.[158] By integrating oversight into a unified framework, policymakers can create a cohesive system that reinvigorates online obscurity as a meaningful protection for victims while streamlining compliance for brokers.

C. The Case for a Centralized, Coordinated Intervention

A centralized, coordinated system would most effectively reallocate the burden of maintaining obscurity away from victims and onto data brokers, while also ensuring the system has ample oversight. Such a system could enable victims to reclaim their obscurity with a single request, reducing the labor currently required to achieve even temporary relief. Once a request is submitted, brokers—not victims—would bear the responsibility for obscuring personal information promptly and permanently. Existing models in both the private and public sectors suggest that this approach is feasible and effective.

The private sector has already experimented with using centralized coordination to address systemic threats to privacy. A compelling example is Stop Non-Consensual Intimate Image Abuse (StopNCII), a global initiative that helps individuals prevent the spread of intimate images that have been shared without their consent on online platforms.[159] StopNCII empowers individuals to leverage the technological capabilities and informational advantages of online platforms to proactively prevent nonconsensual intimate images from circulating across participating platforms.[160] Through a centralized system, individuals can generate hash values—unique digital fingerprints—of their intimate images without sharing the images themselves.[161] Platforms such as Facebook, Instagram, and TikTok use this database to identify and block these images before they are further distributed, mitigating the harm that people might otherwise experience due to the repeated appearance of their images.[162] This approach minimizes the labor required to protect people’s privacy, eliminating the need for individuals to request takedowns across multiple platforms and shifting some responsibility onto the companies to prevent harm.[163]

Policymakers could learn lessons from StopNCII as they craft a centralized mechanism to address brokered abuse. Just as StopNCII enables individuals to take preemptive steps to protect themselves, a centralized obscurity system could allow abuse victims to submit a single request to remove personal information across all covered data brokers. Brokers would then bear the responsibility of ensuring that the flagged data is removed and does not reappear on their platforms. Similarly, StopNCII exemplifies how an empathetic system should be sensitive to how people can invoke its protections. While StopNCII allows people to generate hashes of the photos they want to obscure instead of sharing the actual photos, a system addressing brokered abuse could require victims to submit only the minimum amount of personal information required by brokers to identify data points to obscure.

StopNCII relies on users to generate hashes for the images they seek to take down, but victims of abuse cannot be expected to identify every single piece of information that puts them at risk and requires removal.[164] However, other examples from the private sector show that companies can coordinate to identify potentially harmful content without relying entirely on individual submissions. For example, companies like Pinterest,[165] Instagram,[166] and YouTube[167] collaborate to detect and remove self-harm and suicide-related material.[168] These platforms use centralized tools such as machine-learning algorithms to identify patterns, such as flagged keywords, imagery, and behavior, and share insights across platforms to ensure that content removed from one site is unlikely to reappear on another.[169] Instead of requiring victims to identify every piece of data that endangers them, data brokers could similarly use pooled technological resources to identify and suppress information related to particular individuals.

The public sector also provides precedents for centralized frameworks that redistribute responsibility from individuals to entities better equipped to manage systemic risks. The National Center for Missing & Exploited Children (NCMEC) operates a centralized database of hashed child sexual abuse material (CSAM), and online platforms are legally required to report such content to NCMEC promptly upon discovery.[170] The NCMEC database then feeds back into tools that companies use to detect and block CSAM proactively across cooperating platforms.[171] This system alleviates some of the burden on victims by shifting monitoring and reporting responsibilities onto the companies that host or distribute CSAM. Similarly, the Federal Trade Commission’s (FTC) Do Not Call Registry enables consumers to invoke their right to opt out of telemarketing calls by registering their phone numbers in a central database once.[172] The onus of compliance then transfers to telemarketers, who must design and implement systems that continuously monitor the database and ensure that the registered numbers are not contacted, requiring no further action by consumers invoking the right.[173]

The General Data Protection Regulation’s (GDPR) “right to be forgotten” also offers valuable lessons for a centralized obscurity system.[174] European Union residents may invoke this right to request the deletion of their personal data from a specific company. Once an individual submits a request, the GDPR mandates that the company delete the individual’s data and notify any third parties to whom the data has been disclosed to do the same.[175] Although this process is decentralized, it creates a network effect of data deletion. A centralized obscurity system can harness the same benefits by mandating that brokers notify upstream suppliers and downstream customers that they may be illegally disseminating victim information.

The existence of these initiatives from the private and public sectors suggests that centralized systems can provide a feasible way to address systemic harms. But these examples also suggest that legal mandates might be necessary when private companies fail to act. Data brokers seem unlikely to develop centralized harm-reduction tools on their own initiative. Their profit model thrives on the mass aggregation, sale, and dissemination of personal information, and they operate with minimal interaction with or visibility to the individuals affected by their practices.[176] Brokers also face less reputational risk because their operations are largely opaque to the public, and their incentives are fundamentally misaligned with user safety.[177] Unlike some online platforms that might be more sensitive to public backlash, brokers have continually profited despite the harmful consequences their data sales have had on individuals. This absence of market-driven incentives makes voluntary coordination among brokers highly unlikely, necessitating regulatory intervention to enforce harm-reduction practices. The next Part addresses how such an intervention should be constructed.

III. Designing a Centralized Obscurity System

I’m like, “Why?”—so much unwanted contact and just more headaches, more calling companies, more procedures to just go through. . . . Are you going to have 200 bookmarks of data brokers?

—Ella

In a world where data brokers exist, greater protection for abuse victims could be achieved through a centralized system designed to ensure their personal information remains inaccessible to those who aim to exploit it. This Part evaluates existing and proposed state and federal regulations to create a version of this system, highlighting their gaps and inefficiencies. Building on these insights, we then outline a statutory framework specifically tailored to safeguard abuse victims. Unlike broad policy prescriptions that neglect practical implementation, this Part emphasizes the necessity of aligning regulatory design with operational feasibility and the policy’s overarching goals. To that end, it explores the technical architecture of our proposed centralized system, illustrating how it can effectively shift the burden of managing obscurity from victims to data brokers while promoting robust oversight and accountability.

A. Limitations of Current and Proposed Interventions

Regulatory proposals to protect abuse victims from the risks of broker industry need not start from scratch. Initiatives at the state and federal levels—especially California’s DELETE Act[178] and the proposed federal DELETE Act[179]—represent important attempts to streamline data removal processes and recognize the burdens placed on individuals.[180] However, both fall short in critical ways, either due to express provisions, omissions, or uncertainties left to future rulemaking. By examining these gaps, this Section lays the groundwork for designing a harmonized, comprehensive statutory framework to offer victims greater protection and more effectively redistribute the burden of achieving obscurity from individuals to data brokers.

1. Common Strengths

The California DELETE Act and the proposed federal DELETE Act aim to address the privacy challenges posed by data brokers by creating centralized systems that simplify how individuals manage their personal information.[181] Both bills provide consumers with a streamlined process to request the deletion or cessation of the sale of their personal data, replacing the current fragmented and burdensome approach of contacting multiple data brokers individually. These efforts acknowledge the need to reduce logistical barriers to achieving obscurity in a complex and pervasive data ecosystem.

Under both acts, data brokers—defined as entities that collect personal information from third-party sources and sell or license it—are the primary covered entities. This excludes first-party data collectors who use information solely for their own business purposes, limiting the scope of regulation. In California, the DELETE Act builds on the definitions and obligations established under the CCPA and the California Privacy Rights Act (CPRA). Brokers must register annually with the California Privacy Protection Agency (CPPA), which administers the state’s centralized deletion portal.[182] Similarly, the federal DELETE Act proposes a nationwide broker registry and centralized opt-out system managed by the FTC.

Both bills share several features aimed at improving consumer privacy and accountability in the data broker industry. They provide a centralized portal for consumers to submit a single request for data deletion or cessation of data sales, shifting some responsibility away from individuals.[183] Additionally, both require brokers to maintain compliance records and undergo audits every three years.[184] Separately, they require the FTC to verify the identity of requesters to guard against fraudulent deletions.[185] These provisions acknowledge the systemic nature of data broker harms and represent a partial shift toward holding brokers accountable.

2. Differences

The California DELETE Act and the proposed federal DELETE Act take different approaches to regulating data brokers, revealing critical strengths and weaknesses when evaluated against the twin goals of protecting victims of brokered abuse and shifting the responsibility for achieving obscurity from individuals to brokers.

One major difference lies in the treatment of public information. The California DELETE Act explicitly excludes publicly available data—such as property ownership records, voter registration, or court filings—from the scope of personal information that must be removed. This exclusion, based on definitions established by the CCPA and CPRA, leaves significant holes that undermine protections for victims.[186] By contrast, the proposed federal DELETE Act does not categorically exclude publicly available information, leaving room for future rulemaking by the FTC to include such data within its scope.[187] This difference could make the federal approach significantly more protective, depending on how the FTC defines the boundaries of “covered” data.

The exclusion of publicly available information under California law is particularly problematic for abuse victims. Abusers frequently exploit publicly accessible records to locate or stalk victims, leveraging details like addresses, phone numbers, or workplace information.[188] Although California’s law provides safeguards for certain types of personal information, excluding publicly available data allows brokers to continue amplifying sensitive details, putting victims at risk.[189] Closing this gap is essential for achieving meaningful obscurity and addressing the systemic exploitation of public records by abusers.

A second difference between the two acts is the compliance timeline for data brokers—neither of which adequately protects victims. The California DELETE Act requires brokers to check the registry every forty-five days,[190] while the federal DELETE Act mandates a thirty-one-day compliance period.[191] Although both laws establish ongoing obligations, these timelines are excessively long for individuals in danger, giving abusers ample time to exploit personal information before it is removed. These delays fail to account for the urgency victims face, particularly in situations of imminent threat.

Such long timelines are technologically unnecessary. Data brokers already operate advanced systems capable of processing vast quantities of information quickly.[192] The centralized registries envisioned by these laws are designed to simplify compliance, and brokers could act on deletion requests within far shorter timeframes—potentially within days, if not hours. By allowing brokers such extended leeway, both laws dilute their effectiveness and maintain the burden on victims to remain vigilant in the interim. Shortening these timelines would not only enhance protections for abuse victims but also hold brokers accountable for leveraging their technological capabilities to ensure privacy and safety.

3. Limitations

Although both the California DELETE Act and the proposed federal DELETE Act make strides in regulating data brokers, they each suffer from critical limitations that undermine their effectiveness for abuse victims. Key gaps—such as the lack of a private right of action,[193] the absence of an appeals process for denied requests, [194] and the failure to impose ongoing duties on brokers to ensure deleted data remains off their systems[195]—leave individuals with limited protection and perpetuate the burden of achieving obscurity.

One of the most significant omissions is the lack of a private right of action. Both laws delegate enforcement to government agencies.[196] This setup forces victims to rely on slow and resource-intensive government investigations to address noncompliance, delaying relief for individuals who might face imminent risks.[197] For victims of brokered abuse, whose safety often depends on immediate action, this reliance can result in prolonged exposure to harm.[198] Allowing individuals to directly sue noncompliant brokers would provide an immediate remedy and serve as a stronger deterrent, encouraging brokers to prioritize compliance.[199]

Another critical omission is the lack of an appeals process for denied deletion requests. Broad exceptions—covering legal obligations, fraud prevention, and First Amendment–protected activities—give brokers considerable discretion to deny people’s requests.[200] But, at the same time, neither bill establishes a clear and accessible mechanism for individuals to challenge such denials.[201] Instead, they defer the issue to future rulemaking by the CPPA and FTC, leaving victims with little recourse in the meantime.[202] This gap is especially damaging for abuse victims, as it forces them to navigate a system where unjustified denials can leave their sensitive information exposed indefinitely. A robust appeals mechanism, one complete with defined timelines and requirements for brokers to justify denials, could empower individuals to contest decisions, reducing delays and enhancing accountability.

4. First Amendment Vulnerabilities

Due to the broad scope of their regulatory frameworks, the California DELETE Act and the proposed federal DELETE Act face substantial First Amendment challenges.[203] These vulnerabilities arise from a combination of the Acts’ universal scope, treatment of publicly available information, and selective targeting of data brokers, which collectively weaken the Acts’ ability to withstand constitutional scrutiny.

One critical issue is the universal application of both Acts to all individuals, regardless of their unique need for protection. While this broad scope is intended to promote consumer privacy, it risks overreach by regulating the dissemination of truthful, at times public, information without distinguishing between individuals facing significant risks—such as abuse victims—and those with lesser privacy concerns. Courts might find this lack of tailoring problematic under the First Amendment, as the laws could be deemed more speech restrictive than necessary to achieve a compelling governmental interest.[204]

Additionally, both Acts selectively target data brokers while excluding other entities that make similar commercial sales of identifiable information, such as social media platforms and e-commerce companies. The statutes narrowly define a “data broker” as “a business that knowingly collects and sells to third parties the personal information of a consumer with whom the business does not have a direct relationship.”[205] This definition excludes platforms like Facebook or Google that collect personal information directly from their users and sell identifiable versions of this information to third parties—a gap that allows entities engaged in significant privacy-compromising activities to evade the law’s mandates.[206] The omission is particularly troubling given that these platforms’ sale of user data can contribute toward the same harms the statutes aim to address, such as enabling harassment, stalking, or other forms of abuse.[207]

This selective targeting raises concerns about underinclusivity, as the laws impose obligations on traditional data brokers while allowing other companies that engage in comparable privacy-compromising behaviors to operate without restriction. Courts have previously scrutinized such regulatory disparities under the First Amendment, particularly when they involve the dissemination of truthful information.[208] By failing to cover entities that compile and sell data dossiers, the statutes risk undermining their own objectives and inviting legal challenges.

Together, these issues highlight the DELETE Acts’ vulnerability to First Amendment challenges. While their broad application reflects a commitment to consumer privacy, their lack of precision and gaping exceptions exposes them to legal risks. By contrast, a more narrowly tailored approach—such as the centralized obscurity system proposed in this Article—that closes problematic gaps can address the harms of brokered abuse more effectively while passing constitutional scrutiny.

B. Regulatory Design

This Section proposes a new statutory scheme to address the pernicious harms posed by data brokers and to protect abuse victims more effectively than existing legislative efforts. While the California DELETE Act and the proposed federal DELETE Act represent meaningful progress, their shortcomings leave abuse victims vulnerable.[209] This proposal builds on their strengths while tailoring protections to the specific needs of abuse victims. By narrowing the scope and integrating technical feasibility into regulatory design, this approach reduces First Amendment vulnerabilities, redistributes the burden of obscurity from individuals to data brokers, and offers more effective and enforceable protections. After introducing the centralized system, this Section explores how to design a system with vulnerable populations in mind—from how it should be accessed, to which entities it should cover, how it should be monitored, and how it should be implemented.

1. The Centralized System: An Overview

This statutory intervention creates a centralized system that requires brokers to take on the technical and logistical burden of complying with victim requests to obscure their personal information. At the heart of the system is a central registry maintaining records of individuals who have invoked their obscurity rights. Data brokers must query this database and take immediate, proactive steps to remove or deidentify covered data. This eliminates the fractured, piecemeal nature of existing mechanisms, replacing them with a single point of coordination and enforcement.

Given the interstate nature of the data broker industry and the reality that victims often cross state lines to escape abusers, a federal framework would be preferable to promote comprehensive and uniform protections.[210] Federal legislation avoids jurisdictional gaps and creates consistency across states, preventing brokers from exploiting discrepancies in state laws. Additionally, tying the framework to existing federal statutes like the Violence Against Women Act (VAWA) and the Safe Connections Act could leverage established definitions and enforcement mechanisms, offering a cohesive legal landscape while enhancing support for victims. This approach could also allow victims to take advantage of the system seamlessly while engaging with other victim-protection services, such as changing their address through a state protection program, obtaining a court order, filing a police report, or seeking support from a domestic-abuse hotline or shelter. These points of interaction provide practical opportunities for victims to assert their rights under this statute without additional procedural burdens.

However, in the face of federal inaction, states could still adopt similar statutory proposals to protect their constituent victims. By tailoring this model to fit within state-level programs—such as existing Safe at Home initiatives or domestic-violence protections—states can achieve meaningful reform and establish localized solutions that address the urgent safety needs of their residents.

2. Invoking the Obscurity Right: Whom & How

The statute should endeavor to protect individuals whose safety and well-being are endangered by data brokers disseminating their personal information. The goal is to craft a narrowly tailored yet inclusive framework that prioritizes the needs of the most vulnerable victims while avoiding unnecessary exclusions. Although sound reasons exist to extend similar protections to other groups, our proposal focuses on people who have experienced particular forms of interpersonal abuse.[211]

By drawing on established federal protections, and perhaps filling gaps through independent definitions, the statute could offer both clarity and flexibility to address emerging forms of abuse. Federal law already creates special protections for vulnerable individuals who experience abuse in intimate or familial relationships, while occasionally extending protections to include interpersonal abuse between strangers. VAWA, for example, targets physical, sexual, or psychological harm by intimate partners,[212] while the Safe Connections Act addresses domestic violence, dating violence, sexual assault, stalking, and sex trafficking.[213] Frameworks like these provide definitions that a statute creating a centralized obscurity process could leverage and incorporate.

However, gaps remain. VAWA’s focus on intimate partner violence excludes victims abused by coworkers, acquaintances, or strangers,[214] while both VAWA and the Safe Connections Act arguably neglect newer forms of interpersonal abuse like doxing and nonconsensual sharing of intimate images, at least in certain circumstances.[215] To address gaps and protect victims in diverse and nontraditional abuse contexts, the statute could define additional qualifying acts that fall outside the purview of existing federal laws. By grounding eligibility in established federal definitions and supplementing them with independently enumerated harms, the statutory intervention could provide both consistency and flexibility.

Once eligibility is defined, the next question is how individuals demonstrate that they qualify for the right to opt out. The statute should adopt self-attestation as the preferred method, whereby victims submit a sworn statement affirming their eligibility without further evidentiary requirements.[216] This approach minimizes barriers to access this protection and empowers victims to invoke their rights without requiring documentation that may be difficult, dangerous, or retraumatizing to obtain. It therefore aligns with the statute’s goal of avoiding the major bureaucratic and emotional burdens that obtaining documentation places on the shoulders of vulnerable victims.

While requiring documentation—such as police reports, protective orders, or affidavits from counselors—might theoretically add a layer of verification, it is not an ideal solution. Victims often face significant hurdles in obtaining these materials,[217] whether due to distrust of law enforcement, safety concerns, or the sheer difficulty of navigating bureaucratic systems while coping with trauma. Imposing documentation requirements would create an inequitable system where only those with the resources and ability to produce proof are protected. Furthermore, documentation requirements disproportionately exclude individuals in emergency situations or those who fear retaliation for seeking help. Privacy protections for abuse victims should minimize, not compound, the labor these individuals must undertake.

Self-attestation does not pose a significant risk of misuse in this context for several reasons. First, if an ineligible individual invokes an obscurity right, this does not prejudice anyone else’s legal interests (unlike scenarios involving shared property like vehicles, where granting access to one party necessarily deprives the other). Second, nonvictims are unlikely to exploit this system at scale because the process still requires submitting a sworn statement attesting to their eligibility, which serves as a deterrent to frivolous claims or insufficiently motivated individuals.

Even if there are some risks of misuse, the risks of imposing barriers to access are significant. Raising the threshold for eligibility could disqualify people who do not fit traditional or clear-cut narratives of abuse. The psychology and dynamics of relational abuse—where roles of victim and abuser may shift or overlap—demand a nuanced approach that errs on the side of accessibility. By allowing self-attestation, the statute would ensure that people in complex or nontraditional abuse situations can still seek obscurity. If concerns about misuse arise after implementation, they can be addressed through periodic reviews of the system’s operations, as is done for other comparable interventions.[218] However, the statute should prioritize removing barriers for victims rather than preemptively creating hurdles based on hypothetical concerns.

3. Covered Brokers & Data

The statute should aim to shift the labor of protecting vulnerable individuals onto the entities profiting from the collection and sale of personally identifiable data. To achieve this goal, the statute should employ broad definitions of “data broker” and “covered data” and impose carefully crafted compliance obligations.

Under this statute, a “data broker” should be defined as “any entity that sells or licenses personally identifiable information (PII) to third-parties in a non-deidentified form, regardless of whether the entity has a direct relationship with the individual from whom the data was initially collected.”[219] This definition ensures coverage of platforms like Google, which collect data directly from users but sell it in various formats to third parties.[220] A narrow definition would create dangerous gaps, allowing entities to avoid accountability by operating under alternative business models or by selling smaller quantities of data.[221] Information that undermines people’s obscurity must be comprehensively purged from brokered datasets to ensure safety.[222]

The statute’s definition of “covered data” should take a similarly expansive approach to include both direct and indirect information. Direct PII encompasses traditional identifiers like names, addresses, phone numbers, emails, and Social Security numbers.[223] However, the statute should also cover indirect data that could endanger victims by providing alternative avenues for harm. For example, records related to family members, employment locations, or roommates can enable abusers to locate or target victims indirectly.[224] To address these risks, brokers should be required to use clustering techniques to identify and act on indirect data,[225] guided by thresholds and methodologies developed by experts at agencies like the FTC or the National Institute of Science and Technology. This would reduce the risk that abusers could target victims through more tangential connections, aligning with the statute’s goal of providing comprehensive protection. [226]

Including indirect data in the statute’s scope would impose a minimal burden on third parties, such as family members, whose information may also be removed. The statute should place the responsibility on brokers, streamlining the process and shifting the burden away from victims and their families.[227] Given that brokers already use clustering techniques for commercial purposes,[228] such as creating consumer profiles and linking datasets,[229] this requirement should be feasible. The statute would simply compel brokers to repurpose their existing tools and expertise toward protecting vulnerable individuals, rather than solely pursuing profit.

Covered data should also include publicly available data to seal gaps that would undermine the statute’s protective goals.[230] Information that is technically public can be weaponized by abusers to locate or harm victims. While public records like voter registrations or property deeds might not be inherently harmful, brokers exacerbate the risks they pose by aggregating and centralizing this information, making it instantly accessible at scale.[231] The statute would not erase public records but rather restore the practical obscurity that once limited their accessibility to abusers.

4. Adherence to a Standard of Care

The statute should impose rigorous obligations on data brokers to enhance protections for abuse victims. Key obligations could include prohibitions on dissemination, proactive monitoring, supply chain accountability, and robust compliance measures.

As a threshold requirement, covered data brokers should be compelled to register with the agency tasked with overseeing the statute’s implementation. This builds on models like those in Vermont[232] and California[233] and could provide regulators with insights into the broker ecosystem, facilitating enforcement and the potential for future regulation.

Brokers should be explicitly prohibited from publishing, selling, or disseminating identifiable data tied to registered victims. Like the Do Not Call Registry,[234] this prohibition would not require outright deletion of data; it would mandate that covered data not be disclosed in any identifiable form. Even pseudonymous data, which risks reidentification, should be restricted, with dissemination permitted only in fully deidentified formats.

The statute should also impose a continuing obligation on brokers to monitor their systems to prevent reemergence of covered data. Automated processes should compare newly acquired data against records of registered victims, removing any flagged information before further dissemination. These monitoring obligations should extend beyond previously deleted data to include new details like updated addresses, phone numbers, or employment information tied to registered victims. Such mandates could provide victims long-term, dynamic protection, which would go far beyond the temporary relief offered by one-time removals.

In addition to refraining from disseminating non-deidentified victim data, brokers should also be required to notify entities in their data supply chain—both those they acquire data from and those they sell data to—when a dataset contains records about a registered victim. Inspired by GDPR principles,[235] this would promote compliance throughout the data ecosystem. For example, if a downstream buyer unknowingly receives covered data, the seller would be legally compelled to inform them to prevent further circulation. This would create a cascading effect that reduces the risk of victim information persisting in the ecosystem.

While the statute could allow certain exceptions for lawful obligations, fraud prevention, or protected First Amendment activities, brokers should be forced to maintain an appeals process for disputes over opt-out requests. In contested cases, brokers should default to concealing the data, notifying the victim of their intent to invoke an exception, and providing an opportunity for the victim to challenge the exception’s applicability. Critically, the burden of proof shifts to the broker to justify the exception, reducing the procedural burden on victims.

Finally, brokers should be required to self-attest to compliance, make their books and systems available for impromptu inspection by governing agencies, and submit regular compliance reports detailing actions taken to honor opt-out requests, including records of flagged, deleted, or deidentified data. These measures would allow agencies such as the FTC or Consumer Financial Protection Bureau (CFPB) to audit broker activities, identify patterns of noncompliance, and enforce penalties where necessary.

5. Implementation

The successful implementation of this statutory regime depends on a robust and well-funded infrastructure, overseen by a competent government agency capable of managing its many responsibilities.[236] The agency should handle critical tasks such as maintaining central registries, verifying opt-out requests, monitoring broker compliance, and adapting to emerging risks in the data broker ecosystem. Two central registries are necessary: one for data brokers and another for individuals opting out of data dissemination. The agency should monitor compliance by requiring brokers to regularly query the registry, auditing their activity logs to detect anomalies, and addressing complaints from victims and brokers reporting noncompliance. Additionally, the agency could invest in ongoing research to refine key processes like hash matching, clustering, and deidentification, ensuring the system evolves alongside advancements in the broker ecosystem.

Given the wide-ranging responsibilities of the implementing agency, substantial funding will be essential. However, relying on government appropriations is impractical in the current political climate, especially at the federal level. To address this challenge, the statute should require brokers to pay tiered registration fees, scaled by their size or the volume of data they handle. These fees could create a sustainable revenue source to support the agency’s work, including audits, enforcement, public education campaigns, and even grants for domestic violence shelters or legal aid organizations that assist victims in navigating the opt-out process. Penalties collected from noncompliant brokers could also supplement this fund. This fee-based funding model, inspired by the Universal Service Fund in telecommunications,can support a comprehensive regulatory framework without relying on direct congressional appropriations.[237]

The statute’s penalty scheme should seek to protect victims and promote compliance from brokers. Civil penalties should escalate with the frequency and severity of violations, ensuring that repeat offenders face increasingly harsher consequences. Equally important is the inclusion of a private right of action, which could empower victims to hold brokers directly accountable. A private right of action should ideally allow victims to recover compensation from brokers for the harm they suffer due to the broker’s noncompliance. Even if victims are unable to sue for damages, however, a private right of action could allow victims to seek injunctive relief from noncompliant brokers—a faster avenue than waiting for agency action and cheaper than a civil suit for damages. If a private right of action proves to be politically or legally unfeasible, the tort system could offer an alternative mechanism for accountability. Though this system is imperfect, brokers who fail to meet their obligations might still be held liable under common-law tort doctrines for breaching a duty of care to victims.

C. Technical Design

This proposed statutory scheme provides the foundation for redistributing the burden of achieving obscurity from individuals to brokers, but its success hinges on the implementation of proper technical infrastructure. Without a reliable and carefully designed system to operationalize these rights and protections, even the best-intentioned law risks regulatory impotence. This Section outlines the technical design considerations of a centralized obscurity system predicated upon a government-maintained database of registered victims and interoperable standards to ensure uniform compliance across a fragmented broker ecosystem. The proposed system aims to avoid the pitfalls of a piecemeal approach to privacy self-management, offering a scalable and resilient centralized pathway to meaningful reform.

1. The Need for a Prescriptive Technical Solution

Allowing brokers to design and implement compliance systems independently would almost certainly lead to inconsistency, inefficiency, and opportunities for brokers to circumvent these rules in bad faith. Data brokers operate within a competitive market where incentives to comply rigorously with privacy protections often clash with motives to maximize profits.[238] Historically, self-regulation in industries with significant public interest has resulted in systems designed with inefficiencies—intentional or not—that become entrenched over time.[239] In some cases, companies purposefully design systems that are cumbersome and opaque to use.[240] Once these systems are in place, the architecture ossifies, and companies exploit the narrative that compliance is expensive and burdensome as a way to resist further regulation or even challenge existing mandates.[241]

A striking example is the ongoing failure to mandate true interoperability in health data across electronic health record (EHR) systems.[242] Despite laws like the Health Information Technology for Economic and Clinical Health Act[243] and subsequent interoperability initiatives, the EHR industry has built fragmented systems full of incompatible standards and data silos.[244] This lack of seamless interoperability is not accidental; it is a calculated feature of self-regulation, aimed at maintaining vendor lock-in and avoiding competition.[245] Similarly, the rollout of the CCPA[246] saw companies implement patchwork compliance mechanisms that confused consumers and created obstacles for consumers exercising their rights under the CCPA.[247] These systems were later cited by industry advocates as evidence of compliance being too complicated or costly,[248] fueling lobbying efforts to water down subsequent enforcement or legislative expansion of the CCPA and other privacy statutes.

By learning from these examples, the proposed statutory scheme prioritizes a centralized technical framework to ensure consistency, eliminate inefficiencies, and prevent evasion. A legislated framework for implementation avoids the pitfalls of industry-designed systems, promoting compliance mechanisms that are transparent, effective, and resistant to weaponized inefficiency.

2. Central Victim Opt-Out Registry

The centralized database lies at the heart of this system, maintaining records of individuals who have opted out of having their personal information distributed by data brokers. Its design must prioritize storing only the minimal information necessary, balancing functionality with privacy and security. The database should hold only the information necessary for brokers to identify and act on relevant records in their own systems—such as hashed combinations of names, dates of birth, Social Security numbers, and aliases—rather than a comprehensive repository of every data point a victim wishes to be removed. By limiting the scope of stored data, the database would be less attractive to hackers while still enabling brokers to meet their obligations. The system should also employ prevailing cybersecurity best practices to further secure information in the database.[249]

3. Data Broker Queries to a Centralized Database

Data brokers would interact with the centralized database via a secure API endpoint, matching records to minimize strain on the central system and protect victim privacy. The challenge lies in allowing brokers to see whether the data they possess matches the data in the central registry without either learning the contents of the central registry or sharing all the personal information they possess with the central registry. Application programming interface (API) queries can fix this issue. API queries can employ advanced cryptographic techniques that would allow brokers to compare the contents of their databases with the content in the central victim registry without exposing the information they possess or accessing information from the registry they do not already have.[250] Some techniques could enable brokers to identify matches without the central victim registry seeing their datasets,[251] while others could ensure the database comparisons occur only in encrypted form.[252] These methodologies would enable brokers to fulfill their compliance obligations without further compromising victims’ privacy.

To accommodate the variability in personal data records—such as nicknames, typos, or alternate spellings—the database should support approximate matching techniques.[253] Algorithms like fuzzy hashing[254] and Levenshtein distance[255] could allow brokers to identify close matches rather than relying on exact matches, promoting comprehensive compliance without forcing victims to list all possible variations of their data. Importantly, these approximate matching methods are compatible with the advanced cryptographic protocols the database could use to ensure API queries do not further compromise victim privacy. Brokers could use approximate matching locally with hashed data to identify variations without revealing their full dataset or accessing unrelated records in the central database. This would ensure that variability in data formats does not impede compliance while maintaining robust privacy protections.

Each interaction between a broker and the central database would be logged, capturing timestamps, query metadata, and the broker’s unique identifier. These activity logs would create transparency and accountability, enabling the central database to monitor compliance. Alongside these logs, brokers should also submit compliance reports detailing queries conducted, matches identified, and actions taken, such as records deleted or deidentified. These broker-generated compliance reports would allow the central database to audit broker activities and identify discrepancies or patterns of noncompliance to reinforce the integrity of the framework and the protection of victims’ data.

To streamline the process of pushing registry updates to data brokers, the central database could offer webhook integration.[256] Brokers could subscribe to receive notifications when a registered victim updates or expands their covered data. These notifications would not disclose sensitive information but instead include a broker-specific reference ID and a directive to re-query the database. This approach would foster efficient compliance without exposing unrelated victim data.

4. Identifying Covered Victim Data

Brokers, upon receiving hashed identifiers for individuals who have opted out, should use these hashes to locate and remove or deidentify records containing direct PII, such as Social Security numbers, names, email addresses, and dates of birth. The ability to pinpoint an individual often arises from a combination of elements, such as a name paired with a date of birth or an email address tied to a Social Security number.[257] Using the provided hashed values, brokers could deploy automated matching algorithms to accurately locate and expunge these direct identifiers, ensuring victims’ key identity markers are no longer accessible within their systems.

However, obscurity cannot be achieved by removing direct PII alone. Abusers are often intimately familiar with their victims and can exploit otherwise vague or innocuous data to harm them.[258] To provide meaningful victim obscurity, brokers should also be compelled to identify and address indirect data points that, while not explicitly identifying an individual, could still expose them to harm.[259] Indirect data might include records tied to family members, roommates, or frequent contacts—information that an abuser could exploit to track or target a victim.[260] For instance, even if a victim’s personal address is removed, their residential location could be revealed through records associated with a roommate.

To achieve this kind of comprehensive coverage, brokers could employ the hashed identifiers from the central database as anchor points in their datasets to locate and address indirect or nonobvious data risks. By analyzing patterns and associations, such as shared addresses, linked phone numbers, or overlapping network connections, brokers could identify records indirectly tied to victims who have opted out. For example, if a hashed email address corresponds to a victim, brokers could identify other accounts registered at the same physical address or other individuals linked through shared data points. The government agency should also establish clear, actionable thresholds for clustering proximity, ensuring that brokers strike a balance between privacy and technical feasibility without overreaching into unrelated data.

Placing the burden on brokers to identify and remove indirect data is both practical and justified. Brokers have unparalleled access to vast quantities of data, advanced analytical tools, and the technical expertise required to perform this task.[261] Victims, in contrast, lack both the resources and the visibility into the complex networks of data maintained by brokers, which limits their ability to achieve meaningful obscurity on their own. Moreover, brokers already use sophisticated clustering techniques for commercial purposes, such as building consumer profiles and linking related data across datasets.[262] Applying similar methods to identify indirect information tied to victims is not only feasible but also fair given brokers’ role in undercutting victim obscurity.

Brokers must also prevent obscured data from reentering their systems. Newly ingested datasets should be automatically compared against hashed identifiers already in their possession. If a match with previously removed information is detected, the system should trigger automatic obscurity workflows and notify the central database. This would promote ongoing compliance and protect victims from reemerging risks over time.

5. Deidentification Standards

While deletion of personal data is a powerful tool for achieving victim obscurity, it is not the only means of protecting individuals. In some cases, deidentification could serve as an alternative that balances the need to conceal victims’ data with brokers’ business interests. To be a viable alternative to deletion, deidentification must ensure that data is irreversibly unable to be linked to any individual and incapable of reidentification through direct or indirect methods.[263] This requires adhering to rigorous benchmarks, such as differential privacy standards, which introduce controlled randomness to obscure individual data points while maintaining the statistical integrity of datasets. In reality, though, especially with machine learning, deidentification is often imperfect. The question then becomes: How much deidentification should the law require?

By providing enforceable guidance on acceptable deidentification practices, government agencies could promote consistency across the industry and account for technological advancements that might otherwise render older techniques obsolete. Given that abusers often possess intimate knowledge of their victims,[264] the standards must be designed with the utmost care. Well-informed abusers may be able to reidentify information with fewer specific data points than the average person.[265] So, brokers must ensure that deidentification is robust enough to foil the most dedicated and sophisticated abusers and the government should periodically audit broker deidentification methodologies to ensure brokers are employing the latest state of the art practices.[266] To further streamline the process and improve compliance, the FTC or a similar state agency could offer a centralized validation tool or API that brokers could use to test their deidentification methods against established benchmarks. This tool could promote deidentification practices that are robust, consistent, and aligned with regulatory expectations, providing both accountability and operational clarity.

6. Standards Development Process

The development of technical standards for this framework is just as important as the implementation of the framework itself. A well-structured standards development process not only promotes the system’s technical efficacy but also lends legitimacy and trust to its implementation. To this end, convening a diverse and knowledgeable standard-setting body is paramount. This body should include technical experts, privacy advocates, industry representatives, and government staff, cultivating a balanced approach that reflects the interests of all stakeholders while prioritizing victim protection and privacy. The technical community’s work to mitigate the misuse of Apple AirTags for stalking highlights the importance of involving subject-matter experts.[267] These experts would bring critical insights into how decisions involving technical design have real-world outcomes and can also anticipate potential risks and challenges.

Open standards should be adopted for API protocols, data formatting, clustering, deidentification, and cryptography to ensure cross-industry interoperability. The stakes for this process are particularly high given the immense resources and coordination required to build such a system. Once implemented, the framework will likely become entrenched, making significant redesigns or reversals exceedingly difficult. This reality underscores the importance of designing a solution that is robust and future-proof. An open, transparent, and inclusive standards development process safeguards against industry capture or arbitrage while ensuring that the system’s design is robust, fair, and adaptable to future challenges.

IV. Negotiating First Amendment Challenges

I believe that our concept of records and what needs to be public is not quite keeping up with the pace of technology. What these brokers are offering is not just something that you could go to the courthouse and get; it’s like an aggregation of everything that I didn’t necessarily provide . . . .

—Ella

Implementing a centralized obscurity system for abuse victims entails not only legislative and technical challenges but also constitutional ones. Even if legislators aspire to address brokered abuse, they might fear that constitutional doctrine will thwart their efforts. The regulation of information flows inevitably awakens the First Amendment Balrog.

Data brokers would have lawmakers and the public believe that laws—like the one proposed in this Article—regulating publicly available information face wholesale invalidation, or at the very least must face strict scrutiny.[268] While this Part ultimately maintains that a centralized obscurity proposal should survive even under strict scrutiny, along the way it also challenges the purportedly unavoidable assumption that the First Amendment even requires laws regulating brokers’ use of publicly available information to clear such a hurdle.

First Amendment analysis can be broken into two cascading inquiries: coverage and protection.[269] The coverage inquiry determines whether the First Amendment is even in play, while the protection inquiry subsequently assesses the law’s constitutionality when coverage is established.[270] Asserting that the First Amendment “covers” particular conduct means that First Amendment analysis is required to determine the constitutionality of a law regulating such conduct.[271] Asserting that the First Amendment “protect[s]” such conduct means that the law is unconstitutional.[272] This Part begins by evaluating the coverage inquiry, casting skepticism on the presumption of First Amendment coverage for the regulation of brokers’ sale of publicly available information. It then moves to the protection inquiry to contemplate whether such speech, if covered, is commercial or noncommercial—that is, whether it warrants protection under intermediate or strict scrutiny. Finally, even if subject to strict scrutiny, this Part argues that the proposed centralized obscurity system would survive that standard despite its regulation of publicly available information.

A. Constitutional Coverage: Data Brokers as Navigational Maps?

While the Supreme Court has explained that First Amendment coverage should adapt to evolving media of communication,[273] the data economy raises new questions about what activities the First Amendment covers.[274] As Robert Post theorizes, “First Amendment coverage is triggered by those forms of social interaction that realize First Amendment values . . . [and] extends to [media] that realize First Amendment values.”[275] Scholars have spilled much ink over the animating values of free speech, often centering the protection on three general ideals:[276] (1) marketplace of ideas,[277] (2) individual autonomy,[278] and (3) participatory democracy.[279] Any constitutional challenge to our proposal should begin by assessing whether data brokerage represents a medium that realizes First Amendment values, and thus warrants coverage.[280]

Regardless of where one locates free-speech values, “listener-based educative theory underlies much First Amendment doctrine.”[281] In the context of regulating brokers, listeners’ rights are particularly salient. Listeners’ rights go hand in hand with access to information,[282] and data brokers market themselves as the keyholders to the Library of Alexandria.[283] However, marketplace-of-ideas and participatory-democracy theories of free speech view listeners’ rights in meaningfully different ways that affect the coverage inquiry.

Jane Bambauer contends that data’s potential to inform justifies its classification as speech.[284] According to Bambauer, the coverage question is not whether data is speech in a metaphysical sense, but rather whether the regulation “deliberately interferes with an individual’s effort to learn something new . . . .”[285] In her view, First Amendment coverage should extend to laws that “target[] information-gathering for the very purpose of disrupting it.”[286] While some courts have effectively adopted this view,[287] this coverage analysis arguably privileges a particularly expansive marketplace-of-ideas theory,[288] often to the detriment of public discourse.[289]

Bambauer’s scientific-method framing offers a compelling take on a marketplace-of-ideas theory,[290] but contemporary courts might be increasingly concerned that this will lead to coverage creep and sanitize First Amendment values. Regulating the public’s access to information might not always trigger First Amendment scrutiny under a participatory-democracy view of the First Amendment. Post stresses the importance of the relationship between speaker and listener.[291] To truly serve First Amendment values, he argues, media of communication must do more than “facilitate the communication of particularized messages,” and “the facilitation of communication is not by itself a sufficient reason for social conventions to be valued by the First Amendment.”[292] Under a Postian participatory-democracy theory of free speech, data dossiers might not receive First Amendment protection. Akin to how navigation charts communicate “monologically to their audience,” data brokers’ dossiers speak monologically to their clientele of private parties.[293] Rote conveyance of personal data functions in a similar fashion to a map or other reference source. The consumer, or audience, “assume[s] a position of dependence” and relies on the data as unadulterated fact.[294] Facts, or information, alone do not necessarily constitute constitutionally salient speech.[295] Unless a speaker imbues such facts with an expressive or communicative “use” to express a message, facts alone might not constitute covered speech.[296] Data dossiers, like navigation charts, arguably function as products that lack the kinds of social interactions that realize First Amendment values.[297]

Rather than focusing narrowly on information flows, Post emphasizes the constitutional salience of public discourse.[298] This notion, too, could affect the coverage analysis for data brokerage. Drawing on Supreme Court doctrine,[299] Post raises the “paradox of public discourse,”[300] which posits that public discourse can only perform its constitutional function “if it is conducted with a modicum of civility.”[301] Although demanding civility may constrain speech, sufficiently abusive and alienating public discourse could lead individuals to recoil from engaging in public discourse to influence the construction of public opinion.[302] If incivility is left to fester, public discourse will fail to foster a sense of legitimacy and participation, and the rationale for safeguarding the principle will wane.[303] It is precisely this line of thought that leads Post to the conclusion that the “right to be forgotten” is compatible with the democratic function of public discourse.[304]

Sometimes when you wield a constitutional hammer, everything looks like a nail. And no constitutional right possesses more social and rhetorical power than the First Amendment and freedom of speech.[305] Frederick Schauer refers to this phenomenon as First Amendment magnetism.[306] First Amendment magnetism characterizes the “accelerating attempt to widen the scope of First Amendment coverage to include actions and events traditionally thought to be far removed from any plausible conception of the purposes of a principle of free speech.”[307] However, in an age of rapid First Amendment expansionism, some courts might scrutinize the coverage question to avoid First Amendment creep serving as a tool of deregulation.[308]

It is precisely these deregulatory “perils of Volokhner” that underpin Neil Richards’s contention that privacy regulation and speech regulation need not be in tension.[309] Richards challenges the assumption that information flows constitute speech and therefore fall within the ambit of the First Amendment.[310] In his view, such an absolutist approach to First Amendment coverage fails to adequately question the “constitutional metaphysics of ‘speech.’”[311] Calling “things ‘speech’ or ‘not speech’” might spike judicial anxiety,[312] but courts might be persuaded by the chorus of scholars calling on them to police the boundaries of coverage given the First Amendment’s deregulatory expansion.[313] While privacy must be squared with First Amendment interests, privacy often gets the short end of the stick.[314]

First Amendment questions raised by the digital age invite us to set aside our casebooks and let more elemental constitutional inquiries come to the fore.[315] Even if some contemporary doctrine suggests that data dossiers might be covered, courts should interrogate whether such a conclusion serves the First Amendment’s animating values. Laws like the one proposed in this Article force us to reckon with the costs of First Amendment expansionism, yet they might also provide an opportunity to pump the brakes and demand greater introspection on how constitutional coverage reflects socio-constitutional values.

B. Constitutional Protection: The Commerciality Conundrum

Given the expansion of First Amendment coverage,[316] courts may well extend coverage to data brokers’ sites. Presuming coverage, the protection inquiry begins. Before assessing the constitutionality of a law regulating the dissemination of abuse victims’ data, courts would need to determine the proper level of constitutional scrutiny. If they deem the dissemination of abuse victims’ data to be noncommercial speech, the law must survive strict scrutiny rather than the intermediate scrutiny applied to commercial speech.[317]

1. Commercial Speech: Dossiers v. News

Historically, the First Amendment did not protect commercial speech.[318] However, the Court determined in Virginia State Board of Pharmacy v. Virginia Citizens Consumer Council, Inc. that commercial speech warrants constitutional protection, albeit lesser protection than noncommercial speech.[319] The Court laid out the contours of this diminished protection in Central Hudson, articulating a four-part test.[320] First, commercial speech “must concern lawful activity and not be misleading.”[321] If the speech clears this initial threshold, then the state may only regulate if (1) the “government interest is substantial,” (2) the regulation “directly advances the governmental interest asserted,” and (3) the regulation is “not more extensive than is necessary to serve that interest.”[322] While the Court in Sorrell v. IMS Health Inc. hinted at heightened scrutiny for content-based and viewpoint-based regulation of data flows,[323] the Court’s silence on the contours of this heightened scrutiny has often led lower courts to continue to apply a version of the Central Hudson test.[324]

Meanwhile, noncommercial speech is at the core of canonically protected speech, and once information has entered the public sphere, the First Amendment generally precludes the government from restricting its subsequent use.[325] Once information had been “publicly revealed”[326] and “widely disseminated”[327] to the general public, it became unconstitutional to “restrain its dissemination”[328] and retract it from the “public domain.”[329]

Cox Broadcasting Corp. v. Cohn[330] and Florida Star v. B.J.F.[331] present two key cases in this regard.[332] In Cox, the television station broadcast a rape and murder victim’s name that court records had already “publicly revealed,”[333] while the newspaper in Florida Star published a rape victim’s name derived from a “publicly released police report.”[334] In both cases, the Court held that after the information entered “the public domain,” the First Amendment protected the use of that information.[335]

The Court came to a similar conclusion in Smith v. Daily Mail Publishing Co.,[336] even though the information came from a nongovernmental source. In Smith, two reporters learned the name of a teenage boy who killed his classmate from individuals present at the crime scene.[337] Following the airing of the boy’s name by several radio stations, newspapers printed the name and were indicted under a state law that made it a crime to publish the names of juvenile arrestees without a court’s written approval.[338] The Court held that the First Amendment prohibited the state from punishing the publication of the information.[339] Despite recognizing that prior cases involved the governmental release of information, the Court downplayed this distinction, explaining that the public “cannot be made to rely solely upon the sufferance of government to supply it with information.”[340] The information’s source did not matter as much as the fact that the information had already entered the public domain.[341]

Applying this doctrinal backdrop to this Article’s central proposal, does data brokers’ commercial dissemination of personal data warrant protection as commercial speech or noncommercial speech? At least according to one court, the answer might hinge on whether data brokers serve a newsgathering function that informs the public.[342]

On the one hand, data brokers collect information from diverse sources, collate it, and share it publicly. They charge for access to this information, but the New York Times and countless other publications also disseminate information for profit. Framed in this way, data brokers’ sale of personal information might seem eligible for protection as noncommercial speech, akin to that enjoyed by newspapers. Courts might doubt, however, that the private sale of data dossiers represents a journalistic endeavor. While private dossiers have the potential to inform, they are rarely (if ever) “an effort to engage public opinion.”[343]

The court in Brooks v. Thomson Reuters Corp.[344] expressed these very doubts in a case related to brokered dossiers:

Thomson Reuters is not a journalist performing a “public benefit” by making Plaintiffs’ personal information available to the public. Rather, the company’s dissemination of this information only benefits the private parties who purchase the [company’s] dossiers. All the other cases cited by Thomson Reuters to suggest that there is no privacy right in speech derived from public records are similarly inapposite because they involve journalists disclosing publicly available information to the general public.[345]

The Brooks court draws an appealing distinction between brokers selling data dossiers to private parties and journalists disclosing information to the general public.[346] The Supreme Court drew a similar distinction in Dun & Bradstreet, Inc. v. Greenmoss Builders, Inc.,[347] where a nonmedia information distributor sought the same First Amendment protections as media defendants in defamation actions.[348] The Court found that the sale of credit reports was “speech solely in the individual interest of the speaker and its specific business audience” and was “solely motivated by the desire for profit.”[349] Accordingly, the speech did not address a matter of public concern and received only the diminished protection afforded to commercial speech.[350] Likewise, data brokers—the largest of which are credit reporting agencies[351]—sell dossiers purely from a place of financial self-interest.[352]

Data brokers’ central practice is to sell a product to private parties, not to engage public opinion as a journalistic purveyor of information. Common sense counsels us to look past any journalistic façade that brokers might suggest. Despite the industry’s claims to the contrary, it is anything but clear that brokers should receive noncommercial speech protection rather than the diminished protection afforded to commercial speech.

2. Noncommercial Speech: Passing Strict Scrutiny

Even if courts determine that data brokers’ dissemination of abuse victims’ information—especially publicly collected information—should undergo strict scrutiny, this Article’s central proposal meets this constitutional bar.

To pass strict scrutiny, the government must first demonstrate a compelling government interest.[353] Public health and safety is a classic example of a compelling government interest,[354] and the protection of abuse victims from the primary harms of brokered abuse certainly fits within the ambit of these core governmental concerns.[355]

The government also has a compelling interest in protecting victims from the secondary harms of brokered abuse, including the pressure they feel to withdraw from the public sphere. The privacy and expressive interests of abuse victims too often go unnoticed, as with other marginalized communities.[356] While our proposal might implicate the First Amendment rights of data brokers and those who benefit from information capitalism, obscurity rights might also empower abuse victims to engage in First Amendment activity, rather than suffer the chilling effect of withdrawing from society in hopes of safety.[357]

For abuse victims, privacy might be a necessary precondition for self-expression.[358] While First Amendment doctrine does not generally employ balancing tests,[359] we need not ignore the competing First Amendment interests at play, including their relation to democratic self-governance.[360] Indeed, abuse victims might even refrain from voting—the fundamental right undergirding participatory democracy—for fear that data brokers will scrape their information from public voting rolls and make them readily accessible to their abusers.[361] There is surely a formidable governmental interest in ensuring all citizens feel safe exercising their right to vote. Ultimately, a strict scrutiny analysis will likely hinge on whether the regulation is narrowly tailored rather than whether a compelling government interest exists,[362] but it is vital to foreground the stakes at play here.

Turning to whether our proposal is narrowly tailored, brokers would likely challenge it as both overinclusive and underinclusive.[363] The former will likely rest on a claim that our proposal restricts more speech than necessary to advance its aim of protecting victims from brokered abuse. Brokers might point to the broad definitions of “data broker” and “covered data” to demonstrate the overinclusive sweep of the regulation. While it is true that the proposal includes capacious definitions, it does so to effectively achieve the aim of protecting victims from brokered abuse. To promote safety through obscurity, the regulation must cover the entire supply chain to minimize the potential for data leakage, especially given the potentially severe consequences of such a leak for abuse victims. The same logic justifies expansive coverage of data, given how indirect data might allow determined abusers to locate victims through a proxy. Narrower measures would leave abuse victims at continued risk.

The sweep of our proposed statute is significantly limited by its verification requirement and aggregated, deidentified data carveout. Brokers have raised the lack of verification requirements to argue existing nondisclosure laws are overinclusive.[364] Our proposal, however, restricts access to the centralized obscurity system remedy in two ways. First, the proposal would limit protected persons to those who have experienced specific forms of abuse.[365] Second, the proposal implements a self-attestation regime where victims submit sworn statements affirming their eligibility to access the centralized obscurity system. This dual-layered approach balances the competing need to provide abuse victims unencumbered access to the system’s protections while also ensuring the system adequately limits this obscurity remedy to abuse victims. The proposal also limits covered data to PII, carving out aggregated, deidentified data entirely. Bulk, deidentified transactions do not meaningfully implicate abuse victims’ safety, and they are core to the lucrative marketing and advertising data economy. Therefore, the proposal covers data that meaningfully implicates abuse victim safety while balancing brokers’ business interests.

Brokers have also argued that laws solely regulating commercial disclosures of personal data are underinclusive because public agencies can often still disclose the same personal data.[366] While our proposed statute would not restrict all governmental disclosures of covered personal data, a regulation need not be perfectly tailored to pass strict scrutiny.[367] Here, the functional aim of the regulation is practical obscurity for abuse victims.[368] Data brokers provide frictionless personal data dossiers as a service. The same publicly available personal data may be accessed through FOIA requests, but such processes require tailored requests—often requiring specification of the desired data and the agency that should field the request—and take time to process. Convenience is as central to the product as the data itself. The regulation does not strive to prevent all access to abuse victims’ data but rather stem the tide of abuse, and re-abuse, that arises from instantaneous digital access to troves of frequently refreshed personal data at the click of button for a nominal expense.

In a related domain, the Supreme Court has recognized the significance of practical obscurity, coining the term itself in United States Department of Justice v. Reporters Committee for Freedom of the Press.[369] Writing for the Court, Justice Stevens echoed Ella’s concern about a new age of instantaneous access to collated public information:

Plainly there is a vast difference between the public records that might be found after a diligent search of courthouse files, county archives, and local police stations throughout the country and a computerized summary located in a single clearinghouse of information.[370]

In recognizing this, the Court seemed to consider the medium of dissemination to be as important as the information itself when determining whether “the compilation of otherwise hard-to-obtain information alters the privacy interest implicated by disclosure of that information.”[371] The question is less about the limitation of access to public information and more about the ease of access. Similarly, here, while the regulation would still allow for targeted data requests from public agencies, this reversion to practical-obscurity status quo—where discrete information must be accessed through FOIA requests to specific public agencies—materially advances the government’s compelling interest in protecting victims from abuse.

At bottom, under any application of the First Amendment, this Article’s proposal passes constitutional muster. The First Amendment does not serve as an impenetrable shield for an industry whose business model amplifies danger rather than democratic discourse. When regulation targets the near-frictionless compilation and dissemination of information that places abuse victims in harm’s way, it vindicates rather than violates core First Amendment values.

Conclusion

Brokered abuse represents a fundamental failure of privacy law—an abdication of policymaker responsibility to prioritize human safety over corporate profit. Victims of abuse should not have to navigate an insurmountable maze of data broker opt-out processes to achieve the basic security of online obscurity. This Article underscores the urgent need for an enforceable, centralized obscurity system that redistributes the burden of achieving obscurity from the victims to the data brokers profiting off their vulnerability. By mandating an obscurity system that leverages corporate insights into the informational ecosystem and sophisticated data-processing technologies, regulatory intervention can provide a sustainable solution that promotes victim safety without retraumatizing them.

However, any regulatory intervention must be designed with constitutional resilience in mind, particularly in the face of inevitable First Amendment challenges. The broker industry would surely argue that restrictions on the dissemination of data dossiers, particularly their publicly available components, violate brokers’ right to free speech. To ensure a legally durable regulatory solution to brokered abuse, policymakers must craft a system that is narrowly tailored to achieve the compelling government interest of protecting individuals from stalking, harassment, and violence.

Looking ahead, privacy law should do more to center victims of abuse. If lawmakers fail to consider their unique vulnerabilities, the cycle of harm will only deepen, leaving countless people at risk. Our proposal is no panacea, but it offers regulators a concrete measure to begin addressing the harms of brokered abuse.

Copyright © 2026 Chinmayi Sharma*, Thomas E. Kadri** & Sam Adler***

           *     Associate Professor of Law, Fordham Law School.

          **     Associate Professor, University of Georgia School of Law; Affiliate Faculty, University of Georgia Institute for Women’s & Gender Studies; Legislative & Policy Director, Clinic to End Tech Abuse at Cornell University.

         ***     J.D. Candidate, Fordham Law School. For feedback on earlier versions of this project, we thank Ariana Aboulafia, RonNell Andersen Jones, Jane Bambauer, Elettra Bietti, Hannah Bloch-Wehba, Ryan Calo, Ignacio Cofone, Julie Cohen, Nicki Dell, Amy Gajda, Michael Goodyear, Yael Grauer, Nikolas Guggenberger, Woodrow Hartzog, Mike Hintze, Leigh Honeywell, Ido Kilovaty, Anne Klinefelter, Mark Lemley, Lyrissa Lidsky, Andrew Miller, Christopher Morten, Mark Nottingham, Paul Ohm, Natalia Pires de Vasconcelos, Robert Post, Chris Riley, Ani Satz, Evan Selinger, Scott Skinner-Thompson, Olivier Sylvain, Eugene Volokh, Rachel Vrabec, Ari Waldman, George Wang, Rebecca Wexler, Felix Wu, and Carly Zubrzycki, as well as other participants at the Consumer Law Scholars Conference at Boston University School of Law, Freedom of Expression Scholars Conference at Yale Law School, Privacy Law Scholars Conference at UCLA School of Law and at Northeastern University, UGA-Emory Faculty Workshop, UGA Institute for Women’s Studies Lunch Series, University of North Carolina School of Law Faculty Workshop, Southeastern Junior/Senior Legal Scholars Conference, Journal of Free Speech Law Symposium on Media and Society After Technological Disruption, Junior Law & Tech Scholars Workshop, and UCLA School of Law Institute for Technology Law & Policy Panel on Anonymity and Tech. Authors are listed from most to least Kafkaesque.

       [1]. Interview with Ella (May 31, 2022) [hereinafter Ella Interview]. All subsequent quotations and statements related to Ella’s story are from this interview and will not be cited repeatedly for readability. To protect her anonymity, Ella is a pseudonym.

       [2]. See Thomas E. Kadri, Brokered Abuse, 3 J. Free Speech L.137, 138–39 (2023); Sam Adler, Thomas E. Kadri & Chinmayi Sharma, Brokered Violence: Safety for Sale in the Free Marketplace of Data, Lawfare (Aug. 8, 2025), https://www.lawfaremedia.org/article/brokered-violence--safety-for-sale-in-the-free-marketplace-of-data [https://perma.cc/EE9C-3APW].

[3]. See, e.g., Danielle Keats Citron, Mainstreaming Privacy Torts, 98 Calif. L. Rev. 1805, 1817–19, 1834–35 (2010) (discussing the physical harms that can be associated with information disclosures made by data brokers and other online platforms); Kaveh Waddell, How FamilyTreeNow Makes Stalking Easy, Atlantic (Jan. 17, 2017), https://www.theatlantic.com/technology/archive/2017/01/the-webs-many-search-engines-for-your-personal-information/513323 [https://perma.cc/854H-SBBL] (reporting on how brokered data can facilitate stalking).

[4]. See Danielle Keats Citron, A New Compact for Sexual Privacy, 62 Wm. & Mary L. Rev. 1763, 1788–89 (2021);Margaret B. Kwoka, FOIA, Inc., 65 Duke L.J. 1361, 1376–1401 (2016); David E. Pozen, Transparency’s Ideological Drift, 128 Yale L.J. 100, 125 (2018); Theodore Rostow, What Happens When an Acquaintance Buys Your Data?: A New Privacy Harm in the Age of Data Brokers, 34 Yale J. on Regul., 667, 669 (2017).

       [5]. Amy Gajda, Seek and Hide: The Tangled History of the Right to Privacy 231–41 (2022); Andy Z. Wang, Network Harms, 91 U. Chi. L. Rev. 2093, 2094–95 (2024); Danielle Keats Citron, Reservoirs of Danger: The Evolution of Public and Private Law at the Dawn of the Information Age, 80 S. Cal. L. Rev. 241, 246–51 (2007) (discussing the relationship between private-sector databases and commercial data brokers); Citron, supra note 4, at 1788.

       [6]. Chris Jay Hoofnagle, Big Brother’s Little Helpers: How ChoicePoint and Other Commercial Data Brokers Collect and Package Your Data for Law Enforcement, 29 N.C. J. Int’l L. & Com. Regul. 595, 595 (2004); Citron, supra note 4, at 1789.

[7]. See generally Woodrow Hartzog & Evan Selinger, Surveillance as Loss of Obscurity, 72 Wash. & Lee L. Rev. 1343 (2015) (exploring how obscurity as a privacy interest rests on the difficulty and probability of discovering or understanding information). See also Ignacio N. Cofone & Adriana Z. Robertson, Privacy Harms, 69 Hastings L.J. 1039, 1053 (2018) (reminding us that “informational privacy is really about levels of privacy,” rather than “about having privacy or not”); infra Part I.

[8]. See Ignacio Cofone, Privacy Standing, 2022 U. Ill. L. Rev. 1367, 1403–07 (2022) (interrogating the connection between privacy and physical harm); seealso Danielle Keats Citron & Daniel J. Solove, Privacy Harms, 102 B.U. L. Rev. 793, 831–34 (2022) (offering examples of information disclosures by data brokers and other entities that resulted in violence).

       [9]. See Kadri, supra note 2, at 138.

[10]. See infra Part I.A.1; Kadri, supra note 2, at 150. See generally Cofone, supra note 8, at 1401–07 (outlining how privacy invasions can cause “a distinct set of harms in addition to privacy harms,” including reputational, financial, discriminatory, bodily, and autonomy harms); Danielle Keats Citron, Sexual Privacy, 128 Yale L.J. 1870 (2019) (discussing how networked technologies have facilitated various forms of interpersonal abuse).

[11]. See Daniel J. Solove & Danielle Keats Citron, Risk and Anxiety: A Theory of Data-Breach Harms, 96 Tex. L. Rev. 737, 763 (2018).

     [12]. Mara Hvistendahl, I Tried to Get My Name off People-Search Sites. It Was Nearly Impossible.,Consumer Reps. (Aug. 20, 2020), https://www.consumerreports.org/electronics/personal-information/i-tried-to-get-my-name-off-peoplesearch-sites-it-was-nearly-a0741114794 [https://perma.cc/BF54-Q4KD].

[13]. Id.

     [14]. Kadri, supra note 2, at 153. Danielle Citron and Daniel Solove have offered a similar argument about economic harms resulting from privacy violations. See Citron & Solove, supra note 8, at 835–86 (discussing how some information disclosures can result in a “loss of productivity or time to deal with privacy violations”). There is a subtle difference between the privacy harm in Citron and Solove’s taxonomy and our conception of secondary harms from brokered abuse. While they rightly focus on how privacy violations can cause financial injuries and then also cause a “loss of quality time” as people struggle to deal with the fallout from those privacy violations, we highlight the emotional and financial burdens caused by the legally constructed process of vindicating one’s privacy rights and otherwise engaging in privacy self-management. Compare id., withinfra Part I.A.2.

[15]. See Ari Ezra Waldman, Privacy’s Rights Trap, 117 Nw. U. L. Rev. Online 88, 91–92 (2022).

     [16]. Daniel J. Solove, Privacy Self-Management and the Consent Dilemma, 126 Harv. L. Rev. 1880, 1880–83, 1888–93 (2013).

[17]. Id. at 1880.

[18]. Seegenerally Ella Corren, Gaining or Losing Control? An Empirical Study on the Real Use of Data Control Right and Policy Implications, 109 Iowa L. Rev. 2017 (2024). See also Solove, supra note 16, at 1883–93; Ella Corren, The Consent Burden in Consumer and Digital Markets, 36 Harv. J.L. & Tech. 551, 564–67 (2023) [hereinafter Corren, The Consent Burden].

[19]. See infra Part I.

[20]. See Woodrow Hartzog, What is Privacy? That’s the Wrong Question, 88 U. Chi. L. Rev. 1677, 1683 (2021) (lamenting that few privacy laws “are aimed at disrupting power disparities between people and companies” or “protecting individuals from harassment”).

[21]. Seegenerally Danielle Keats Citron, Hate Crimes in Cyberspace (2014) (documenting the myriad burdens experienced by victims of technology-enabled abuse).

[22]. See infra Part II; cf. Citron, supra note 5, at 283–87 (discussing how the operators of private databases are the least-cost avoiders to protect people’s information from hacks and leaks). For foundational work on the law-and-economics concept of least-cost avoiders, see generallyGuido Calabresi, The Costs of Accidents: A Legal and Economic Analysis (1970) andGuido Calabresi & Jon T. Hirschoff, Toward a Test for Strict Liability in Torts, 81 Yale L.J. 1055 (1972).

     [23]. It is important to stress that our proposal is not the sole legal or technological measure that could disrupt the broker industry or address its harms. Other regulatory measures can—and likely should—be pursued in tandem. See, e.g., Danielle Keats Citron, Intimate Privacy in a Post-Roe World, 75 Fla. L. Rev. 1033, 1062–71 (2023) (outlining reforms to curb the corporate collection and sale of intimate data); Helen Nissenbaum, Katherine Strandburg & Salomé Viljoen, The Great Regulatory Dodge, 37 Harv. J.L. & Tech. 1231, 1261–64 (2023) (proposing approaches to creating more contextually sensitive and comprehensive privacy laws). We will return to this point at the very end of this Article.

[24]. See generally Fed. Trade Comm’n, Data Brokers: A Call for Transparency and Accountability (2014) (examining the lack of transparency and regulation to address risks posed by the broker industry). See alsoinfra Part I.A.1.

     [25]. Other scholars have proposed centralized processes to regulate information flows in related contexts. Danielle Citron, for example, has suggested that a “one-stop shop for deletion” could apply more narrowly to “intimate data” and any company “holding intimate data.” Danielle Keats Citron, The Fight for Privacy: Protecting Dignity, Identity, and Love in the Digital Age 164–65 (2022); see also Citron, supra note 4, at 1768 (defining “intimate privacy” as being concerned with “information about, and access to, the body, particularly the parts of the body associated with sex, gender, sexuality, and reproduction”). Lauren Willis, meanwhile, has raised the idea of a do-not-track default rule that would bar websites from tracking consumer internet use unless the consumer had signed up for a government-run “Track Me” registry. See Lauren E. Willis, When Nudges Fail: Slippery Defaults, 80 U. Chi. L. Rev. 1155, 1218–19 (2013).

     [26]. S.B. 362, 2023 Leg., Reg. Sess. (Cal. 2023) (California’s DELETE Act); see also infra Part III.A.

     [27]. Legis. B. 602, 109th Leg., 1st Sess. (Neb. 2025) (proposed Data Elimination and Limiting Extensive Tracking and Exchange Act); S.B. 2121, 89th Leg., Reg. Sess. (Tex. 2025); S.B. 1343, 89th Leg., Reg. Sess. (Tex. 2025); H.B. 4, 88th Leg., Reg. Sess. (Tex. 2023); H.B. 121, 2024 Gen. Assemb., Reg. Sess. (Vt. 2024) (vetoed by Vermont Governor June 2024); see also Suzanne Smalley, Delete-Your-Data Laws Have a Perennial Problem: Data Brokers Who Fail to Register, Record (Oct. 17, 2023), https://therecord.media/state--registries-california-vermont [https://perma.cc/3MXN-J6ZD].

     [28]. H.R. 4311, 118th Cong. § 2(b) (2023); see also infra Part III.A.

[29]. See infra Parts III.A.4, IV.

[30]. See Frederick Schauer, The Politics and Incentives of First Amendment Coverage, 56 Wm. & Mary L. Rev. 1613, 1617 (2015); Amanda Shanor, First Amendment Coverage, 93 N.Y.U. L. Rev. 318, 322 (2018); Robert Post & Amanda Shanor, Commentary, Adam Smith’s First Amendment, 128 Harv. L. Rev. F. 165, 166–67 (2015) (“Across the country, plaintiffs are using the First Amendment to challenge commercial regulations, in matters ranging from public health to data privacy.”).

     [31]. See Letter from Philip Recht, Partner, Mayer Brown LLP, to Kesha Ram Hinsdale, Sen., Vt. Gen. Assemb. 4–7 (Apr. 2, 2024).

[32]. See Cal. Civ. Code § 1798.99.86 (West 2025); H.R. 4311 § (2)(a)(1)(A).

[33]. See generally Ari Ezra Waldman, Industry Unbound: The Inside Story of Privacy, Data, and Corporate Power (2021) (interrogating the corporate influence that can weaken privacy legislation).

[34]. See, e.g., Evan Selinger & Woodrow Hartzog, Obscurity and Privacy, in Spaces for the Future: A Companion to Philosophy of Technology 119 (Joseph C. Pitt & Ashley Shew eds., 2018); Woodrow Hartzog & Frederic Stutzman, Obscurity by Design, 88 Wash. L. Rev. 385 (2013) [hereinafter Hartzog & Stutzman, Obscurity by Design]; Woodrow Hartzog & Frederic Stutzman, The Case for Online Obscurity, 101 Calif. L. Rev. 1 (2013) [hereinafter Hartzog & Stutzman, Case for Online Obscurity].

[35]. See generally, e.g., Thomas E. Kadri, Networks of Empathy, 2020 Utah L. Rev. 1075; Kadri, supra note 2; Janet X. Chen, Allison McDonald, Yixin Zou, Emily Tseng, Kevin Roundy, Acar Tamersoy, Florian Schaub, Thomas Ristenpart & Nicola Dell, Trauma-Informed Computing: Towards Safer Technology Experiences for All, 2022 Procs. CHI Conf. on Hum. Factors Computing Sys. 1; Diana Freed, Jackeline Palmer, Diana Minchala, Karen Levy, Thomas Ristenpart & Nicola Dell, “A Stalker’s Paradise”: How Intimate Partner Abusers Exploit Technology, 2018 Procs. CHI Conf. on Hum. Factors Computing Sys. 1.

[36]. See, e.g., Neil Richards & Woodrow Hartzog, Taking Trust Seriously in Privacy Law, 19 Stan. Tech. L. Rev. 431, 444 (2016) (arguing that people cannot adequately make choices to protect their information); Solove, supra note 16, at 1882–83.

     [37]. Genevieve Lakier, The First Amendment’s Real Lochner Problem, 87 U. Chi. L. Rev. 1241, 1241 (2020); see also Evelyn Douek & Genevieve Lakier, Lochner.com?, 138 Harv. L. Rev. 100, 103 (2024); Amanda Shanor, The New Lochner, 2016 Wis. L. Rev. 133.

[38]. See infra Part I.A.2.

[39]. See infra Part II; see also Citron, supra note 25, at 14 (asserting that the broker industry generates $200 billion annually).

[40]. See infra Parts II, III.B–C; see also Citron, supra note 25, at 164–65.

     [41]. Often, but not always. See generally Thomas E. Kadri, Platforms as Blackacres, 68 UCLA L. Rev. 1184, 1222–49 (2022) (discussing First Amendment doctrine governing information that has entered the public sphere); Daniel J. Solove, Access and Aggregation: Public Records, Privacy and the Constitution, 86 Minn. L. Rev. 1137, 1200–17 (2002); Molly Cinnamon, You Have the Right to Be Deleted: First Amendment Challenges to Data Broker Deletion Laws, 9 Geo. L. Tech. Rev. 492 (2025). Courts, too, are no longer avoiding these questions, in part because legislators are calling this a First Amendment question. Seegenerally Kratovil v. City of New Brunswick, 336 A.3d 201 (N.J. 2025); Adler, Kadri & Sharma, supra note 2.

[42]. See infra Part IV.A.

[43]. See infra Part IV.B.

[44]. See infra Part IV.C.

[45]. See id.

[46]. See Citron, supra note 5, at 283–87; sources cited supra note 22.

[47]. See infra Parts II.A.4, II.A.5.

[48]. See infra Parts III.A.1–A.4.

[49]. See infra Part III.C.

[50]. See infra Part IV.

     [51]. This Article focuses on the concept of obscurity within broader privacy law discourse. Obscurity offers victims a more expansive and operational remedy to the harms of data broker enabled surveillance. Seegenerally Hartzog & Stutzman, Case for Online Obscurity, supra note 34; Hartzog & Stutzman, Obscurity by Design, supra note 34; Hartzog & Selinger, supra note 7; Woodrow Hartzog & Evan Selinger, Obscurity: A Better Way to Think About Your Data Than ‘Privacy,’ Atlantic (Jan. 17, 2013), https://www.theatlantic.com/technology/archive/2013/01/obscurity-a-better-way-to-think-about-your-data-than-privacy/267283/ [https://perma.cc/PJ5Y-VME4].

[52]. See generally Waldman, supra note 15 (critiquing the law’s reliance on individualistic privacy protections); Solove, supra note 16, at 1880 (raising concerns about privacy self-management).

[53]. See Cofone & Robertson, supra note 7, at 1049–55; Cofone, supra note 8, at 1367; Citron & Solove, supra note 8, at 830–61; see also Scott Skinner-Thompson, Agonistic Privacy & Equitable Democracy, 131 Yale L.J.F. 454, 456 (2021).

     [54]. See Solove, supra note 16, at 1881.

[55]. See Kadri, supra note 35, at 1078–80, 1118–19 (arguing that empathy should be a guiding principle in regulating tech-enabled abuse).

[56]. See Frank Pasquale, Opinion, The Dark Market for Personal Data, N.Y. Times (Oct. 16, 2014), https://www.nytimes.com/2014/10/17/opinion/the-dark-market-for-personal-data.html [https://perma.cc/M4JQ-CHA5].

[57]. See Kadri, supra note 2, at 142.

     [58]. Kadri, supra note 2, at 138; see also Hartzog & Selinger, supra note 7, at 1355–69 (building out the connection between privacy and obscurity).

     [59]. Citron, supra note 3, at 1817–19, 1834–35; Kadri, supra note 2, at 150.

[60]. See Salomé Viljoen, A Relational Theory of Data Governance, 131 Yale L.J. 573, 588 n.19 (2021); see also Citron, supra note 25, at 14 (asserting that “[t]he data-brokerage industry generates 200 billion dollars annually”). For important early scholarship on brokers, see generally Hoofnagle, supra note 6; Daniel J. Solove & Chris Jay Hoofnagle, A Model Regime of Privacy Protection, 2006 U. Ill. L. Rev. 357, 367 (2006); Citron, supra note 5; Citron, supra note 3. For more contemporary reporting, see Adi Robertson, The Long, Weird History of Companies That Put Your Life Online, Verge (Mar. 21, 2017), https://www.theverge.com/2017/3/21/14945884/people-search-sites-history-privacy-regulation [https://perma.cc/Z9J8-HU9G]; Yael Grauer, What Are ‘Data Brokers,’ and Why Are They Scooping Up Information About You?, Vice (Mar. 27, 2018), https://www.vice.com/en/article/what-are-s-and-how-to-stop-my-private-data-collection [https://perma.cc/2WRK-7KB7].

     [61]. Emile Ayoub & Elizabeth Goitein, Closing the Data Broker Loophole, Brennan Ctr. for Just. (Feb. 13, 2024), https://www.brennancenter.org/our-work/research-reports/closing-data-broker-loophole [https://perma.cc/DZC6-Q734].

[62]. See Justin Sherman, People Search Data Brokers, Stalking, and ‘Publicly Available Information’ Carve-Outs, Lawfare (Oct. 30, 2023), https://www.lawfaremedia.org/article/people-search-s-stalking-and-publicly-available-information-carve-outs [https://perma.cc/EE4V-2XQR]; see also Ashley Kuempel, The Invisible Middleman: A Critique and Call for Reform of the Data Broker Industry, 36 Nw. J. Int’l L. & Bus.207, 210 (2016).

[63]. See Sherman, supra note 62.

[64]. See Kwoka, supra note 4, at 1379–414.

     [65]. Gajda, supra note 5, at 231–41; Ayoub & Goitein, supra note 61.

[66]. See Amanda Levendowski, Resisting Face Surveillance with Copyright Law, 100 N.C. L. Rev. 1015, 1018, 1022–35 (2022); Neil Richards & Woodrow Hartzog, The Pathologies of Digital Consent, 96 Wash. U. L. Rev. 1461, 1485 (2019); Woodrow Hartzog & Evan Selinger, Opinion, Why You Can No Longer Get Lost in the Crowd, N.Y. Times (Apr. 17, 2019), https://www.nytimes.com/2019/04/17/opinion/data-privacy.html [https://perma.cc/TL4A-VS2E].

[67]. See Justin Sherman, How Shady Companies Guess Your Religion, Sexual Orientation, and Mental Health, Slate (Apr. 26, 2023), https://slate.com/technology/2023/04/data-broker-inference-privacy-legislation.html [https://perma.cc/45VZ-5SRL]. See generally Alicia Solow-Niederman, Information Privacy and the Inference Economy, 117 Nw. U. L. Rev. 357 (2022) (calling for privacy law to adapt considering how machine learning enables such inferences).

     [68]. Gajda, supra note 5, at 231–41.

[69]. See Kadri, supra note 41, at 1184–87; Thomas E. Kadri, Digital Gatekeepers, 99 Tex. L. Rev. 951, 977–82 (2021).

[70]. See Andrew Wade, Note, The Clocks are Striking Thirteen: Congress, Not Courts, Must Save Us from Government Surveillance via Data Brokers, 102 Tex. L. Rev. 1099, 1106 (2024).

[71]. See Dave, What Are Data Brokers?, DeleteMe, https://help.joindeleteme.com/hc/en-us/articles/8319769261203-What-are-Data-Brokers [https://perma.cc/3F88-HTND].

     [72]. Urbano Reviglio, The Untamed and Discreet Role of Data Brokers in Surveillance Capitalism: A Transnational and Interdisciplinary Overview, 11 Internet Pol’y Rev. 1, 15 (2022) (figure illustrating eight different risks of data brokers being under-regulated). On the possible benefits of some types of data brokerage, see generallyJennifer Barrett Glasgow, Data Brokers: Should They Be Reviled or Revered?, in TheCambridge Handbook of Consumer Privacy 25 (Evan Selinger, Jules Polonetsky & Omer Tene eds., 2018) (surveying ostensible benefits that brokers bring to the economy, innovation, and consumers).

[73]. Cf. Selinger & Hartzog, supra note 34, at 120, 123–25 (describing how technological infrastructure and innovation remove friction to access personal information and corrode obscurity).

[74]. See The Amy Boyer Case, Elec. Priv. Info. Ctr. (June 15, 2006), https://archive.epic.org/privacy/boyer [https://perma.cc/LS52-FW3J]; Sherman, supra note 62; Adler, Kadri & Sharma, supra note 2.

[75]. See generally Eugene Volokh, Cheap Speech and What It Will Do, 104 Yale L.J. 1805 (1995) (predicting that the “cheap” speech enabled by digital technologies will alter information flows).

[76]. See Citron, supra note 10.

     [77]. Gajda, supra note 5, at 52–66.

[78]. Seeid. at 231–41; Sherman, supra note 62.

     [79]. Hoofnagle, supra note 6, at 595.

[80]. See Chen et al., supra note 35, at 1.

     [81]. Joseph Cox, Candy Crush, Tinder, MyFitnessPal: See the Thousands of Apps Hijacked to Spy on Your Location, Wired (Jan. 9, 2025), https://www.wired.com/story/gravy-location-data-app-leak-rtb [https://perma.cc/KNN2-YXQ7].

[82]. See Justin Sherman, Credit Reporting Agencies Don’t Just Report Credit Scores, Duke Sanford Tech Pol’y Program (Nov. 9, 2022), https://techpolicy.sanford.duke.edu/blogroll/credit-reporting-agencies-dont-just-report-credit-scores [https://perma.cc/R675-DV89].

[83]. See Scottie Andrew, For Abuse Victims, Registering to Vote Brings a Dangerous Tradeoff, CNN (Oct. 27, 2020), https://www.cnn.com/2020/10/27/us/domestic-violence-voting-election-privacy-trnd/index.html [https://perma.cc/M6NZ-46XA].

     [84]. Nicole Froio, Should Abuse Survivors Have to Disappear from the Internet?, Verge (Dec. 6, 2021), https://www.theverge.com/22812890/domestic-abuse-survivors-online-presence-spyware-recommendations [https://perma.cc/KFA3-WZQV].

[85]. See Kadri, supra note 2, at 151.

[86]. Seeid. at 143.

[87]. See Hvistendahl, supra note 12.

     [88]. Solove, supra note 16, at 1880–83; seealso Waldman, supra note 15, at 89–90; Kadri, supra note 2, at 151.

[89]. See Solove, supra note 16, at 1880–83.

[90]. See id. at 1888; Corren, The Consent Burden, supra note 18, at 564–67.

     [91]. Kadri, supra note 2, at 143, 151–54; Solove, supra note 16, at 1880–81.

[92]. See Solove, supra note 16, at 1888.

     [93]. Hvistendahl, supra note 12 (“No two of these convoluted procedures seem to be alike. People who track the problem estimate that it can take from six business days to two weeks of full-time work to delete your data from data brokers’ sites.”).

[94]. Id. (“Some sites asked me to enter a current phone number or email address to remove my data, which felt like extortion. Others asked me to register and create a password to ‘control’ my information, without giving me the option to delete it entirely. A few even required me to pick up the phone, send snail mail, or—get this—fax in my request. Where do you even find a fax machine these days?”).

[95]. See Kejsi Take, Kevin Gallagher, Andrea Forte, Damon McCoy & Rachel Greenstadt, “It Feels Like Whack-a-Mole”: User Experiences of Data Removal from People Search Websites, 3 Procs. on Priv. Enhancing Techs. 159, 166–69 (2022).

[96]. See id. at 167 (noting that by requiring victims to provide information like their legal names, opt-out processes “disproportionately affect[] some people more than others, for example, those who change their name to better fit their gender identity”).

[97]. See Hvistendahl, supra note 12.

[98]. Id.

[99]. See id.

[100]. See id. (“I found my information reappearing online, too. Five months after opting out from one data broker, my profile reappeared. When I clicked on my name, the page showed a satellite photo of a house where I had once lived.”); Yael Grauer, Victoria Kauffman & Leigh Honeywell, Consumer Reps., Data Defense: Evaluating People-Search Site Removal Services 10 (2024) (“As a whole, people-search removal services are largely ineffective. . . . [W]ithout exception, information about each participant still appeared on some of the 13 people-search sites at the one-week, one-month, and four-month intervals.”).

[101]. See Kadri, supra note 2, at 153.

[102]. See Grauer, Kauffman & Honeywell, supra note 100, at 5; Hvistendahl, supra note 12; Take et al., supra note 95, at 170.

[103]. See Privacy Bee, https://privacybee.com [https://perma.cc/BB7H-6STT]; DeleteMe Plans, DeleteMe, https://joindeleteme.com/privacy-protection-plans [https://perma.cc/HA77-3JWC].

[104]. See Grauer, Kauffman & Honeywell, supra note 100; Hvistendahl, supra note 12.

   [105]. For example, the yearly price for DeleteMe starts at $129 per year. See DeleteMe Plans, supra note 103.

[106]. See Hvistendahl, supra note 12; Kadri, supra note 2, at 153.

[107]. See Grauer, Kauffman & Honeywell, supra note 100, at 5; Hvistendahl, supra note 12.

   [108]. Hvistendahl, supra note 12.

[109]. See Kadri, supra note 2, at 153.

[110]. See Hvistendahl, supra note 12.

[111]. See Brittany A. Martin, The Unregulated Underground Market for Your Data: Providing Adequate Protections for Consumer Privacy in the Modern Era, 105 Iowa L. Rev. 865, 867 (2020).

[112]. See Take et al., supra note 95, at 166–70.

[113]. See id. at 171–72.

[114]. See Kadri, supra note 2, at 153.

[115]. See Hvistendahl, supra note 12 (reporting how one victim “started her quest” to remove her data “hoping to distance herself from a traumatizing situation, but instead she was continually forced to relive it” and reshare her story).

[116]. Seeid.

[117]. See Kadri, supra note 2, at 153.

[118]. See Hvistendahl, supra note 12.

[119]. See Corren, The Consent Burden, supra note 18, at 551 (arguing that various privacy laws’ reliance on self-managed consent “enables and legitimizes digital surveillance and other consumer exploitations”). See generally Daniel J. Solove, A Brief History of Information Privacy Law, in Proskauer on Privacy 1-1 (Kristen J. Mathews ed., 2d ed. 2016) (discussing the development of the patchwork of privacy laws in the United States).

[120]. See Kadri, supra note 2, at 142–48; Michael Kans, Data Brokers and National Security, Lawfare (Apr. 29, 2021), https://www.lawfaremedia.org/article/data-brokers-and-national-security [https://perma.cc/8QHR-BMXH].

[121]. See Kadri, supra note 2, at 152.

[122]. See Corren, The Consent Burden, supra note 18, at 551.

[123]. See, e.g., Ga. Code Ann. § 16-5-90 (criminalizing the offense of “stalking”); Cal. Penal Code § 653.2 (criminalizing doxing).

[124]. See Kadri, supra note 2, at 142.

[125]. Id. at 142–43.

   [126]. Cal. Gov’t Code § 6208.1; see Kadri, supra note 2, at 142–43. Without scienter requirements, however, these laws risk running afoul of constitutional protections such as the First Amendment. See Kadri, supra note 2, at 142–43.

   [127]. Kadri, supra note 2, at 144–45.

[128]. See Vt. Stat. Ann. tit. 9, § 2446; Cal. Civ. Code § 1798.99.82.

   [129]. Cal. Civ. Code § 1798.115.

[130]. See Waldman, supra note 15, at 88 (arguing that “the history of using individual [privacy] rights to solve structural problems proves how rights crowd out necessary reform”); Pozen, supra note 4, at 135–41 (contending that soft-touch and targeted transparency mandates have “evolved into a stock substitute for more robust and direct regulation” to protect consumers).

[131]. See, e.g., Cal. Penal Code § 502 (outlawing “[u]nauthorized access to computers, computer systems, and computer data”).

[132]. See Cal. Civ. Code §§ 1798.100–.199.100; Cal. Code Regs. tit. 11, §§ 7000–7304 (collectively, CCPA).

[133]. See Cal. Civ. Code § 1798.140.

[134]. Id.

   [135]. Kadri, supra note 2, at 146.

[136]. See, e.g., Citron, supra note 3, at 1826–28 (discussing the limits of the four major privacy torts to address broker disclosures); Neil M. Richards & Daniel J. Solove, Prosser’s Privacy Law: A Mixed Legacy, 98 Calif. L. Rev. 1887, 1919 (2010) (explaining that disclosing a person’s home address would likely fail to satisfy the “highly offensive to a reasonable person” requirement of the disclosure tort); see alsoGM Agrees to 5-Year Ban on Selling Drivers’ Location Data, Reuters (Jan. 17, 2025), https://www.reuters.com/business/autos-transportation/ftc-bans-gm-disclosing-driver-consumer-data-consumer-reporting-agencies-2025-01-16 [https://perma.cc/G9WB-MCBR].

   [137]. Cal. Civ. Code § 1798.120.

[138]. See Cal. Gov’t Code § 6208.1(b)(1); see also Kadri, supra note 2, at 148–49.

[139]. See Cal. Gov’t Code § 6208.1(a)(2)(B); see also Kadri, supra note 2, at 148–49.

[140]. See Kadri, supra note 2, at 148–49.

[141]. See Kadri, supra note 2, at 142, 153; Solove & Hoofnagle, supra note 60, at 367.

[142].For key conceptual and critical work on informational capitalism, seegenerally Julie E. Cohen, Between Truth and Power: The Legal Constructions of Informational Capitalism (2019); see also Waldman, supra note 33; Amy Kapczynski, The Law of Informational Capitalism, 129 Yale L.J. 1460 (2020).

   [143]. To be clear, we are far from the first to propose connecting privacy and safety. See generally, e.g., Citron, supra note 21; A. Michael Froomkin & Zak Colangelo, Privacy as Safety, 95 Wash. L. Rev. 141 (2020); A. Michael Froomkin, Phillip J. Arencibia & P. Zak Colangelo-Trenner, Safety as Privacy, 64 Ariz. L. Rev. 921 (2022). We hope to build on prior scholarly foundations for this connection and propose a legislative implementation of it.

[144]. See Kadri, supra note 2, at 154.

[145]. Seeid.

[146]. See Solove, supra note 18, at 1881;Corren, The Consent Burden, supra note 18, at 551; Take et al., supra note 95, at 166–70.

[147]. See Take et al., supra note 95, at 166–70.

   [148]. Kadri, supra note 2, at 152–54.

[149]. See Citron, supra note 5, at 290–92.

   [150]. Citron, supra note 5, at 291.

[151]. See sources cited, supra note 22.

[152]. See Citron, supra note 5, at 283–87 (supplementing the fairness argument described in supra text accompanying note 150 with one sounding in economic efficiency).

   [153]. Citron, supra note 5, at 285–86.

[154]. See Citron, supra note 25, at 11–14; Rostow, supra note 4, at 674.

[155]. See Reviglio, supra note 72, at 12 (warning that “once personal information has been packaged, sold and resold, it may live indefinitely in the servers run by the data broker industry”).

[156]. See S.B. 362, 2023 Leg., Reg. Sess. (Cal. 2023) (“Beginning January 1, 2028, and every three years thereafter, a data broker shall undergo an audit by an independent third party to determine compliance with this section.”).

[157]. See generally Fed. Trade Comm’n, supra note 24.

[158]. See id. at 53.

   [159]. SeeHow StopNCII.org Works, StopNCII.Org, https://stopncii.org/how-it-works/ [https://perma.cc/DZM5-WRHH]; see also Brenda Dvoskin & Thomas E. Kadri, Safe Sex in the Age of Big Tech Feminism, 39 Harv. J.L. & Tech. 59 (2025); Thomas E. Kadri, Juridical Discourse for Platforms, 136 Harv. L. Rev. F. 163, 200–01 (2022).

[160]. See About Us, StopNCII.Org, https://stopncii.org/about-us [https://perma.cc/6L27-936F].

[161]. How StopNCII.org Works, supra note 159.

[162]. See Industry Partners, StopNCII.Org, https://stopncii.org/partners/industry-partners [https://perma.cc/ZS6W-4LQV].

[163]. See How StopNCII.org Works, supra note 159.

[164]. See id.

[165]. SeeSuicide, Self-Harm, and Domestic Violence Prevention,Pinterest: Help Ctr., https://help.pinterest.com/en/article/suicide-and-self-harm-prevention [https://perma.cc/3JQ3-MQ9T].

[166]. See Adam Mosseri, Changes We’re Making to Do More to Support and Protect the Most Vulnerable People Who Use Instagram, Instagram(Feb. 7, 2019), https://about.instagram.com/blog/announcements/supporting-and-protecting-vulnerable-people-on-instagram [https://perma.cc/36X6-X5KD].

[167]. See Suicide, Self-Harm, and Eating Disorders Policy,YouTube Help, https://support.google.com/youtube/answer/2802245?hl=en [https://perma.cc/7VZK-J3XP].

   [168]. Kalhan Rosenblatt & Maya Eaglin, Meta Teams up with Snap and TikTok to Address Self-Harm Content, NBC News (Sep. 12, 2024),https://www.nbcnews.com/tech/social-media/meta-teams-snap-tiktok-address-self-harm-content-rcna170838 [https://perma.cc/3LNQ-2BBD].

[169]. Suicide Prevention, Meta: Safety Ctr.,https://about.meta.com/actions/safety/topics/wellbeing/suicideprevention [https://perma.cc/9WYZ-KJB6].

[170]. See 18 U.S.C. § 2258A; see also United States v. Keith, 980 F. Supp. 2d 33, 37–39 (D. Mass. 2013).

[171]. See The Tech Coalition Empowers Industry to Combat Online Child Sexual Abuse with Expanded PhotoDNA Licensing, Tech Coal. (Jan. 27, 2025), https://technologycoalition.org/news/the-tech-coalition-empowers-industry-to-combat-online-child-sexual-abuse-with-expanded-photodna-licensing [https://perma.cc/6PLA-ES4W].

[172]. National Do Not Call Registry, Fed. Trade Comm’n, https://www.donotcall.gov [https://perma.cc/N4AU-2B2P]; Lauren E. Willis, Why Not Privacy by Default?, 29 Berkeley Tech. L.J. 61, 108 (2014).

   [173]. Chris Jay Hoofnagle, Privacy Self-Regulation: A Decade of Disappointment, in Consumer Protection in the Age of the ‘Information Economy’ 379, 380–83 (Jane K. Winn ed., 2006).

   [174]. Regulation 2016/679, of the European Parliament and 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), 2016 O.J. (L 119) ch. 3, art. 17 (EU) [hereinafter GDPR].

   [175]. Id.

   [176]. Fed. Trade Comm’n, Protecting Consumer Privacy in an Era of Rapid Change 68 (2012), https://www.ftc.gov/sites/default/files/documents/reports/federal-trade-commission-report-protecting-consumer-privacy-era-rapid-change-recommendations/120326privacyreport.pdf [https://perma.cc/KY6E-W39M].

[177]. See id.

   [178]. S.B. 362, 2023 Leg., Reg. Sess. (Cal. 2023).

   [179]. H.R. 4311, 118th Cong. (2023).

   [180]. Cal. S.B. 362; H.R. 4311.

   [181]. H.R. 4311; Cal. Civ. Code § 1798.99.86 (added by California’s DELETE Act).

   [182]. Cal. Civ. Code § 1798.99.82(a).

   [183]. Cal. Civ. Code § 1798.99.86(a)(2), (b)(1); H.R. 4311 § 2(b)(1)(A)(ii); H.R. 4311 § 2(b)(1)(B)(i).

   [184]. Cal. Civ. Code § 1798.99.86(e)(1) (added by California’s DELETE Act); H.R. 4311 § 2(b)(2)(C)(i).

   [185]. Cal. Civ. Code § 1798.99.85 (added by California’s DELETE Act); H.R. 4311 § 2(b)(2)(A)(i).

   [186]. See Cal. Civ. Code § 1798.140(v)(2).

   [187]. See H.R. 4311 § 2(b)(2)(A)(ii).

   [188]. Sherman, supra note 62.

   [189]. See Cal. Civ. Code § 1798.140(v)(2).

[190]. Id. § 1798.99.86(c)(1)(A).

   [191]. H.R. 4311 § 2(b)(1)(C)(i).

[192]. See Kuempel, supra note 62, at 219–21.

[193]. See Wade, supra note 70, at 1129–30 (“Because the Delete Act lacks a private cause of action, residents cannot hold non-compliant brokers accountable themselves; they must trust that the California Privacy Protection Agency will do it for them—a needlessly risky bet.”).

[194]. See generally S.B. 362, 2023 Leg., Reg. Sess. (Cal. 2023); H.R. 4311.

[195]. See generally Nicole A. Ozer, Golden State Sword: The History and Future of California’s Constitutional Right to Privacy to Defend and Promote Rights, Justice, and Democracy in the Modern Digital Age, 39 Berkeley Tech. L.J. 963, 1069 (2024) (arguing for more restrictive data protection laws).

   [196]. Specifically, the CPPA and the California Attorney General for the DELETE Act, and the FTC and potentially state attorneys general for the federal DELETE Act. Cal. Civ. Code § 1798.99.82; H.R. 4311 § 2(c); see alsoAnalysis of the California Delete Act (SB 362) – Signed by Governor Newsom into Law, Tom Kemp (Oct. 10, 2023), https://www.tomkemp.ai/blog/2023/10/10/analysis-of-the-california-delete-act-sb-362-signed-into-law [https://perma.cc/J8EU-SB43].

[197]. See, e.g., Citron & Solove, supra note 8, at 822 (“The main benefit of a private right of action in a law is to encourage private enforcement of that law because government agencies often lack the resources to enforce a law rigorously and consistently enough.”); see also Ozer, supra note 195, at 1071 (“[G]overnment enforcers have limited bandwidth and sometimes-conflicting internal interests related to government surveillance and consumer privacy.”).

[198]. See Kadri, supra note 2, at 152.

   [199]. When California considered the CCPA in 2018, Attorney General Xavier Becerra wrote to then Assembly member Ed Chau and Senator Robert Hertzberg emphasizing the need for a private right of action. See Letter from Xavier Beccera, Att’y Gen., to Ed Chau, Assemb., Cal. St. Assemb. & Robert Hertzberg, Sen., Cal. St. Senate (Aug. 22, 2018); see also Peter C. Ormerod, A Private Enforcement Remedy for Information Misuse, 60 B.C. L. Rev. 1893, 1941–46 (2019) (describing advantages of state-law private enforcement remedy for data misuse).

   [200]. H.R. 4311 § 2(b)(2)(A)(ii); Cal. Civ. Code § 1798.99.86.

[201]. See H.R. 4311 § 2(b)(2)(A)(ii); Cal. Civ. Code § 1798.99.86.

[202]. See Cal. Civ. Code§ 1798.99.87.

[203]. See generally Ozer, supra note 195, at 1034–35 (stating that technology companies use First Amendment claims to challenge privacy laws).

[204]. See, e.g., NetChoice, LLC v. Bonta, 113 F.4th 1101, 1121 (9th Cir. 2024) (finding a provision of the California Age-Appropriate Design Code Act likely to fail strict scrutiny, as the state could have “employed less restrictive means to accomplish its protective goals”).

   [205]. Cal. Civ. Code § 1798.99.80; H.R. 4311 § 2(f)(3)(B) (excluding entities with direct relationships to individuals whose data they sell from the definition of data broker).

[206]. See Amnesty Int’l, Surveillance Giants: How the Business Model of Google and Facebook Threatens Human Rights 10 (2019), https://www.amnesty.org/en/documents/pol30/1404/2019/en [https://perma.cc/Q4HN-VPC7]; Fed. Trade Comm’n, A Look Behind the Screens: Examining the Data Practices of Social Media and Video Streaming Services 37 (2024), https://www.ftc.gov/system/files/ftc_gov/pdf/Social-Media-6b-Report-9-11-2024.pdf [https://perma.cc/6KU7-8SWZ]; Kurt Knutsson, How the Delete Act Misses Big Tech Culprits in a Law Designed to Protect Consumers, Fox News (Oct. 19, 2023), https://www.foxnews.com/tech/delete-act-misses-big-tech-culprits-law-designed-protect-consumers [https://perma.cc/E27U-KJGC] (“[T]he concerning culprits of social media companies like Meta’s Facebook and Instagram were given a pass and not included in the Delete Act signed into law by Gov. Gavin Newsom.”).

[207]. Seegenerally Fed. Trade Comm’n, supra note 24.

[208]. See, e.g., Sorrell v. IMS Health Inc., 564 U.S. 552, 564 (2011); see also G.S. Hans, No Exit: Ten Years of “Privacy vs. Speech” Post-Sorrell, 65 Wash. U. J.L. & Pol’y 19, 32–37 (2021) (collecting cases considering challenges to privacy laws post-Sorrell).

[209]. See supra Part III.A.

[210]. See Catherine Stupp, Patchwork of State Privacy Laws Remains After Latest Failed Bid for Federal Law, Wall St. J. (Aug. 27, 2024), https://www.wsj.com/articles/patchwork-of-state-privacy-laws-remains-after-latest-failed-bid-for-federal-law-2a1a020d [https://perma.cc/J5HX-8YN3].

[211]. See Citron, supra note 25, at 164–65 (suggesting that a centralized data-deletion system could protect people when companies collect their “intimate” information related to sex, gender, sexuality, and reproduction).

   [212]. Violence Against Women Act of 1994, 42 U.S.C. §§ 13925–14045d.

   [213]. Safe Connections Act of 2022, 47 U.S.C. § 345.

[214].Immigrant Legal Res. Ctr., Community Explainer: Who Is Eligible For VAWA? 1–3 (2022), https://www.ilrc.org/sites/default/files/2023-02/Who%20is%20Eligible%20for%20VAWA%3F.pdf [https://perma.cc/ZMH2-LD62].

[215]. See 42 U.S.C. §§ 13925–14045d; 47 U.S.C. § 345.

[216]. See, e.g., N.Y. Pub. Serv. Law § 48-A (providing an example of how self-attestation works in a similar statutory context).

[217]. See Calling the Police Shouldn’t Be Another Barrier, DomesticShelters.org (Nov. 7, 2016), https://www.domesticshelters.org/articles/escaping-violence/calling-the-police-shouldn-t-be-another-barrier [https://perma.cc/R3ML-NW74].

[218]. See, e.g., National Do Not Call Registry, supra note 172 (featuring a feedback provision if you still receive unwanted calls).

[219].This proposed definition reflects the findings of FTC reports concerning commercial surveillance and data brokers. See Fed. Trade Comm’n, supra note 176, at 68; Fed. Trade Comm’n, supra note 24, at 2–3, 5.

[220]. See Chris J. Hoofnagle & Jan Whittington, Free: Accounting for the Costs of the Internet’s Most Popular Price, 61 UCLA L. Rev. 606, 628 (2014) (explaining Google’s business model of exchanging a free search engine for information); see also Alexander Tsesis, The Right to Be Forgotten and Erasure: Privacy, Data Brokers, and the Indefinite Retention of Data, 48 Wake Forest L. Rev. 101, 105 (2014) (“Popular companies like Facebook, Amazon, and Google can retain users’ data indefinitely and sell it to other companies.”).

[221]. See Samuel W. Buell, Good Faith and Law Evasion, 58 UCLA L. Rev. 611, 614 (2011) (“Narrow and hard-edged rules of law create space for evasion.”).

[222]. See Kadri, supra note 2, at 149–54.

[223]. See Jules Polonetsky, Omer Tene & Kelsey Finch, Shades of Gray: Seeing the Full Spectrum of Practical Data De-Identification, 56 Santa Clara L. Rev. 593, 605 (2016); see also Lauren A. Di Lella, Comment, Accept All Cookies: Opting-in to a Comprehensive Federal Data Privacy Framework and Opting-out of a Disparate State Regulatory Regime, 68 Vill. L. Rev. 511, 513–14 (2023) (defining PII as the focus of data privacy protections).

[224]. See Paul M. Schwartz & Daniel J. Solove, The PII Problem: Privacy and a New Concept of Personally Identifiable Information, 86 N.Y.U. L. Rev. 1814, 1841–43, 1847 (2011) (explaining how identified data can often be re-identified, especially as more data points become available).

   [225]. For a background on data clustering, see generallyData Clustering: Algorithms and Applications (Charu C. Aggarwal & Chandan K. Reddy eds., 2014).

[226]. See, e.g.,Sherman, supra note 62 (explaining how aggregating data facilities stalking and abuse).

[227]. See Kuempel, supra note 62, at 221–23.

[228]. See Tal Z. Zarsky, “Mine Your Own Business!”: Making the Case for the Implications of the Data Mining of Personal Information in the Forum of Public Opinion, 5 Yale J.L. & Tech. 1, 9–11 (2003).

[229]. See Kuempel, supra note 62, at 219–21.

[230]. See Rostow, supra note 4, at 674, 670–72.

[231]. See Fed. Trade Comm’n, supra note 24, at 3, 48–49.

[232]. See VT. Stat. Ann. tit. 9, § 2446.

[233]. See Cal. Civ. Code § 1798.99.82.

   [234]. National Do Not Call Registry, supra note 172.

[235]. See generally GDPR, supra note 174, art. 17 (exploring a similar idea within the GDPR’s right to erasure).

   [236]. The FTC and the CFPB are well positioned to oversee a federal statutory scheme protecting personal and sensitive data. The FTC, as the primary agency protecting consumers from deceptive and unfair business practices, has already brought multiple enforcement actions against data brokers for abusive data practices. See Press Release, Fed. Trade Comm’n, FTC Takes Action Against Mobilewalla for Collecting and Selling Sensitive Location Data (Dec. 3, 2024), https://www.ftc.gov/news-events/news/press-releases/2024/12/ftc-takes-action-against-mobilewalla-collecting-selling-sensitive-location-data [https://perma.cc/7W2D-FY7M] (announcing an action against a data broker to prohibit the sale of sensitive location data); Press Release, Fed. Trade Comm’n, FTC Cracks Down on Mass Data Collectors: A Closer Look at Avast, X-Mode, and InMarket (Mar. 4, 2024), https://www.ftc.gov/policy/advocacy-research/tech-at-ftc/2024/03/ftc-cracks-down-mass-data-collectors-closer-look-avast-x-mode-inmarket [https://perma.cc/X5SZ-HJP2] (announcing actions against data brokers for the sale of sensitive location data). Additionally, the CFPB, an agency with a specialized focus on protecting consumers’ financial information, recently proposed a rule to protect the public’s personal and financial information. See Press Release, Consumer Fin. Prot. Bureau, CFPB Proposes Rule to Stop Data Brokers from Selling Sensitive Personal Data to Scammers, Stalkers, and Spies (Dec. 3, 2024), https://www.consumerfinance.gov/about-us/newsroom/cfpb-proposes-rule-to-stop-data-brokers-from-selling-sensitive-personal-data-to-scammers-stalkers-and-spies [https://perma.cc/778Y-XC4J].

   [237]. However, constitutional challenges to such fee-based regimes might undercut their feasibility moving forward. See generally Consumers’ Rsch. v. FCC, 109 F.4th 743 (5th Cir. 2024), cert. granted sub nom., Schs., Health & Librs. Broadband Coal. v. Consumers’ Rsch., 145 S. Ct. 587 (2024).

[238]. See Rostow, supra note 4, at 670; Data Brokers, Elec. Priv. Info. Ctr., https://epic.org/issues/consumer-privacy/data-brokers [https://perma.cc/3W9C-F2F4].

[239]. See Alyssa Wong, Regulatory Gaps and Democratic Oversight: On AI and Self-Regulation, U. of Toro.: Schwartz Reisman Inst. for Tech. & Soc’y (Sep. 21, 2023), https://srinstitute.utoronto.ca/news/tech-self-regulation-democratic-oversight [https://perma.cc/9FW9-HXW2]; Corporate Self-Regulation Is a Global Crisis, Hum. Rts. Watch (Nov. 14, 2017), https://www.hrw.org/news/2017/11/14/corporate-self-regulation-global-crisis [https://perma.cc/GN9H-UAU9]; Andreja Marusic & Madelynne Grace Wagner, How Companies Like Yum! Brands Can Improve Compliance Through Self-Regulation, World Bank Blogs (Feb. 20, 2018), https://blogs.worldbank.org/en/psd/how-companies-yum-brands-can-improve-compliance-through-self-regulation [https://perma.cc/AE25-5HYU].

[240]. See Marusic & Wagner, supra note 239.

   [241]. Douglas C. Michael, Admin. Conf. of the U.S., Federal Agency Use of Audited Self-Regulation as a Regulatory Technique 21 n.81 (1993) (citing Eugene Bardach & Robert A. Kagan, Going by the Book: The Problem of Regulatory Unreasonableness 234–38 (1982)), https://www.acus.gov/sites/default/files/documents/1994-01%20The%20Use%20of​%20Audited%20Self-Regulation%20as%20a%20Regulatory%20Technique.pdf [https://perma.cc/RX32-UZM2].

   [242]. Carleen M. Zubrzycki, The Abortion Interoperability Trap, 132 Yale L.J.F. 197, 212 (2022).

   [243]. Health Information Technology for Economic and Clinical Health Act, Pub. L. No. 111-15, 123 Stat. 115, 226 (2009).

   [244]. See Julia Adler-Milstein & Eric Pfeifer, Information Blocking: Is It Occurring and What Policy Strategies Can Address It?, 99 Milbank Q. 303, 123 (2021) (“Among the 8 specific forms of information blocking in which EHR vendors may engage, 49% of respondents reported that vendors routinely or often deploy products with limited interoperability.”); U.S. Gov’t Accountability Off., GAO-15-817, Electronic Health Records: Nonfederal Efforts to Help Achieve Health Information Interoperability 1–2, 5 (2015) (explaining that data standards dictate technical specifications for system design and data transfers that are necessary for interoperability and identifying the lack of health data standardization as a key barrier to achieving EHR interoperability).

   [245]. William J. Gordon & Kenneth D. Mandl, The 21st Century Cures Act: A Competitive Apps Market and the Risk of Innovation Blocking, 22 J. Med. Internet Rsch. 1, 1–2 (2020) (describing anticompetitive vendor behavior that might be taken to circumvent interoperability mandate); Bryan Cleveland, Note, Using the Law to Correct the Market: The Electronic Health Record (EHR) Incentives Program, 29 Harv. J.L. & Tech. 291, 311 (2015) (describing vendor lock-in by design).

   [246]. Cal. Civ. Code §§ 1798.100–.199.100.

[247].Philip N. Yannella & Timothy W. Dickens, New State Privacy Laws Creating Complicated Patchwork of Privacy Obligations, BlankRome (June 7, 2024), https://www.blankrome.com/publications/new-state-privacy-laws-creating-complicated-patchwork-privacy-obligations [https://perma.cc/M8BZ-NQCE]; Andrew Blustein, With CCPA Looming, Publishers Are Confused and Consumers Are Unlikely to Share Their Data, Drum (Oct. 3, 2019), https://www.thedrum.com/news/2019/10/03/with-ccpa-looming-publishers-are-confused-and-consumers-are-unlikely-share-their [https://perma.cc/ZW84-L58Z].

[248]. See Lauren Feiner, California’s New Privacy Law Could Cost Companies a Total of $55 Billion to Get in Compliance, CNBC (Oct. 5, 2019), https://www.cnbc.com/2019/10/05/california-consumer-privacy-act-ccpa-could-cost-companies-55-billion.html [https://perma.cc/5PUW-9LBV].

   [249]. To enhance security further, all stored identifiers can be cryptographically hashed with a process known as “salting,” where each value is combined with a unique random string before hashing. This ensures that even if the database is compromised, the hashed data cannot be reverse engineered into its original form without the “salt,” or the unique random string associated with the record. Andrew Hughes, Encryption vs. Hashing vs. Salting - What’s the Difference?, PingIdentity (Dec. 19, 2024), https://www.pingidentity.com/en/resources/blog/post/encryption-vs-hashing-vs-salting.html [https://perma.cc/DJN4-LWE7].

   [250]. Advanced cryptographic methods such as Private Set Intersection, homomorphic encryption, and Bloom filters ensure privacy-preserving queries. In these processes, the database and brokers compare hashed identifiers without revealing any additional data that the broker did not already possess. See generally Mike Rosulek, Or. State Univ., A Brief Overview of Private Set Intersection (Apr. 19, 2021), https://csrc.nist.gov/presentations/2021/stppa2-psi (on file with the California Law Review); What Is Homomorphic Encryption?, IBM, https://www.ibm.com/think/topics/homomorphic-encryption [https://perma.cc/Z48A-KNLT]; Tristan Garwood, Saving Money and Protecting Privacy with Bloom Filters, Localytics Eng (Aug. 27, 2018), https://eng.localytics.com/saving-money-protecting-privacy-with-bloom-filters [https://perma.cc/R2D6-3FV7].

   [251]. Ulf Mattsson, Privacy-Preserving Analytics and Secure Multiparty Computation, ISACA (Mar. 17, 2021), https://www.isaca.org/resources/isaca-journal/issues/2021/volume-2/privacy-preserving-analytics-and-secure-multiparty-computation [https://perma.cc/9ZLW-E85N].

[252]. What Is Homomorphic Encryption?, Supermicro, https://www.supermicro.com/en/glossary/homomorphic-encryption [https://perma.cc/ZJD5-UURH].

[253]. See generally Madhurima Nath, Fuzzy Matching Algorithms, Medium (Jan. 8, 2024), https://medium.com/@m.nath/fuzzy-matching-algorithms-81914b1bc498 [https://perma.cc/GVP7-G8RC].

[254]. See id.

[255]. See Saurabh Gupta, Piyushank Gupta, Anup Kumar & Mohd. Wasim, Privacy Preserving Optimized Fuzzy Like Search over Encrypted Data Using Phonology, 184 Int’l J. Comput. Applications 45, 46 (2022) (explaining that privacy can be preserved in large databases of sensitive information by searching for particular records with algorithms that “look for holistic patterns in data segregated from various sources” to identify a close, approximate (fuzzy) match rather than an exact match); id. at 48 (“Fuzzy matching using edit distance approach compares two strings to quantify how similar or dissimilar two strings are. Levenshtein distance technique was selected to achieve the fuzzy search which is most widely used edit distance algorithm.”).

   [256]. For an explainer on webhooks, see What is a webhook?, Red Hat, https://www.redhat.com/en/topics/automation/what-is-a-webhook [https://perma.cc/Z44F-AX3F] (last updated Feb. 1, 2024) (explaining that webhooks are a method by which developers can connect two systems and trigger an activity, including download of new information, when a predefined event happens such as new entries to a database).

   [257]. Jake Frankenfield, Personally Identifiable Information (PII): Definition, Types, and Examples, Investopedia (Aug. 1, 2025), https://www.investopedia.com/terms/p/personally-identifiable-information-pii.asp [https://perma.cc/76RE-DUZT].

[258]. See Kadri, supra note 2, at 138–39.

[259]. See generally Solon Barocas & Karen Levy, Privacy Dependencies, 95 Wash. L. Rev. 555 (2020) (exploring how a person’s privacy often depends on the decisions and disclosures of other people).

[260]. See Myra Luna Lucero & Kailee Kodama Muscente, Understanding Identifiable Data, Tchrs. Coll.: Colum. Univ. (June 29, 2020), https://www.tc.columbia.edu/institutional-review-board/irb-blog/2020/understanding-identifiable-data- [https://perma.cc/N6BG-Q9TQ].

   [261]. Protecting Americans From Harmful Data Broker Practices,89 Fed. Reg.101402 (proposed Dec. 13, 2024) (to be codified at 12 C.F.R. pt. 1022) (describing the advanced technological capabilities brokers possess); Elijah Greisz, Transparency Without Teeth: An Empirical Understanding of Data Broker Regulation, 92 U. Chi. L. Rev. 1077, 1085–87 (2025) (describing the sophisticated and technologically advanced data broker economy); Pauline T. Kim & Erika Hanson,People Analytics and the Regulation of Information Under the Fair Credit Reporting Act,61 St. Louis U. L.J. 17, 27 (2016) (describing the harvesting of large datasets for the creation of personal profiles);Danielle Keats Citron & Frank Pasquale, The Scored Society: Due Process for Automated Predictions,89 Wash. L. Rev. 1, 22 (2014) (describing the broker practice of making inferences about individuals based on data).

[262]. See Off. of Oversight & Investigations Majority Staff, Comm. on Com., Sci. & Transp., A Review of the Data Broker Industry: Collection, Use, and Sale of Consumer Data for Marketing Purposes 23 (2013).

[263]. See, e.g.,Simson L. Garfinkel, U.S. Dep’t of Com., Nat’l Inst. Standards & Tech., De-Identification of Personal Information iii, 1 (2015) (defining de-identification as a “collection of approaches” to remove “identifying information from a dataset so that individual data cannot be linked with specific individuals”).

[264]. See Kadri, supra note 2, at 138, 140–41 (showing how data brokers can fuel abuse with personal information, such as home addresses and even intimate images).

[265]. See, e.g., Ira S. Rubinstein & Woodrow Hartzog, Anonymization and Risk, 91 Wash. L. Rev. 703, 710–11 (2016) (demonstrating that entities with auxiliary information can link datasets to individuals more easily).

[266]. See, e.g., Information for Data Brokers, Cal. Priv. Prot. Agency, https://cppa.ca.gov/data_brokers [https://perma.cc/75K5-P5SV].

   [267]. Apple and Google Deliver Support for Unwanted Tracking Alerts in iOS and Android, Apple: Newsroom (May 13, 2024), https://www.apple.com/newsroom/2024/05/apple-and-google-deliver-support-for-unwanted-tracking-alerts-in-ios-and-android/ [https://perma.cc/N8G8-TU9E].

   [268]. Recht, supra note 31, at 4–7.

[269]. See generally Frederick Schauer, Categories and the First Amendment: A Play in Three Acts, 34 Vand. L. Rev. 265, 267 (1981) (describing the category of coverage); Shanor, supra note 30, at 324–30.

   [270]. Robert Post, Encryption Source Code and the First Amendment, 15 Berkeley Tech. L.J. 713, 714 (2000).

[271]. Id.

[272]. Id.

   [273]. Brown v. Ent. Merchs. Ass’n, 564 U.S. 786, 790 (2011).

[274]. See Neil M. Richards, Reconciling Data Privacy and the First Amendment, 52 UCLA L. Rev. 1149, 1150–51 (2005).

   [275]. Post, supra note 270, at 716.

[276]. See generally AlexanderTsesis, FreeSpeechConstitutionalism, 2015 U. Ill. L. Rev. 1015 (2015) (exploring the three dominant normative rationales for free speech in the United States).

   [277]. Born from Justice Oliver Wendell Holmes’s canonical dissent in Abrams v. United States, the “marketplace of ideas” theory of free speech champions the “free trade in ideas” as the premier driver of truth in a society predicated upon democratic self-government. 250 U.S. 616, 630 (1919) (Holmes, J., dissenting). While some scholars support Justice Holmes’s view, others raise concerns about the theory’s blind spots. See Eugene Volokh, In Defense of the Marketplace of Ideas / Search for Truth as a Theory of Free Speech Protection, 97 Va. L. Rev. 595, 596–97 (2011); Robert Post, Participatory Democracy and Free Speech, 97 Va. L. Rev. 477, 478–80 (2011).

[278]. See C. Edwin Baker, Human Liberty and Freedom of Speech 47–69 (1989) (arguing that speech is protected because it “promotes both the speaker’s self-fulfillment and the speaker’s ability to participate in change”); Thomas I. Emerson, Toward a General Theory of the First Amendment, 72 Yale L.J. 877, 879 (1963) (explaining freedom of expression’s role in “the achievement of self-realization”).

   [279]. This Article focuses on a participatory democracy theory of free speech rather than the broader democratic self-governance theory. Robert Post complicates the democratic self-governance theory by introducing his own nuanced offshoot of the theory rooted in participatory democracy. See Post, supra note 277, at 478. While democratic self-governance and participatory democracy may seem interchangeable, Post distinguishes between a Meiklejohnian view of democratic self-governance and his own theory of participatory democracy. See Robert Post, Reconciling Theory and Doctrine in First Amendment Jurisprudence, 88 Calif. L. Rev. 2353, 2368–69 (2000) [hereinafter Post, Reconciling Theory and Doctrine].

   [280]. Why scrutinize First Amendment values rather than simply examine and apply First Amendment doctrine? First Amendment doctrine is notoriously incoherent, and many view this incoherence as a product of doctrinal divergence from animating First Amendment values. See Robert Post, Recuperating First Amendment Doctrine, 47 Stan. L. Rev. 1249, 1249–50 (1995); Post, Reconciling Theory and Doctrine, supra note 279, at 2365; Shanor, supra note 30, at 322–23. A more elemental inquiry is warranted before presuming First Amendment doctrine extends coverage to brokers’ platforms as new media of communication. See Reno v. ACLU, 521 U.S. 844, 885 (1997); Packingham v. North Carolina, 582 U.S. 98, 107–08 (2017).

   [281]. Thomas E. Kadri, Drawing Trump Naked: Curbing the Right of Publicity to Protect Public Discourse, 78 Md. L. Rev. 899, 917 (2019).

[282]. See id. at 912–17; Leslie Kendrick, Are Speech Rights for Speakers?, 103 Va. L. Rev. 1767, 1789 (2017); Meir Dan-Cohen, Freedoms of Collective Speech: A Theory of Protected Communications by Organizations, Communities, and the State, 79 Calif. L. Rev. 1229, 1233 (1991); Morgan N. Weiland, Expanding the Periphery and Threatening the Core: The Ascendant Libertarian Speech Tradition, 69 Stan. L. Rev. 1389, 1451 (2017).

[283]. See Matthew Crain, The Limits of Transparency: Data Brokers and Commodification, 20 New Media & Soc’y 88, 90 (2018).

   [284]. Jane Bambauer, Is Data Speech?, 66 Stan. L. Rev. 57, 61 (2014).

[285]. Id. at 60.

   [286]. Jane R. Bambauer, The Empirical First Amendment, 78 Ohio St. L.J. 947, 955 (2017).

[287]. See Ashutosh Bhagwat, Sorrell v. IMS Health: Details, Detailing, and the Death of Privacy, 36 Vt. L. Rev. 855, 862–63 (2012); see also, e.g., IMS Health Inc. v. Sorrell, 630 F.3d 263, 271–72 (2d Cir. 2010).

   [288]. Bambauer’s coverage analysis privileges an expansive view of the “marketplace of ideas” theory. Meiklejohn’s approach to democratic self-governance is “quite analogous to the theory of the marketplace of ideas.” Post, Reconciling Theory and Doctrine, supra note 279, at 2369. But Meiklejohn’s approach foregrounds the aphorism that “[w]hat is essential is not that everyone shall speak, but that everything worth saying shall be said.” Alexander Meiklejohn, Political Freedom: The Constitutional Powers of the People 26 (1960). A Meiklejohnian approach aims to distinguish “between cognitive and noncognitive aspects of speech” and to extend “less constitutional protection” to the latter. Cass R. Sunstein, Pornography and the First Amendment, 1986 Duke L.J. 589, 603.

   [289]. For background on the concept of “public discourse” as an animating principle of the First Amendment, see Robert C. Post, Citizens Divided: Campaign Finance Reform and the Constitution 49 (2014) (“I shall use the term publicdiscourse to describe the communicative processes by which persons participate in the formation of public opinion.”); Robert C. Post, Constitutional Domains: Democracy, Community, Management 7 (1995) (defining public discourse as “an open structure of communication” in which there can be “reconciliation of individual and collective autonomy”); Robert C. Post, The Constitutional Status of Commercial Speech, 48 UCLA L. Rev. 1, 7 (2000) [hereinafter Post, Constitutional Status of Commercial Speech] (“Public discourse is comprised of those processes of communication that must remain open to the participation of citizens if democratic legitimacy is to be maintained.”); Robert Post, Meiklejohn’s Mistake: Individual Autonomy and the Reform of Public Discourse, 64 U. Colo. L. Rev. 1109, 1115–16 (1993) (using the term “public discourse” to refer to the “communicative processes sufficient to instill in citizens a sense of participation, legitimacy, and identification”).

[290]. See Bambauer, supra note 286, at 948–50.

[291]. See Post, supra note 280, at 1254–55.

[292]. Id. at 1254.

[293]. See id. (invoking navigation charts as an example of media that communicate particularized messages that do not get First Amendment protection). For an analysis and rejection of First Amendment protection for software navigation and digital map programs, see Tim Wu, Machine Speech, 161 U. Pa. L. Rev. 1495, 1525 (2013) (“I believe that these technologies are still unprotected tools, because their communications perform a function unrelated to the communication of ideas, namely, telling someone how to get from A to B.”).

[294]. See Post, supra note 280, at 1254.

[295]. See Raymond Shih Ray Ku, Free Speech and Abortion: The First Amendment Case Against Compelled Motherhood, 43 Cardozo L. Rev. 2105, 2126–27 (2022) (articulating that, generally, selling information is not considered speech and that the Sorrell Court relied upon a “simplistic and reductionist interpretation of the First Amendment” which “ignores the reality that the Amendment does not protect all speech or apply simply because an activity may be labeled as speech”).

[296]. See Spence v. Washington, 418 U.S. 405, 409–10 (1974).

[297]. Cf. Post, supra note 280, at 1253–55 (1995) (explaining that First Amendment values presuppose a dialogic and independent relationship between speaker and audience and navigation charts, for example, lack First Amendment protection because they speak monologically to their audience); see also Ashutosh Bhagwat, Details: Specific Facts and the First Amendment, 86 S. Cal. L. Rev. 1, 40 (2012) (“[W]hile personal details sometimes play a key role in forms of self-governance, the relationship is often far more distant.”).

[298]. See Post, Reconciling Theory and Doctrine, supra note 279, at 2371–72.

   [299]. Boos v. Barry, 485 U.S. 312, 322 (1988) (plurality opinion).

[300]. See Robert C. Post, Data Privacy and Dignitary Privacy: GoogleSpain, The Right to Be Forgotten, and the Construction of the Public Sphere, 67 Duke L.J. 981, 1009 (2018) [hereinafter Post, Data Privacy and Dignitary Privacy] (citing Robert C. Post, The Constitutional Concept of Public Discourse: Outrageous Opinion, Democratic Deliberation, and Hustler Magazine v. Falwell, 103 Harv. L. Rev. 601, 640–44, 680–84 (1990) [hereinafter Post, The Constitutional Concept of Public Discourse]); Post, The Constitutional Concept of Public Discourse, supra,at 624–26, 640–44, 680–84 (elucidating the concept of public discourse and civility norms present in First Amendment doctrine).

[301].Post, Data Privacy and Dignitary Privacy, supra note 300, at 1009.

   [302]. Kadri, supra note 281, at 948–49.

[303]. Id.

[304]. See Post, Data Privacy and Dignitary Privacy, supra note 300, at 1008–09.

[305]. See Frederick Schauer, Boundaries of the First Amendment, 117 Harv. L. Rev. 1765, 1790 (2004).

[306]. Id. at 1789.

   [307]. Schauer, supra note 30, at 1617.

[308]. See Shanor, supra note 30, at 322.

[309]. See Richards, supra note 274, at 1165–66 (criticizing the deregulatory effect of First Amendment arguments advanced by Eugene Volokh).

[310]. Id. at1169.

[311]. Id. at1168.

[312]. Id. at 1171.

[313]. See Shanor, supra note 30, at 322; Post & Shanor, supra note 30, at 166–67.

   [314]. Richards, supra note 274, at 1179.

[315]. See Schauer, supra note 305, at 1777–78.

[316]. See Shanor, supra note 30, at 322.

[317]. See Cent. Hudson Gas & Elec. Corp. v. Pub. Serv. Comm’n, 447 U.S. 557, 562–64 (1980).

[318]. See Valentine v. Chrestensen, 316 U.S. 52, 54 (1942) (“[T]he Constitution imposes no such restraint on government as respects purely commercial advertising.”).

[319]. See 425 U.S. 748, 770–71 (1976).

[320]. See 447 U.S. at 566.

[321]. Id.

[322]. Id.

[323]. See 564 U.S. 552, 570 (2011); Bhagwat, supra note 287, at 859–60.

[324]. See Robert L. Kerr, Desperately Seeking Coherence: The Lower Courts Struggle to Determine the Meaning of Sorrell for the Commercial Speech Doctrine, 7 U. Balt. J. Media L. & Ethics 1, 4 (2019) (“[L]ower courts today are drawing upon a dizzying array of approaches and emphases in seeking to articulate and apply Sorrell.”); see also N.J. Dep’t of Lab. & Workforce Dev. v. Crest Ultrasonics, 82 A.3d 258, 268 (N.J. Super. Ct. App. Div. 2014) (“[T]he Court has not clearly elucidated what that ‘heightened scrutiny’ might entail. In the wake of the Supreme Court’s post-Sorrell silence and inaction, many federal and state courts are continuing to apply the standard set forth in Central Hudson.”).

[325]. Cf. Eugene Volokh, No Take-Backs, No Do-Overs, No Data Replevin, Reason: Volokh Conspiracy (June 13, 2019), https://reason.com/2019/06/13/no-take-backs-no-do-overs-no-data-replevin [https://perma.cc/ZW63-NLTW] (remarking on a case where the court held that the plaintiff couldn’t rely on the remedy of replevin to “take back” what he already said in a digital recording).

   [326]. Cox Broad. Corp. v. Cohn, 420 U.S. 469, 471 (1975).

   [327]. Okla. Publ’g Co. v. Dist. Ct., 430 U.S. 308, 310 (1977) (per curiam).

   [328]. Smith v. Daily Mail Publ’g Co., 443 U.S. 97, 103 (1979).

   [329]. Fla. Star v. B.J.F., 491 U.S. 524, 538 (1989).

   [330]. See generally 420 U.S. 469.

   [331]. See generally 491 U.S. 524.

[332]. Cox, 420 U.S. at 471, 474 (assessing the constitutionality of imposing civil liability under a state law recognizing tortious invasion of privacy); Fla. Star, 491 U.S. at 526 (assessing the constitutionality of imposing civil liability under a state law making it illegal to “print, publish, or broadcast” the names of victims of sexual offenses (quoting Fla. Stat. § 794.03 (1987))).

[333]. Cox, 420 U.S. at 471, 473–74.

[334]. Fla. Star, 491 U.S. at 526–27.

[335]. Cox, 420 U.S. at 495–96 (finding that the information entered the “the public domain” because the records containing the information were “open to public inspection” and had been “released to the public”); Fla. Star, 491 U.S. at 524, 527 (determining that because the department did not “restrict access either to the pressroom or to the reports made available therein” the information entered “the public domain”).

   [336]. 443 U.S. 97 (1979).

[337]. Seeid. at 99.

[338]. Seeid. at 99–100.

[339]. Seeid. at 104.

[340]. Id. at 104.

[341]. See id.

   [342]. Brooks v. Thomson Reuters Corp., No. 21-cv-01418-EMC, 2021 WL 3621837, at *9 (N.D. Cal. Aug. 16, 2021).

[343]. See Post, Constitutional Status of Commercial Speech, supra note 289, at 8 (suggesting that the distinction between public discourse and commercial speech rests upon a commonsense evaluation as to whether “the utterance of a particular speaker should be understood as an effort to engage public opinion or instead simply to sell products”).

   [344]. 2021 WL 3621837, at *9.

[345]. Id.

[346]. See id.

   [347]. See generally 472 U.S. 749 (1985).

[348]. See id. at 753.

[349]. Id. at 762.

[350]. See id.

[351]. See Justin Sherman, Credit Reporting Agencies Don’t Just Report Credit Scores, Duke Sanford Tech Pol’y Program (Nov. 9, 2022), https://techpolicy.sanford.duke.edu/blogroll/credit-reporting-agencies-don’t-just-report-credit-scores [https://perma.cc/R675-DV89].

[352]. SeeData Brokers, supra note 238 (“For these companies, consumers are the product, not the customer.”).

[353]. See, e.g., Austin v. Mich. Chamber of Com., 494 U.S. 652, 655 (1990).

[354]. See J. Morris Clark, Guidelines for the Free Exercise Clause, 83 Harv. L. Rev. 327, 330–31 (1969) (“The purpose of almost any law can be traced back to one or another of the fundamental concerns of government: public health and safety, public peace and order, defense, revenue.”).

[355]. Seesupra Part I.B.1.

[356]. See Scott Skinner-Thompson, Privacy at the Margins 8 (2021); see also Faye Vasilopoulos, Hanging by a Thread: Meta’s New Platform, Threads, Sheds Light on the Slow Unraveling of Individual Privacy, 58 U. Ill. Chi. L. Rev. 473, 482 (2024).

[357]. See Danielle Keats Citron & Jonathon W. Penney, When Law Frees Us to Speak, 87 Fordham L. Rev. 2317, 2318–20 (2019); see also Jonathon W. Penney, Understanding Chilling Effects, 106 Minn. L. Rev. 1451, 1478–79 (2022); Mary Anne Franks, Free Speech Black Hole: Can the Internet Escape the Gravitational Pull of the First Amendment?, Knight First Amend. Inst. at Colum. Univ. (Aug. 21, 2019), https://knightcolumbia.org/content/the-free-speech-black-hole-can-the-internet-escape-the-gravitational-pull-of-the-first-amendment [https://perma.cc/T4KQ-RUW2].

[358]. Cf. Danielle Keats Citron, Intimate Privacy’s Protection Enables Free Speech, 2 J. Free Speech L. 3, 3 (2022) (“[I]ntimate privacy is an essential precondition for self-expression.”).

   [359]. In the limited public employer-employee context, the Court utilizes the Pickering test. See Rankin v. McPherson, 483 U.S. 378, 384–85 (1987) (balancing the speech interests of a public employee in commenting on matters of public concern with the State’s interest, as an employer, in promoting the efficiency of public services (citing Pickering v. Bd. of Educ., 391 U.S. 563, 568 (1968))); see also Anna Tichy, Gillis v. Miller, 64 N.Y.L. Sch. L. Rev. 115, 129 (2020); Abby Ward, In Defense of Pickering: When A Public Employee’s Social Media Speech, Particularly Political Speech, Conflicts with Their Employer’s Public Service, 108 Minn. L. Rev. 1643, 1700 (2024).

[360]. See Skinner-Thompson, supra note 53, at 456 (“Although visibility comes with risks for members of marginalized groups, controlled visibility through privacy protections has the potential to serve important antisubordination goals and lead to broader societal participation of entire communities in the public square. Given that public space may deny the existence of nonnormative identities, that participation may by itself be radical and politically transformative.”); see also Mary Anne Franks, Democratic Surveillance, 30 Harv. J.L. & Tech. 425, 430 (2017) (“A democratic conception of privacy, by emphasizing the experiences of those most vulnerable to its violation, offers the best chance of securing privacy for all.”).

[361]. See Andrew, supra note 83; see also Ira S. Rubinstein, Voter Privacy in the Age of Big Data, Wis. L. Rev. 861, 896–97 (2014) (claiming that a breach of political data results in harms such as a declining faith in publicly supervised political processes).

   [362]. When arguing that New Jersey’s Daniel’s Law violated the First Amendment, the class of data brokers did not even bother to challenge whether the government had a compelling interest. See Plaintiffs’ Memorandum of Law in Opposition to Defendants’ Consolidated Motion to Dismiss Plaintiffs’ Complaint at 41–42, Atlas Data Priv. Corp. v. We Inform, LLC, 758 F. Supp. 3d 322 (D.N.J. 2024) (No. 24-4037), 2024 WL 4905924 [hereinafter Atlas Data Privacy Brief].

[363]. Cf.id. at 28, 31–36 (rebutting data brokers’ arguments that New Jersey’s Daniel’s Law is both overinclusive and underinclusive).

[364]. Seeid. at 32 (“[W]hether the statute contains a verification requirement does not alter the scope of its coverage or the amount of speech it restricts.”).

[365]. Seesupra Part III.B.1.

[366]. See Atlas Data Privacy Brief, supra note 362, at 33–36.

[367]. See Burson v. Freeman, 504 U.S. 191, 209 (1992) (plurality opinion); see also Bd. of Trs. of the State Univ. of N.Y. v. Fox, 492 U.S. 469, 480 (1989) (“What our decisions require is . . . a fit that is not necessarily perfect, but reasonable . . . .”); Matthew Passalacqua, Something’s Brewing Within the Commercial Speech Doctrine, 46 Valparaiso U. L. Rev. 607, 642 (2012).

   [368]. Practical obscurity describes the functionally obscure state of scattered, uncollated, and noncomputerized information in contrast to streamlined, digital access to that very same information in an aggregated form. See Nancy S. Marder, From “Practical Obscurity” to Web Disclosure: A New Understanding of Public Information, 59 Syracuse L. Rev. 441, 441–43 (2009).

[369]. See 489 U.S. 749, 762–64 (1989).

[370]. Id. at 764.

[371]. Id.

Table of Contents Show

Introduction

In January 2025, as President Donald J. Trump assumed the presidency for a second term, one of his earliest and most consequential acts was to institute a series of economic emergency actions, relying on a familiar instrument: the tariff.[1] Since taking office, the President has issued at least twenty-three executive orders, four memoranda, and three proclamations adjusting tariff rates.[2] Although the Trump administration has defended these moves as legal and necessary for national security,[3] this presidential deployment of tariffs presents a striking departure from the constitutional architecture governing foreign commerce. That U.S. tariff rate crescendos and diminuendos are now wielded principally by the White House signals not merely a shift in practice, but a profound evolution in the understanding of federal trade authority—one that has largely escaped sustained scholarly and judicial scrutiny.[4]

This Article addresses a question that has been obscured in recent debates over trade, diplomacy, and presidential power: What are the constitutional bounds of the executive’s authority over foreign commerce? Article I, Section 8 of the Constitution unambiguously assigns to Congress the power “[t]o lay and collect . . . Duties” and “[t]o regulate Commerce with foreign Nations.”[5] These enumerated powers—collectively referred to herein as the “foreign commerce power”—represent a core component of the legislative branch’s authority. Yet in contemporary practice, the executive has come to exercise an expansive range of commercial powers: imposing tariffs, negotiating trade agreements, prohibiting certain business transactions, administering sanctions, and modifying export controls.[6] The executive branch takes many of these actions with little to no statutory authorization and only the most perfunctory gestures in the direction of congressional involvement.[7]

Through a novel review of eighteenth-century British law, constitutional drafting history, and records from the early decades of the Republic, we show how the Framers’ intent was clear: The President was not intended to have any constitutional authority to regulate foreign commerce. No drafter saw the President as the “trader in chief.”[8] No evidence from the early days of practice among the political branches suggests that the early policymakers did either.[9]

Given this clear separation, the President’s commercial acts must necessarily be based on congressional delegations. But existing statutes often do not authorize, or at best offer only tepid support for, the President’s foreign commercial policies. The President and his defenders often point instead to a presidential foreign affairs power, which they trace to Article II, as a foundation on which he can rely in whole or in part in undertaking policies that are reshaping the U.S. economy and its relationship to the rest of the world. Indeed, the prevailing justification offered by the executive branch relies on a hybridized theory, which grounds the President’s actions simultaneously in congressional delegations—delegations that often do not explicitly mention and have never been used to impose the President’s chosen instrument, such as tariffs or international agreements—and in an implied Article II general foreign affairs power. In its opening brief before the U.S. Court of Appeals for the Federal Circuit in V.O.S. Selections, Inc. v. Trump—the primary case challenging President Trump’s imposition of tariffs on imports from nearly every country under the International Emergency Economic Powers Act (IEEPA)—the government wrote, “Congress has long delegated broad tariff authority to the President to supplement the President’s Article II powers over foreign affairs. Presidents have exercised that authority across many administrations to impose tariffs that . . . protect national security, foster economic prosperity, and facilitate negotiations with foreign counterparts.”[10] This quote is among the more aggressive framings of an argument that the government regularly makes throughout foreign commerce cases: Foreign commercial delegations should be read more broadly than ordinary economic delegations in light of the President’s overlapping constitutional power over foreign affairs.[11]

Put differently, arguments about the scope of the government’s foreign commerce powers are typically contests over how broadly to read statutory delegations. And it is true that Congress has codified far-reaching delegations to the executive to manage and to administer our trade law regime.[12] Myriad statutes now on the books offer the President and executive branch agencies considerable authority to adjust U.S. trade policy.[13] But contrary to the executive branch’s position, those delegations do not combine with any perceived executive foreign affairs power to make the President’s foreign commerce powers quasi-constitutional. The government’s framing attempts to change how courts and other actors interpret foreign commerce delegations by filtering statutory interpretation through a constitutional power that does not exist. This reasoning reflects a broader jurisprudential and scholarly trend that treats foreign commerce as part of foreign affairs.[14] Such a synthesis, in which commercial regulation is increasingly subsumed within the general rubric of executive foreign relations power, as this Article contends, rests on a misreading of constitutional history, text, and structure.

Part of the problem, both in theory and in practice, is an epistemological distinction between two groups of scholars. One group considers the Commerce Clause for its reach—that is, what is it Congress can regulate inside or outside of the United States? Another group is concerned with the President’s authority over foreign policy. The former are often constitutional law scholars who focus on the Interstate Commerce Clause case law.[15] They concentrate on the substance of commerce and, if they consider the Foreign Commerce Clause at all, they extrapolate from the Interstate Commerce Clause to inform their views. The latter are foreign relations scholars who study and critique the President’s role in foreign policy.[16] These two lines of literature and styles of thinking are difficult to reconcile because they rarely speak the same language. Constitutional law scholars tend to dismiss the Foreign Commerce Clause, except when Congress seeks to regulate private activity beyond U.S. borders. Foreign relations scholars often assume that whatever the President’s foreign policy power, it encompasses activities that are commercial in nature; rarely do these scholars acknowledge that there is a Foreign Commerce Clause in Article I that could prove an exception to their Article II analyses.[17] These opposing approaches have left a gap in our collective understanding. We bring together those two conversations to cross what we see as a substantive and procedural divide that has obscured the critical legal fact that the executive branch has no power to regulate foreign commerce under the Constitution.

This Article makes three contributions. First, it corrects the prevailing misconception in both doctrine and scholarship that the President’s constitutional powers extend to foreign commerce. A careful examination of constitutional drafting history and early republican practice reveals a consensus among the Framers: Foreign commerce was a legislative power. Further, the historical record shows that the President did not gain any foreign commercial powers in situations where commercial regulation was incident to a war or other foreign diplomatic entanglement. By the eighteenth century, it was well established in Britain that only Parliament could approve tariffs or implement commercial provisions of treaties, including peace treaties—a history with which the Framers would have been familiar.[18] Debates throughout the 1780s over both the Articles of Confederation and the Constitutional Convention show an overwhelming concern that foreign commercial policy remain subject to majoritarian, if not super-majoritarian, voting to protect the diverse range of economic interests in the new nation.[19]

While the Framers acknowledged the national security implications of foreign commerce, they viewed the allocation of power as primarily an issue of federalism: Should it be the state legislatures or the federal legislature that set foreign commercial policy, and if the latter, should supermajority rules apply? The thought that a single official could set foreign commercial policy for the nation did not arise and would have been antithetical to any of the frameworks that they discussed during this period.[20] Further, the Napoleonic Wars forced the United States to confront a number of situations during the early years of the Republic in which foreign commercial policy was inextricably bound up in the general foreign policy question of whether the United States should remain neutral, engage in limited hostilities, or go to war.[21] Despite this link, Presidents from Washington to Madison deferred to Congress in establishing the rules governing foreign commerce with the European powers. The Supreme Court’s conclusion in its consolidated decision in Learning Resources, Inc. v. Trump and Trump v. V.O.S. Selections (hereinafter Learning Resources) that “the President enjoys no inherent authority to impose tariffs” because the Framers “did not vest any part of the taxing power in the Executive Branch” rests on this historical record.[22] At the same time, though, the Supreme Court expressly left open the question of whether the President might be able to impose tariffs pursuant to his constitutional warmaking powers.[23] The historical record from the years immediately after the Constitution’s ratification suggests that the answer to that question is no.

Second, this Article explains how the expansion of executive commercial power occurred through a combination of broad statutory delegations and an evolving theory of executive supremacy in foreign affairs. The latter misconstrues the reach of the President’s power based on executive practices that emerged in the twentieth century. This co-optation has had the effect of enhancing the executive’s perception of its own authority and allowing the President to invoke sweeping justifications for his global commercial actions. While this same claim might be made about foreign affairs more generally, the trend is particularly startling in foreign commerce. In other areas where twentieth-century practices accord the President a dominance inconsistent with Founding-era practices—such as the use of force and the negotiation and withdrawal of international agreements—the Constitution divides power between the President and Congress. But the Constitution vests the entirety of the federal government’s foreign commercial powers in Congress.

Third, this Article situates these developments within broader structural concerns, arguing that the erosion of the foreign commerce power’s legislative foundation poses significant risks not only for the constitutional separation of powers but also for the legitimacy of the United States’ participation in global economic governance—even beyond cross-border commercial matters. Questions about the content, reach, and scope of the foreign commerce power must urgently be addressed. While the Supreme Court’s decision in Learning Resources clarified that the President has no inherent power to impose tariffs, the Court’s decision addressed only tariffs imposed during peacetime, leaving aside both the question of tariffs imposed during wartime and broader foreign commerce issues that remain points of contention. To give only two examples: First, since the decision was issued on February 20, 2026, President Trump has relied on a different law, Section 122 of the Trade Act of 1974, to impose a 10 percent tariff on nearly all products from all countries. Two lawsuits have been filed challenging those tariffs.[24] Second, President Trump has threatened trade embargoes against U.S. allies that do not support his recent war efforts against Iran[25]—a war which has closed the Strait of Hormuz, shutting off much of the world’s oil supply and causing one of the largest commercial disruptions in the last 50 years.[26] Subsequent action to address the commercial consequences of the war may force the courts to address the relationship between the President’s warmaking powers and foreign commerce, which the Supreme Court reserved in Learning Resources. This Article fills the gap in the literature that has heretofore overlooked the constitutional history that informs these disputes.

The structure of this Article proceeds as follows. Part I surveys the existing doctrinal and scholarly treatment of the Foreign Commerce Clause, identifying the lacunae in the literature that have permitted executive overreach to flourish largely unchecked. Although the commerce clauses have been the subject of extensive judicial interpretation,[27] little attention has been devoted to the Foreign Commerce Clause as a site of separation-of-powers conflict. We review how the work that has been done sheds light on some aspects of this puzzle while obscuring others.

Part II develops the theoretical framework underpinning this inquiry. We argue that the Constitution contains no general “foreign affairs” executive power; rather, some commentators have maintained that such a power has emerged through executive practice and judicial acquiescence. We critique the reliance on “historical gloss”—by which we mean the notion that longstanding institutional practice can supplement or supplant constitutional text and the original understanding—as a foundation for executive authority over foreign commerce. The Roberts Court has exhibited skepticism toward arguments resting on historical events that contradict the understanding of the Constitution at the Founding and is unlikely to rely on historical gloss to determine the allocation of the foreign commerce power absent clear evidence from the Founding Era.

Part III comprises the Article’s core. We return to materials that guided the Framers’ thinking and draw upon originalist sources to establish that the Framers intended foreign commerce to be regulated by Congress, even when foreign entanglements and outright war are involved. The records of the Constitutional Convention and the ratification debates reveal that the primary concern animating the delegation of trade authority to Congress was the failure of the Articles of Confederation to manage interstate and international trade effectively. The presidency was largely absent from the debates surrounding trade powers, reflecting a consensus that the authority to impose tariffs and negotiate commercial terms would remain firmly within Congress’s grasp. Part III unpacks this multifaceted history and asks what “regulate commerce with foreign nations” meant in the late 1700s, demonstrating how that exercise was both a foreign affairs and legislative prerogative. We bookend this discussion of constitutional drafting with an examination of relevant eighteenth-century British and early American practices.

Part IV traces the evolution of foreign commerce authority in the nineteenth century, a period in which the executive remained largely deferential to congressional primacy in trade matters. Although Congress began to enact statutes delegating discrete authorities to the President—particularly in relation to tariff adjustments and the negotiation of reciprocal trade agreements—these measures preserved the structural understanding that foreign commerce was fundamentally a legislative domain, according to which Congress assigned tasks to the executive only where and when it saw fit. Nevertheless, this period laid the groundwork for modern confusion, as the executive’s increasing involvement in foreign affairs provided a pretext for the subsequent conflation of trade and diplomacy.

Part V examines the contemporary constitutional crisis precipitated by this shift. We document how both the Trump and Biden administrations have asserted expansive executive authority over trade matters, often grounding their claims in Article II’s foreign affairs powers. These assertions are especially evident in two domains: tariff policy and the negotiation of international economic agreements. We maintain that such practices, which have shown no sign of abating in the weeks since the Supreme Court’s decision in Learning Resources, exceed constitutional bounds and risk further entrenching a model of presidential economic unilateralism that is antithetical to the separation of powers and constitutional design.

This Article concludes by assessing the broader implications of these trends. As executive power over foreign commerce continues to expand, constitutional boundaries are being redrawn in ways that privilege expediency over accountability. The consequences extend beyond trade. In areas such as immigration and transnational regulation, the perceived elasticity of the foreign commerce power has enabled the executive to circumvent traditional checks, thereby distorting both domestic governance and international economic cooperation.

In sum, this Article urges a constitutional recalibration. The foreign commerce power is not a vestigial relic but a critical component of the Constitution’s allocation of authority. Its neglect has permitted the rise of a presidential foreign commerce power untethered from text, history, or principle. Restoring the primacy of Congress in this domain is essential not only for constitutional fidelity but for the preservation of a democratic trade policy in an era of global economic upheaval.

I. Mapping Foreign Commercial Domains

To set the stage for the Article’s historical argument concerning the scope of the foreign commerce power, this Part begins by offering an overview of how the courts and the academy have interpreted that power in the last two hundred years—and how those interpretations are overly limited relative to the foreign commerce power’s present importance. We focus in this Part on the Foreign Commerce Clause in Article I, Section 8, Clause 3.

The case law and scholarly work that have considered the Foreign Commerce Clause can be divided into two primary strands. One strand of scholarship concerns the central preoccupation of the Founders in drafting not just that language of the Constitution but also in convening the Constitutional Convention: the division of authority between the states and the federal government. Part I.A lays out this federalism-focused scholarship and related case law. A second strand examines the meaning of the words “regulate” and “commerce,” while assuming that, among the branches of the federal government, Congress is responsible for that undertaking.[28] As relevant to foreign commerce, the primary concern of this literature is the reach of congressional authority on economic matters abroad. Part I.B summarizes this “exterritoriality” canon and the relevant case law. These two Sections illustrate that in the federalism-oriented cases, plaintiffs have invoked the Clause as a sword to challenge state regulations alleged to intrude upon foreign commerce. By contrast, in the extraterritoriality cases, defendants have deployed the Clause as a shield against the application of expansive federal criminal statutes—most notably in prosecutions for sex offenses.

A third body of work touches on the Foreign Commerce Clause only incidentally, but in a way that is important to this study. This work is carried out by academics who are primarily international law scholars. Much of the trade law literature of the last several decades concentrates on the rights and obligations of governments under international law, particularly at the World Trade Organization. A small handful of international law scholars have considered U.S. domestic law as it relates to international commercial activity.[29] We examine this feature and its bugs in Part I.C.

While the literature on federalism, extraterritoriality, and international trade is useful in understanding certain aspects of the foreign commerce power, the collective work to date has offered little clarity on the executive-legislative relationship with respect to foreign commerce. The limited focus on these three strands has obfuscated any analysis on the critical question of today: What is the scope of the executive’s constitutional authority over foreign commerce?

A. Federalism

Although policing the boundaries of executive power may be the most important constitutional question of the twenty-first century regarding the foreign commerce power, the dominant question of the eighteenth century concerned policing the boundaries of federal power vis-à-vis the states. Federalism was the driving force behind the drafting of Article I of the Constitution, and arguably the primary theme at the Constitutional Convention as a general matter, which we discuss in Part II. Thus, it should come as no surprise that in the early decades of the nation, policing the line between state authority and federal power would preoccupy jurists and policymakers more than any other Commerce Clause issue.

As a result, the earliest case law on the Commerce Clause was almost exclusively about the other two elements of the Clause—interstate commerce and commerce with the Indian tribes—rather than the foreign element. Any treatment of the Foreign Commerce Clause was incidental to the other two, or in service of doctrinal theories that guided the courts with respect to the other two.[30]

During this early period, the Interstate Commerce Clause was at the center of frequent disputes—far more frequent than its sibling Foreign Commerce Clause.[31] Nevertheless, to establish the breadth of the Commerce Clause, Chief Justice Marshall—in the foundational case guiding interstate commerce jurisprudence, Gibbons v. Ogden (1824)—set out a unified theory of the Clause that allowed him to also speak to the reach of foreign commerce. Chief Justice Marshall affirmed that Congress’s power over foreign commerce “does not stop at the jurisdictional lines of the Several States.”[32] Highlighting the need for federal supremacy in foreign commerce, as well as interstate commerce, the Chief Justice interpreted the “commerce of the United States with foreign nations” as a federal power Congress could exercise “whenever the subject exists,” including within a state.[33]

Nineteenth century scholarship on the commerce power similarly reflected the dominance of the Interstate Commerce Clause and the tendency to lump the Foreign Commerce Clause into this analysis via incidental references to foreign commerce. Writing in 1887 in the first issue of the first volume of the Harvard Law Review, Louis Greeley commented that “[n]o class of cases [is] more perplexing than those involving the distinction between the power of Congress to regulate foreign or interstate commerce and the so-called police power of the States.”[34] According to Greeley, the courts had not wrestled with the Commerce Clause with any consistent principles.[35] Like the courts, Greeley never differentiates between the regulation of foreign commerce and the regulation of interstate commerce. Every reference he makes to the engagement of the states and Congress with relevant legislation treats both foreign and interstate commerce equally. And yet, all the cases he discusses are occupied almost exclusively with interstate commerce. Modern scholarship has likewise focused on interstate commerce cases.[36]

Perhaps unsurprisingly given the nature of global commerce at the time, it took several more decades before substantive questions about the regulation of commerce with foreign nations crept into the courts. By the early part of the twentieth century, cases dealing with the Foreign Commerce Clause and federalism became more common.[37] The courts began to treat foreign commerce as its own separate line of jurisprudence.[38]

Once courts began to engage with foreign commerce as a distinct constitutional question, they frequently grounded Congress’s plenary authority in the imperatives of foreign relations for a federal system. Judicial opinions emphasized the necessity of a coherent, unified national policy—one that could not be undermined by divergent state regulation. In such cases, the Foreign Commerce Clause functioned as a sword, invalidating state measures or foreclosing state claims in areas touching transnational business and trade. Board of Trustees v. United States (1933) illustrates this point. There, the Supreme Court considered whether Illinois’s regulatory regime could supersede federal tariff policy. The Court concluded that “[i]n international relations and with respect to foreign intercourse and trade the people of the United States act through a single government with unified and adequate national power,”[39] and further that “[i]t is an essential attribute of the power that it is exclusive and plenary. As an exclusive power, its exercise may not be limited, qualified, or impeded to any extent by state action.”[40] These holdings extended beyond mere preemption analysis; rather, they reflected a broader constitutional rationale—rooted in the need for a unified foreign policy—that informed the Court’s understanding of the Clause and the foreign commerce power generally.

By the latter half of the twentieth century, federalism-oriented Foreign Commerce Clause jurisprudence had evolved to address what some scholars term the “inward-looking” foreign commerce power: Congress’s authority to regulate domestic activity bearing on foreign commerce or carrying foreign commercial implications.[41] These decisions likewise embraced the view that the foreign commerce power must be construed expansively to equip the federal government to advance the country’s foreign policy agenda, including where it impacts internal policies.[42] Starting in the 1970s, the Supreme Court articulated more explicitly the proposition that the Framers intended the foreign commerce power to displace contrary state action, consistent with the idea of manifesting “one voice”:

The Framers of the Constitution thus sought to alleviate three main concerns by committing sole power to lay imposts and duties on imports in the Federal Government, with no concurrent state power: the Federal Government must speak with one voice when regulating commercial relations with foreign governments, and tariffs, which might affect foreign relations, could not be implemented by the States consistently with that exclusive power; import revenues were to be the major source of revenue of the Federal Government and should not be diverted to the States . . . .[43]

Most of these cases involve taxes imposed by states or related challenges to state action. Those defending the federal government’s exclusive prerogatives in this space rely on the foreign commerce power to overcome any claim by the state to the contrary. The most recent appearance of the “one-voice” theory of federal authority with respect to the foreign commerce power in the courts is from 1993.[44]

But federalism problems are not the most important foreign commerce problems today. More recently, scholars and the courts have asked just how far the federal government’s plenary and exclusive reach may travel abroad, as the next Section explores.

B. Extraterritoriality

Whereas the federalism-focused foreign commerce cases and literature thereon concentrate on the relationship between Congress and the states, another track of literature and judicial engagement concerning the foreign commerce power examines the reach of U.S. law abroad. These works question what Congress can regulate beyond the boundaries of the United States and the extraterritorial scope of U.S. law.[45] In these cases, which often deal with statutes where Congress has criminalized conduct occurring abroad, defendants seek to characterize those laws as beyond the scope of congressional authority and interfering with foreign sovereignty. Anthony Colangelo has referred to these laws as manifestations of Congress’s outward-looking regulatory powers.[46] Colangelo argues that while Congress may exercise expansive authority within the United States, its power to regulate conduct within foreign nations is constitutionally narrower.[47]

Congress has in recent years sought to regulate conduct abroad in myriad ways, many of which have come under scrutiny for their expansive extraterritorial reach. For example, the 2003 Prosecutorial Remedies and Other Tools to End the Exploitation of Children Today Act, which is designed to combat international child-sex tourism and which criminalizes travel abroad for the purpose of engaging in illicit sexual activity, is one such law.[48] The law also prohibits “travel[] in foreign commerce, with intent to engage in any illicit sexual conduct,” among other restrictions.[49] Other criminal statutes also seek to reach beyond U.S. borders in similar fashion.[50] Similar issues arise in intellectual property law, such as the Lanham Act.[51] The components of these extraterritorial statutes have been a subject of debate for many decades, but the Court has yet to set out a clear test for what triggers an analysis of any given law under the Foreign Commerce Clause, let alone how to evaluate a law under the Foreign Commerce Clause.[52]

With respect to the substance of the Foreign Commerce Clause as applied to these statutes, Justice Thomas has noted, writing in dissent from a denial of grant of certiorari in 2017, that the Supreme Court has never thoroughly explored the scope of the Clause with respect to its application to extraterritorial statutes, especially not with respect to those statutes that seek to regulate conduct occurring entirely outside the United States.[53] Even if the Court had done so, though, this analysis is irrelevant for understanding the allocation of power between the political branches.

In sum, while extraterritorial statutory analysis remains a significant field, it is, in effect, a distinct doctrinal domain that offers little guidance on the pressing question about executive authority. The prevailing approach taken by the courts of appeals has been expansive: permitting Congress to regulate economic activity abroad “if it has a substantial effect on this Nation’s foreign commerce.”[54] Yet these courts have not addressed the parallel inquiry as applied to the executive, leaving unresolved the scope—if any—of presidential authority in this sphere.

C.   International Trade Law

A final place where one would expect to see some treatment of the Foreign Commerce Clause is in international trade law, which is quite expressly commerce with foreign nations. Although, as the name suggests, the field is one primarily of international law, some international trade law scholars—such as ourselves—have taken up the U.S. statutory and constitutional elements that encroach or interact with international treaties, including the World Trade Organization (WTO) Agreements and free trade agreements (FTAs).[55] However, these studies only superficially address the question at hand. They do not explore how Congress’s foreign commerce power shapes the bounds of the President’s authority. They are occupied with the legal terms of the United States’s participation in international agreements and organizations, as well as, to some degree, with state regulatory actions that may encroach on U.S. obligations under these instruments.[56]

Scholarship at the intersection of international trade law and the U.S. foreign commerce power also comes up episodically. For example, at the time of the creation of the WTO, when the federal government signed the WTO Agreements and made changes to U.S. law to implement them, a wave of academic commentary examined the interplay between international and domestic law, including some studies related to the separation of powers and constitutional law.[57] This moment in time coincided with the entry into force of the North American Free Trade Agreement (NAFTA), which also triggered explorations into the relationship between congressional-executive agreements (that happened to be related to trade) and the Treaty Clause in Article II.[58]

During the first Trump administration, an upsurge of scholarship confronted the novel issues surfaced by the President’s use of expansive delegations in foreign commercial affairs, especially his imposition of sweeping new tariffs, in some cases in reliance on statutes never used for that purpose, which then faced scrutiny at the WTO.[59] The legal question at the international level was whether the U.S. actions violated WTO rules or whether the United States could legitimately rely on the so-called national security exception under those rules.[60] The scholarly works that engaged with the U.S. domestic law questions and the case law that ensued concentrated on the delegation issues posed by the statutes on which the President relied to impose tariffs as well as on the procedural and administrative law questions that the executive branch’s processes presented.[61]

Apart from these newsworthy moments in time, little attention has been paid to the intersections between international law and domestic law on these issues. Neither the case law nor the academic writing has confronted the words of the two clauses comprising Congress’s foreign commerce power nor its historical pedigree, particularly in the context of the constitutional dimensions of presidential trade regulation. Further, a theme that runs through these strands of scholarship is that any separation of powers question that could possibly arise in respect of the regulation of commerce with foreign nations is one that ought to be left to foreign affairs scholars.[62] Thus, even more recent manifestations of foreign commercial policymaking have not triggered the investigation that we take up in Part III.

***

The failure to determine the scope of the executive’s constitutional authority over foreign commerce carries significant modern implications. It leaves courts with a sparse body of precedent—particularly precedent grounded in originalist methodology—when confronting the separation of powers questions that have arisen with new urgency in the Trump era. The absence of originalist engagement with such issues tends to steer judicial reasoning away from historical and textual analysis and toward prudential considerations. Lastly, it entrenches the erroneous conflation of foreign affairs and foreign commerce, a categorical mistake that we examine next.

In Part II, we explain how the limited attention paid to the foreign commerce power sets in relief this Article’s unpacking of the scholarly and doctrinal relationship between foreign affairs and foreign commerce. We demonstrate how the foreign relations literature has failed to sufficiently disaggregate foreign commerce from its generalist examination of foreign affairs powers. While scholars have readily tackled separation of powers questions as they relate to U.S. foreign policy, they have tended to treat much of that policy as a uniform whole, without regard to the Article I references to Congress’s foreign commerce powers.

II. The Generalist Foreign Affairs Power

In this Part, we outline and critique current theories about foreign affairs powers as applied to foreign commerce. We make two main arguments, one substantive and the other methodological. The substantive point is that foreign relations law scholars have tended to treat foreign affairs powers as an undifferentiated whole, paying only lip service to the foreign commerce power. The result has been a trans-substantive foreign affairs doctrine that is detached from both the text of the Constitution and the historical practice in relation to foreign commerce. The methodological point is that a significant strand of the foreign relations law literature rests on an understanding of the role of history that departs from the current Supreme Court’s use of history. Since at least Youngstown Sheet & Tube Co. v. Sawyer (1952), the Supreme Court and scholars have emphasized the role that historical practice can play in determining the allocation of powers between the branches and thus in evaluating the legality of executive action.[63] However, the Roberts Court has largely rejected claims built solely on modern practices, preferring to focus on historical evidence that speaks to the Founders’ original understanding of the Constitution.[64] As a consequence, much of the historical practice regularly cited in support of a broad executive power, particularly in the commercial context, no longer fits within the Supreme Court’s guidance about relevant history.

The purpose of making this critique is to explain how separation of powers issues that arise in the context of foreign commerce have been overlooked. As we discussed in Part I, existing studies and case law on the foreign commerce power have not focused on separation of powers issues. Instead, when separation of powers issues come up, courts and foreign commerce scholars have turned to the literature on foreign affairs, which is consumed by separation of powers questions. At first blush, this reliance on the foreign affairs literature to answer separation of powers questions might make sense as a matter of relative competence.

Unfortunately, though, the general foreign affairs literature too often grounds its analysis in doctrines related to war and diplomatic powers—domains in which the Constitution expressly divides power between the President and Congress. Foreign commerce, by contrast, is constitutionally distinct: The text vests the entirety of that power in Congress. As a result, reasoning drawn from the broader foreign affairs context overlooks the unique treatment the Constitution gives foreign commerce. This oversight is further compounded when scholars anchor their arguments in twentieth-century practices that themselves have drifted from the Constitution’s text and its original understanding.

A. (Un)Enumerating Foreign Affairs

The U.S. Constitution does not contain any reference to a general “foreign affairs” power.[65] It instead commits certain enumerated powers to Congress and others to the President. For example, in addition to power “to lay and collect Taxes, Duties, Imposts and Excises” and “[t]o regulate Commerce with foreign Nations,” the Constitution also empowers Congress “[t]o define and punish Piracies and Felonies committed on the high Seas, and Offences against the Law of Nations,”[66] “[t]o declare War,”[67] “[t]o raise and support Armies,” “[t]o provide and maintain a Navy,”[68] and “[t]o make Rules for the Government and Regulation of the land and naval Forces.”[69] The President, for his part, is the “Commander in Chief of the Army and Navy;”[70] has the power, “with the Advice and Consent of the Senate, to make Treaties” and “appoint Ambassadors;”[71] and has the responsibility to “receive Ambassadors” from foreign countries.[72]

This allocation of powers does not tell us nearly as much in practice as we might like to know.[73] Some of the open questions involve the interaction between Congress’s powers and the President’s. Nowhere has this interaction been sharper than war powers. When may the President commit the armed forces to combat without congressional approval, and when does the scale of a military operation become so large that it is tantamount to a “war” requiring congressional authorization? In other cases, the Constitution appears not to specify which branch exercises certain powers that clearly a government must exercise, such as the power to recognize foreign nations.[74] In still other cases, one or both branches have devised methods of conducting foreign affairs other than those contemplated in the Constitution. For instance, since the nation’s earliest days, Presidents have concluded “executive agreements”—treaties under international law that the President does not submit to the Senate for its advice and consent prior to entry into force.[75]

Perhaps surprisingly, given the wide range of issues that fall under the umbrella of foreign affairs and the Constitution’s division of authority over these issues between the President and Congress, judges, practitioners, and scholars have reached a constitutional settlement in which the executive branch dominates virtually all foreign affairs issues regardless of subject matter.[76] This executive branch dominance has multiple contributing factors. Constitutionally, executive branch lawyers and scholars have devised new theories of executive power. Scholars have identified a “residual foreign affairs power” in Article I’s vesting of the “executive power” in the President.[77] Under this theory, foreign affairs powers that are not explicitly allocated to either branch are granted to the executive via the Vesting Clause on the grounds that the concept of executive power in the eighteenth century included substantial authority over foreign affairs.[78] Some of these approaches to executive power, such as the one espoused by Saikrishna Prakash and Michael Ramsey,[79] continue to account for Congress’s enumerated powers acting as a limit on executive dominance. However, the development of the unitary executive theory—under which congressional efforts to control the internal operations of the executive branch run afoul of the President’s constitutional prerogatives—has expanded the tools executive branch lawyers have to assert constitutional freedom to act in foreign affairs.[80] At the statutory level, faced with complicated foreign policy problems to which they wished to react quickly, Presidents of both parties have “constru[ed] laws that were enacted to constrain executive authority as authorizing executive action.”[81] The structure of the executive branch, both in terms of its low internal transaction costs to decision-making, as well as its ability to proactively publicize and push for its preferred legal interpretations, creates a situation in which executive claims of power are often the default to which the other branches react.[82]

Confronted with sustained executive efforts to expand executive authority, both Congress and the courts have, in significant respects, acquiesced.[83] In Congress’s case, while it has periodically enacted statutes that aim to reclaim its place in the foreign affairs landscape, rarely has the legislature undertaken the sustained oversight necessary to ensure executive compliance with statutory limits.[84] Partisan polarization has compounded this problem, diminishing the incentives for members to constrain Presidents of their own party.[85] The judiciary has further inhibited Congress through a selective deployment of formalist and functionalist reasoning. On the formalist side, the Supreme Court has struck down mechanisms like the legislative veto that facilitated congressional supervision of foreign affairs powers, on the ground that Congress must legislate only through bicameralism and presentment.[86] But when assessing executive branch action, the courts have instead turned to functionalist considerations, holding that the executive branch’s advantages in terms of expertise, speed of decision-making, and secrecy justify its primacy in foreign affairs.[87]

These theories of constitutional and statutory interpretation typically take “foreign affairs” as the correct level of abstraction at which to analyze the allocation of power.[88] Yet, of note for our purposes here, foreign affairs cases and scholarship pay little attention to the role of Congress’s commercial powers in analyzing the distribution of foreign affairs powers. Take United States v. Curtiss-Wright Export Corp. (1936).[89] Curtiss-Wright is most famous for its statement that “the President [i]s the sole organ of the federal government in the field of international relations.”[90] As the Supreme Court itself has noted, this language is dicta because Congress had authorized the President’s action challenged in this case: “the prohibition of the sale of arms and munitions of war in the United States” to Bolivia and Paraguay.[91] Nevertheless, the Court’s statement could be interpreted to mean that the President would have had the authority, acting alone, to impose the restriction on sale. That implied assertion is remarkable given the Constitution’s explicit allocation of authority over commerce to Congress, as well as the lack of any connection between a border war involving two South American countries and any of the President’s enumerated constitutional powers.

At bottom, the general nature of the “foreign affairs” power may rest on the general nature of diplomacy, as conceived in constitutional discourse. As Jean Galbraith has discussed, “[a] core assumption of the executive branch is that the President possesses exclusive constitutional powers with respect to diplomacy.”[92] Moreover, the executive’s understanding of what constitutes diplomacy is “panoramic,” encompassing, as the Office of Legal Counsel (OLC) has put it with respect to negotiations, “any subject that has bearing on the national interest.”[93] The President’s purported exclusive control over diplomacy—a power both absent from the text of the Constitution and historically unsupported, as Galbraith explains[94]—has swallowed efforts to engage in the kind of nuanced analysis of the allocation of powers among the branches, and the specific powers being invoked by particular executive branch actions, that the Constitution’s text and its original understanding require.[95]

B. Historical Gloss of a General Foreign Affairs Power

The creation of a general foreign affairs power owes much to a historical methodology—often referred to as “historical gloss”—that takes any consistent historical practice, no matter when it emerged, as relevant for constitutional analysis. Courts have relied on this kind of methodology in deciding separation of powers cases, and it is a favorite technique of the executive branch, especially OLC, when evaluating the legality of its own practices. In terms of judicial precedent, the most prominent articulations of historical gloss come from two concurring opinions in Youngstown. Justice Frankfurter famously wrote that:

[A] systematic, unbroken, executive practice, long pursued to the knowledge of the Congress and never before questioned, engaged in by Presidents who have also sworn to uphold the Constitution, making as it were such exercise of power part of the structure of our government, may be treated as a gloss on “executive Power” vested in the President by § 1 of Art. II.[96]

Justice Jackson’s even more famous concurrence elaborates upon that statement. He postulated a three-part framework for evaluating executive branch action. When the executive branch acts with congressional authorization, it acts with “the strongest of presumptions.”[97] When the President acts against the expressed will of Congress, his power is “at its lowest ebb.”[98] But in the middle, what Jackson referred to as the “zone of twilight,” the distribution of power is uncertain and “congressional inertia, indifference or quiescence may sometimes, at least as a practical matter, enable, if not invite, measures on independent presidential responsibility.”[99] From these statements, the Supreme Court has constructed a doctrine of acquiescence, whereby congressional inaction in the face of a consistent executive practice can form the basis for a conclusion that executive branch action is lawful.[100] In this way, Justice Jackson’s Youngstown concurrence—like Curtiss-Wright, with which it is often contrasted—can invite courts and scholars to pay only lip service to the actual constitutional allocation of authority over a particular subject, focusing instead on what the President and Congress did at a particular moment in time.

Several prominent commentators, including Curtis Bradley, Trevor Morrison, and Neil Siegel have sought to provide a more theoretical foundation for what might be called the doctrine of historical gloss.[101] The key idea is that arguments about historical practice, especially between the political branches, have been prominent in scholarship, executive branch practices, and judicial decisions regarding the allocation of powers among the branches.[102] These arguments are especially important in foreign affairs, where a ream of significant twentieth century cases—including Youngstown, Curtiss-Wright, and Dames & Moore v. Regan (1981)—rely on claims about consistent historical practices in the years leading up to those disputes.[103] Significantly, though, the historical practice does not necessarily have to be located at any particular point in the nation’s history to be legally relevant. By contrast, a different approach to history known as constitutional “liquidation” prioritizes the historical practices of the Founding Era on the theory that the Founding Generation has unique insights into original meaning.[104] Historical gloss and constitutional liquidation are thus related insofar as both would give weight to a consistent historical practice emerging in the early years of the nation. But historical gloss could permit that original understanding to be varied by subsequent consistent practice, whereas the most “narrow” conception of liquidation would privilege early constitutional interpretations as “‘permanent’ expositions of constitutional meaning.”[105] As we explain in Part IV and elsewhere, the former is exactly what has occurred with the foreign commerce power.[106]

As an interpretative methodology, historical gloss has been extraordinarily important in the creation of a general foreign affairs power. Withdrawal from treaties offers a useful example. The claim that the President enjoys the power to unilaterally withdraw from treaties is based on twentieth-century practice; commentators and the Restatement (Fourth) of the Foreign Relations Law of the United States agree that prior to the twentieth century, seeking congressional consent to withdraw from international agreements was common, if not the norm.[107] But in this area as in others, consistent executive claims to foreign affairs powers, combined with internal gridlock that prevents Congress from effectively responding to executive action, has led to a steady accretion of historical precedent in favor of an expansive executive power.[108]

Historical gloss sits uncomfortably with the Roberts Court’s approach to history. The Roberts Court has frequently employed what Leah Litman has referred to as an “antinovelty” canon.[109] As she puts it, “Every Justice on the Supreme Court has joined an opinion promoting the idea that legislative novelty is evidence of a constitutional defect.”[110] The antinovelty canon has significant implications for the kinds of historical practice that are relevant to constitutional interpretation: “When the Court has used antinovelty rhetoric to invalidate a statute, it has frequently discounted recently enacted statutes as ‘not relevant,’ instead focusing on whether the statute is similar to statutes enacted in the early 1800s.”[111] In other words, historical practice after the Founding Era has not been probative of constitutional meaning in the Roberts Court’s eyes and cannot be used to vary the constitutional practices of the Founding Generation.

For example, the Roberts Court has invoked the antinovelty canon to strike down statutes that structure agencies in ways that the Court has found lack a sufficient historical pedigree. In Free Enterprise Fund v. Public Company Accounting Oversight Board (2010), the Court found that for-cause removal protections for the Public Company Accounting Oversight Board (PCAOB)—a second layer of for-cause removal after the for-cause removal provisions that apply to commissioners of the Securities and Exchange Commission—were unconstitutional.[112] In so holding, the Justices in the majority on this point concluded that “the most telling indication of the severe constitutional problem with the PCAOB is the lack of historical precedent for this entity.”[113] Since Free Enterprise Fund, both the Supreme Court and the lower courts have extended that holding to situations in which the administrative structure Congress has created has a considerably longer historical pedigree, just not one that dates to the Founding.[114] This progression has culminated in what most have understood to be the Court’s likely overruling (or at least significant limiting) of Humphrey’s Executor in Trump v. Slaughter (2025),[115] foreshadowed in an emergency order in Trump v. Wilcox (2025).[116] The result is that over a century’s worth of congressional and judicial precedent upholding the constitutionality of for-cause removal protections for independent agencies will be replaced by the Court’s original understanding that the President must be free to dismiss principal officers unless the agency has an analog from the Founding Era, as the majority concluded that the Federal Reserve does.[117]

Nor is there any doubt that the antinovelty canon applies to foreign affairs and national security issues. In Zivotofsky v. Kerry (2015), the Supreme Court struck down a congressional statute allowing U.S. citizens born in Jerusalem to have Israel listed as their place of birth on the grounds that such a statute unconstitutionally interfered with the President’s recognition power.[118] In so doing, the Court noted that, while “[o]n occasion, the President has chosen . . . to consult and coordinate with Congress . . . . the most striking thing about the history of recognition is what is absent from it: a situation like this one, where Congress has enacted a statute contrary to the President’s formal and considered statement concerning recognition.”[119] In Biden v. Nebraska (2023), the Supreme Court applied the antinovelty canon to the interpretation of the HEROES Act, a statute that authorizes the Secretary of Education to “waive or modify any statutory or regulatory provision applicable to the student financial assistance programs under title IV of the [Education Act] as the Secretary deems necessary in connection with a war or other military operation or national emergency.”[120] Despite the fact that the statute is only triggered by “war or other military operation or national emergency,” the Court concluded that because “[t]he Secretary has never previously claimed powers of this magnitude under the HEROES Act,” the Secretary lacked the authority to grant the waiver at issue in that case.[121]

***

The importance that the Supreme Court has placed on Founding Era history and its willingness to discount more recent historical practices, either because they are recent or because they are too occasional or modest in scope, calls into question much of the conventional wisdom about the foreign commerce power. The broad exercise of executive power over foreign commerce seen today did not begin to emerge until the late nineteenth century, as Part IV explains. But as we argue in Part III, this view of the foreign commerce power lacks a meaningful historical basis in the Founding Era.

As a result, contemporary reliance on the foreign affairs literature to resolve questions concerning the constitutional allocation of the foreign commerce power is often misplaced. Much of the modern scholarship gestures toward Congress’s exclusive authority over commerce, only to subsume it within a broader and largely undifferentiated conception of a general foreign affairs power. This analytical conflation is further compounded by the elevation of historical practices—many of which crystallized only in the twentieth century—into sources of constitutional meaning. But the Roberts Court’s increasing emphasis on originalist methodology gives scholars and jurists an opportunity to reassess. With that in mind, in Part III we undertake a close examination of the original understanding of the foreign commerce power, and in particular the allocation of institutional authority where questions of foreign commerce intersected with the President’s foreign affairs powers in matters of war or national security.

III. Foreign Commerce at the Founding

At the time of the Constitutional Convention, ratification, and early Republic, both the Framers and the general public agreed that the Articles of Confederation suffered from a major flaw: They did not allow the federal government to regulate commerce on behalf of the United States.[122] That flaw brought the young nation to the brink of collapse right after its fragile birth.

This Part takes a fresh look at the history leading up to the Convention, the constitutional drafts, the final text, and the historical practice in the early years of the Republic. We analyze how the Framers thought about the content and distribution of the foreign commerce power. There is little explicit discussion in the record to help the modern policymaker understand the Framers’ views on the separation of powers on this issue, but that silence does not mean that the record lacks all guidance. In Part III.A, we examine the seventeenth- and eighteenth-century British practice, with which the Framers would have been familiar, and which is often thought to provide the basis for the President’s generalist foreign affairs power. That history explicitly dealt with the distribution of the foreign commerce power between the Crown and Parliament and shows two things: first, that Parliament controlled taxation, including tariffs, and the regulation of foreign commerce, and second, that the intersection of war and peace with foreign commerce did not alter parliamentary supremacy over commerce.

Parts III.B and III.C look at the Articles of Confederation, the Constitutional Convention, and the resulting constitutional text. These sources also make clear that the foreign commerce power was viewed as a legislative power. The thought of an executive officer wielding that power is simply absent. That relative silence is the dog that didn’t bark. Coming out of the experience with the Articles of Confederation, the Framers recognized the allocation of the foreign commerce power, like the power to regulate interstate commerce, as a federalism issue. There was therefore little reason for them to discuss the separation of powers between the executive and the legislature. That is, the question they were dealing with was whether Congress or the states should have primary authority over foreign commerce, not whether Congress or the President would.

Finally, Part III.D looks at how the Founding Generation implemented the foreign commerce power during the period from roughly 1789 to 1820. In practice, the President was an important part of the foreign commerce apparatus, both because he negotiated commercial treaties with the Senate’s consent and because Congress delegated administrative power to the President. But when the President’s own constitutional powers intersected with the foreign commerce power, the historical record shows that both branches understood Congress’s commerce power to take precedence over the President’s own powers.

A. Parliamentary Control over Taxation and Commerce

Any inquiry into the original understanding of the constitutional powers to impose tariffs and regulate foreign commerce appropriately begins with eighteenth-century British practice. The general foreign affairs power claimed by the modern executive, as noted above, is often traced to the British monarch’s royal prerogative.[123] Proponents of this view contend that the Vesting Clause of Article II incorporates the monarch’s traditional powers over foreign affairs in a residual fashion: Those powers included within the royal prerogative but not expressly allocated by the Constitution pass to the President via the Vesting Clause.[124] To be sure, this claim is far from a consensus view among scholars, with many viewing Article II’s Vesting Clause as merely the power to implement (i.e., “execute”) laws passed by Congress.[125] For present purposes, this discussion accepts the “royal residuum,” and we ask whether, even on that premise, any power over foreign commerce could have passed to the President.

The answer is no. The eighteenth-century British monarch was no tariff man.[126] Tariffs are, of course, just taxes,[127] and Parliament had asserted control over taxation, including tariffs, for centuries before U.S. independence. The roots of parliamentary control over taxation go back as far as the thirteenth century, when the English nobility acted as an assembly whose consent to taxes of a non-feudal nature was necessary and sometimes withheld.[128] The early Stuart kings, particularly Charles I, attempted to claim increased executive power over taxes and especially tariffs. Since the fifteenth century, Parliament had given each successive English monarch a life grant of tonnage and poundage, a form of customs duty.[129] But when Charles I ascended the throne in 1625, the House of Commons approved tonnage and poundage for only a single year.[130] Among other predations, Charles resorted to collecting tonnage and poundage without parliamentary authority.[131] In the years that followed, parliamentary leaders cited among their list of grievances against Charles “the taking of Tonnage and Powndage without graunte from the Parliament.”[132] On June 22, 1641, Parliament made it unambiguously unlawful to collect tonnage and poundage without parliamentary consent.[133] The Glorious Revolution and the resulting Bill of Rights in 1689 further solidified parliamentary control over taxation.[134] In short, the tariff power, like the tax power of which it was part, rested squarely with Parliament. The Stuart kings’ efforts to tax and tariff without parliamentary assent led to the execution of one king (Charles I) and the overthrow of another (James II).[135]

More generally, the royal prerogative in the eighteenth century included:

whether and how to make war or peace; regulating captures and prizes of war; sending and receiving ambassadors; entering into and ending treaties and alliances; initiating and determining the content of any other foreign policy communications; [and] appointing and commissioning civil, army, and navy officers, and directing operations of the militia, army, navy, and senior civil officials . . . .[136]

Notably absent from this list is any substantive power over not only tariffs, but foreign commerce more generally. To the extent that the monarch enjoyed any power over the regulation of foreign commerce, it flowed from either his power to make treaties or his ability to make war and peace. Yet even when the monarch’s royal prerogatives over those subjects overlapped with foreign commerce, the historical record demonstrates that the royal prerogatives to make treaties and to declare war and peace did not limit parliamentary control over commerce when the two intersected.

The best example of this primacy is the saga of the 1713 Peace of Utrecht, which ended the War of the Spanish Succession—the first major pan-European war of the eighteenth century.[137] As backdrop, the British system of treaty-making has long followed a dualist model.[138] While the power to make treaties falls within the royal prerogative, treaties cannot have domestic legal effect in Britain unless Parliament enacts any necessary laws.[139] In effect, the monarch controlled Britain’s international legal obligations, but the monarch lacked the authority to give those international obligations domestic legal effect unless Parliament so legislated.[140]

The Peace of Utrecht comprised a series of treaties among the belligerent powers. Among those, Britain and France concluded a treaty of peace and friendship to end hostilities, as well as a related treaty of commerce. While it was generally accepted that the peace treaty itself fell within the royal prerogative “and so could be presented to Parliament as a fait accompli, implementation of the commerce treaty’s principal provisions required legislation.”[141] More specifically, articles 8 and 9 of the commerce treaty would have extended most-favored-nation status to France and would have eliminated the increased duties that Britain had imposed on France over the course of the preceding decades.[142] The House of Commons rejected that legislation on June 18, 1713.[143] As a consequence, the British government was unable to ratify those central provisions of the treaty.[144]

The Commons vote came on the heels of two competing trends. First, British commercial interests, in particular the wool and wine industries, opposed the commerce treaty because it would expose them to French competition that would ultimately undermine British commercial interests.[145] The second was the government’s own assessment of Britain’s geopolitical interests. Those interests were twofold. The government, of course, was using the commerce treaty as part of an inducement to end the war with France. But by liberalizing trade with France, the government also sought an advantage over the Netherlands, which it viewed as its chief economic rival.[146] The vote on the commerce treaty thus pitted the executive’s assertion of foreign affairs imperatives against the domestic economic interests impacted by the regulation of foreign commerce.

The parliamentary vote against the bill to implement the commerce treaty thus stands for two propositions. First, parliamentary control of commerce prevailed over royal or governmental powers concerning treaties, war, and peace. That the commerce treaty was closely tied to both the end of a war and broader British foreign policy objectives did not diminish parliamentary control over foreign commerce. Second, both politically and legally, the eighteenth-century British citizen and politician understood that domestic economic interests could take precedence over foreign policy concerns where Parliament so determined. In short, the royal prerogative could not intrude upon parliamentary authority in matters of commerce. Where foreign affairs and foreign commerce were in tension, it was Parliament, not the monarch or government, that held the final authority to determine the proper ordering of those priorities.

B. A Confederation Without Coordinated Commerce

Given the British tradition of parliamentary control of commerce, even when Parliament’s regulation of commerce intersected with the Crown’s administration of foreign policy, legislative primacy in matters of commerce was the starting point for the American Republic. The question for the new nation was not whether an executive should have a say over foreign commerce; the issue was whether that power should reside in the federal or state legislature(s).

The Articles of Confederation addressed this problem. The Articles explicitly gave the federal Congress the authority to enter into treaties “provided that no treaty of commerce shall be made whereby the legislative power of the respective states shall be restrained from imposing such imposts and duties on foreigners as their own people are subjected to, or from prohibiting the exportation or importation of any species of goods or commodities whatsoever.”[147] The Articles prohibited states from imposing duties in violation of treaties made or “already proposed” by Congress with France and Spain,[148] but also required that any treaties be ratified by nine states before coming into force.[149] In short, the power to regulate foreign commerce and set duties on imports was primarily an issue for the states, with Congress playing a secondary role largely limited by state prerogatives.[150]

This arrangement quickly became untenable for managing the states’ effective cooperation, revenue collection, and unity. In particular, there were insufficient funds for an army and insufficient funds to command respect on the world stage.[151] George Washington complained that it was impossible to make commercial treaties under the Articles of Confederation.[152] John Jay argued that the conflicting commercial interests of the North and South would lead to separate treaties with foreign powers and make them satellite states to European powers.[153] Madison realized that, for the economy to function, the regulation of commerce had to be under a single authority.[154]

More generally, under the Articles, Congress did not have sufficient power over foreign commerce to enable the national government to credibly engage with other nations.[155] During this time, nearly all the colonies enacted their own trade laws with different motivations among them. In the 1780s, there were extensive difficulties in regulating tariffs across those different states in relation to the federal government’s needs. These related in part to revenue but also to formulating a national response to the foreign commercial policies of other nations.[156]

For instance, in 1783, the British excluded American ships from the British West Indies.[157] Virginia wanted to retaliate, and its legislature voted to authorize Congress to do so by prohibiting imports from the British West Indies.[158] The need for the support of the other states themselves, however, doomed Virginia’s proposal.[159]

Calls for centralizing the foreign commerce power did not involve the creation of a central executive to manage foreign commerce. Rather, the difficulties under the Articles of Confederation prompted calls to transfer control over foreign commerce to Congress. St. George Tucker, an American officer during the Revolutionary War and later a law professor at William and Mary, wrote that the failure of Virginia’s proposal was seen as evidence of “the propriety of investing congress with power over” foreign commerce.[160] In 1784, Congress appointed a committee to examine possible amendments to the Articles of Confederation, specifically with respect to the allocation of the commerce power.[161] Future President James Monroe led the committee, which in 1785 recommended that Congress have “‘the sole and exclusive’ authority of ‘regulating trade of the States, as well with foreign nations, as with each other, and of laying such imposts and duties upon imports and exports as might be necessary for the purpose.’”[162] The report did not go anywhere at the time, in part due to regional differences over commercial policy.[163]

But momentum for granting Congress authority over commerce was building. Alexander Hamilton, Thomas Jefferson, and James Madison in particular became strong advocates for reform and ultimately important designers of Congress’s foreign commerce powers. Nor was there any doubt that the Framers understood that congressional primacy over foreign commerce involved primary responsibility for an issue central to diplomacy and national security. Reflecting on that period several decades later, James Madison wrote that if the federal government had not developed the power to regulate foreign commerce, the American people would “present the solitary and strange spectacle of a nation disarming itself of a power exercised by every nation as a shield against the effect of the power as used by other nations.”[164] During the subsequent ratification debates, Hugh Williamson described the relationship between the authority to regulate commerce and national security even more bluntly: Because of its inability to make commercial treaties and regulate trade consistently, the United States had become “the prey of every nation . . . like a dark cloud, without cohesion or firmness . . . ready to be torn asunder and scattered abroad by every breeze of external violence, or internal commotion.”[165] Even those opposed to ratification agreed that “[w]e were suffering from the restrictions of foreign nations, who had shackled our commerce, while we were unable to retaliate: and all now agreed that it would be advantageous to the union to enlarge the powers of Congress; that they should be enabled in the amplest manner to regulate commerce.”[166] As St. George Tucker put it, “[N]o opposition was ever made, to these branches of the authority of congress, when the question respecting the adoption of the constitution of the United States was agitated.”[167]

The topic of foreign commerce thus became a uniting force for delegates to the Constitutional Convention, regardless of their otherwise disparate positions on questions of federalism and separation of powers. Rufus King reported in his Convention notes for June 16, 1787, that “[e]very one is impressed with the idea of a general regulation of trade and commerce. Can congress do this?”[168] Constitution skeptic Roger Sherman included it in the “few” authorities that he thought the new federal government should possess.[169] Alexander Hamilton similarly enshrined it as one of the four “principal purposes” of the federal government in The Federalist No. 23.[170] During the ratification debates, William Bingham posited that the proposed power over foreign and domestic commerce would attract even critics to the federalist cause.[171] In short, granting Congress control of foreign commerce—including its diplomatic and national security dimensions—became an animating purpose of the Constitutional Convention.

C. The Debate, Drafts, and the Text

Debates at the Constitutional Convention picked up where concerns with the Articles of Confederation left off. While there was a consensus that Congress should be given power over foreign commerce, the debates evidenced concern that the way in which Congress exercised that power should reflect the broad range of the new nation’s economic interests. No single actor or faction should be able to dictate foreign commercial policy.

Much of the debate at the Convention focused on crafting a compromise between Southern and Northern states around the regulation of commerce.[172] In 1787, Madison, in a letter to Jefferson, wrote that the question of congressional power over trade could be disaggregated among power over exports, power over imports, and power over enslaved people.[173] While primarily about federalism, these debates still helpfully reflect some of the content that the drafters sought to include in their definition of “commerce.” Madison recounts how some delegates sought to give Congress “unlimited power” over all these subjects: Some sought a qualification of power excepting exports and enslaved people; some wanted to exempt exports alone; and some demanded a two-thirds majority in both houses over the commerce power.[174]

Specifically, South Carolina delegate, Charles Pinckney, proposed that any legislation related to commerce should have a two-thirds majority vote to pass.[175] This position was tied to the southern states’ desire for free trade, low shipping costs, and legal protection of slaveholding and slave trafficking.[176] However, Pinckney was not successful, as the proposal was vigorously opposed by Alexander Hamilton, James Madison, and delegates from Pennsylvania and Massachusetts.[177] That being said, the proponents of supermajority protections for state commercial interests did prevail in another clause relevant to commerce: Article II’s Treaty Clause, which requires the consent of two-thirds of the Senate.[178]

As Delahunty writes, “[t]he debate at the Philadelphia Convention [was] over whether a bare majority or a supermajority of each House was required to enact foreign commerce regulations[, which] demonstrates that the Framers intended such regulation to be made by a legislative body, rather than an executive or judicial one.”[179] In modern separation of powers terms, the issue was an intra-branch issue over how easy it should be for the legislature to regulate commerce. There was no interbranch debate about the President’s role. Indeed, presidential power over commerce would have been inconsistent with the Framers’ general concern that the commerce power should be exercised in a way that reflected the nation’s diversity of economic interests.

With respect to the substance of the power, the earliest proposed drafts stated that Congress would have “the exclusive Power of regulating Trade and levying Imposts.”[180] Shortly thereafter, subsequent drafts separated the revenue provisions from the rest of trade. For example, the New Jersey Amendments of June 15, 1787, gave Congress power to pass acts for raising revenue “by levying a duty or duties on all goods or merchandizes of foreign growth or manufacture imported into any part of the U. States.”[181] The draft also empowered Congress “to pass Acts for the regulation of trade & commerce . . . with foreign nations as with each other.”[182]

Later drafts ultimately separated the power to tax and regulate foreign commerce into two clauses, a distinction preserved in the final text.[183] The word “trade” was also dropped in those revisions.[184] By August 1787, the language “to regulate Commerce with foreign nations, and among the several states” was adopted without objection.[185]

The result was three clauses relevant to our understanding of the power over foreign commerce: the Tariff Clause in Article I, Section 8, Clause 1; the Foreign Commerce Clause in Article I, Section 8, Clause 3; and the Treaty Clause in Article II, Section 2, Clause 2. Notably, in the debates and records of the Constitutional Convention, the Framers made no mention of any executive role in Clause 1 and 3 of Article I, Section 8.[186] To understand the implications of that silence, it is necessary to return to the text and the historical meaning of its key terms. What did the phrases “to regulate,” “commerce,” and “with foreign nations” signify to the drafters? And what constitutional authority did they intend to confer through the Foreign Commerce Clause? The answers illuminate the scope of legislative authority and, by negative implication, the constitutional limits on the executive.

Crucially, the constitutional text disaggregates the foreign commerce power into two provisions: the Tariff Clause and the Foreign Commerce Clause. As noted above, these powers were conceptually and functionally united in the Framers’ earlier drafts, but the Framers later separated them. Contemporary understanding treated tariffs as paradigmatic instruments of foreign commerce regulation. The Framers may have only formally separated them to clarify Congress’s revenue-generating powers.[187] Nevertheless, to avoid surplusage, the distinction must mean that the two—Congress’s power to impose duties and Congress’s power to regulate commerce with foreign nations—are not entirely coterminous. This reading is important in thinking about Congress’s subsequent delegations to the executive branch, to which we return in Part V.

What, then, did the Framers intend by granting Congress the power “to regulate Commerce with foreign Nations”? A closer textual and historical examination of the Clause’s constituent terms—“commerce,” “regulate,” and “with foreign nations”—offers clarity. Modern scholarship has helped reconstruct the Clause’s original meaning. In his important work on the Commerce Clause, Randy Barnett demonstrates that the term “commerce” was used with remarkable consistency throughout the Founding Era to denote “exchange” or “trade”—as distinct from production or agriculture.[188] Contemporary dictionaries and the writings of the Framers confirm that “commerce” was broadly understood as economic intercourse.[189] James Madison himself noted the interchangeability of “trade” and “commerce” in both public discourse and formal writings.[190] On this view, the regulation of commerce with foreign nations would naturally include laws governing the importation and exportation of goods, the terms of navigation, and the conditions under which mercantile enterprises operated.

Albert Abel similarly identified three principal categories encompassed within the concept of foreign commerce: the fiscal regulation of imports and exports (i.e., duties), rules governing navigation, and the structure of mercantile enterprises.[191] According to Abel, historical evidence suggests that these domains were the only ones widely regarded as falling within the scope of the Commerce Clause at the time of its ratification.[192] To this, other commentators have added contemporaneous concerns over industry protection.[193] Economic nationalism, particularly in the form of nascent protectionist measures, animated much of the Founders’ trade thinking. Legislative attention to “unemployed sailors, shipwrights, and iron workers” reflected less of a concern for the welfare of individuals in those trades and more of a broader ambition to fortify domestic industries vulnerable to British competition.[194] In this light, the Clause was a flexible tool designed to empower Congress to respond to foreign economic threats, including, when necessary, through retaliatory measures.[195] As Madison would later note, the regulation of commerce was instrumental to advancing American industry and ensuring reciprocal treatment abroad.[196]

The final textual element—“with foreign Nations”—is particularly revealing. It confirmed that the Framers understood Congress to play an integral role in shaping the nation’s commercial diplomacy. Far from relegating international economic policy to the executive branch, the Constitution assigned Congress a central role in policing “extranational traffic.”[197] While the executive would enforce these policies, the legislature would define them. Congress was vested with the power to engage the world through commercial regulation, exercising sovereign judgment over both domestic and international economic ordering. This understanding underscores a broader constitutional logic: The power to regulate foreign commerce was not merely ancillary to fiscal policy but a cornerstone of the new nation’s international authority. Insofar as the executive now claims independent power to determine trade policy, such assertions stand at odds with the history, drafts, final text, and design of the Constitution.

D. The Post-Ratification Practices of the Founding Generation

Because our interest lies chiefly with determining whether the executive branch can claim any constitutional power over foreign commerce—particularly when there is a tension between Congress and the President’s foreign affairs powers—our examination of the Republic’s early years focuses on episodes where the President’s Article II powers intersected with Congress’s Article I authority. This overlap arose in two ways. First, the President, acting with the Senate’s advice and consent, could employ the treaty power to make commercial treaties that narrowed the practical significance of Congress’s foreign commerce powers. Second, the question arose as to whether the President’s other constitutional authorities in foreign affairs, particularly in periods of international armed conflict, might take precedence over Congress’s authority to regulate commerce during such times.

As one of us has observed, in the Republic’s earliest days, as a practical matter “[n]either branch of government had a clear monopoly as both worked to create alliances and revenue for the newly independent United States.”[198] After the Constitution’s ratification, one of Congress’s first acts was to establish a tariff schedule and create a customs agency to implement it.[199] For its part, the executive branch negotiated treaties of friendship, navigation, and commerce with several nations.[200] Those agreements conferred on treaty partners special trading privileges, especially with regard to tariffs.[201] Through the advice and consent process, the President and the Senate jointly fashioned trading rules that supplemented the statutory framework enacted by the full Congress.[202] Yet, as one historian noted, during these years there already existed a “perennial anxiety” that “a president could transform his diplomatic authority into a virtual power to legislate” with the tariff “topp[ing] th[e] list” of potential problem areas.[203] This concern suggests that, although responsibility for foreign commerce was in practice shared, the Founding Generation understood that the treaty power was the primary constitutional vehicle through which the President could exert influence over the regulation of foreign commerce.[204]

In the early years of the Republic, the question of whether the President and the Senate could use treaties to set tariff rates and establish other commercial regulations without the ordinary participation of the House, was as much an intra-branch issue as an interbranch one. The Jay Treaty (1794) with Great Britain, the Embargo Act of 1807, and the Nonintercourse Act of 1809 were pivotal in determining the boundaries between the treaty-making power and congressional authority to regulate foreign commerce. Much of that debate took place in Congress itself, rather than in the courts. The Federalists generally took the position that Congress’s authority to regulate commerce did not give Congress the power to approve commercial treaties. Jeremiah Smith (Federalist – N.H.) asserted:

When it is said that Congress shall have the power to do certain things, for example, to regulate commerce with foreign nations, it means no more than this, that Congress shall have all Legislative power over this subject, not the entire power over foreign commerce is given to Congress.[205]

On the other hand, the Democratic Republicans held that it was the duty of Congress to assess the constitutionality of treaties made by the President through their Commerce Clause authority. William Lyman (Democratic Republican – Mass.) expounded that:

The present question . . . was not, whether the House should make Treaties, but whether the President and Senate should make laws; all the power contended for on the part of the House was the power of self-preservation; it was a repelling power, a power to prevent the President and Senate, under the color of making Treaties, from making all the laws.[206]

Although the House eventually stood down from blocking the Jay Treaty, the principles behind congressional authority to approve commerce treaties were sufficiently unsettled that, in 1816, a standoff between the House and Senate occurred over another commercial treaty with Great Britain. This too ended in a retreat for the House, but not a total surrender of its foreign commerce power over commercial treaties.[207]

The other way the President’s constitutional powers implicated foreign commerce was when commercial issues intersected with the President’s foreign affairs powers. European wars sparked by the French Revolution caused an episode in which the President’s power over foreign affairs and Congress’s power over commerce were put to the test. In April 1793, President Washington issued his famous Neutrality Proclamation, which, in part and as relevant here, prohibited “carrying to any [of the belligerent nations] those articles which are deemed contraband by the modern usage of nations.”[208] This limitation was a response to reports that Great Britain intended to close French ports to U.S. commerce.[209] The question was thus how the United States should respond. Secretary of State Thomas Jefferson hoped that the United States would respond by closing U.S. ports to exports from nations participating in the blockade of France.[210] As the name suggests, the proclamation ultimately adopted a more evenhanded approach.

The constitutionality of the Neutrality Proclamation was the subject of an exchange of essays in 1793 by Alexander Hamilton (writing as Pacificus) and James Madison (writing as Helvidius). Hamilton’s essays defending the proclamation almost completely overlooked the issue of commerce within the broader foreign affairs framing. His first essay, which most directly took on the constitutional question, advanced the now-familiar argument that Article II vests the President with “the executive power” while Article I vests Congress only with the “legislative powers herein granted.”[211] Hamilton argued that the executive power thus included powers related to war and peace that were “subject only to the exceptions and qu[a]lifications” in the Constitution.[212] Those qualifications, he thought, were the Senate’s role in treaty-making and appointments, and Congress’s power “to declare war, and grant letters of marque and reprisal.”[213] Moreover, Hamilton argued that it was for the President, as the organ of the United States in foreign affairs, to interpret the treaties of the United States and implement them.[214]

The main thrust of Madison’s response was to push back against the expansive characterization of executive power and, more specifically, to argue that some of the powers that Hamilton characterized as “executive”—such as treaty-making and declaring war—were better characterized as legislative powers.[215] Throughout the essays, Madison made clear that he was primarily concerned with war powers, implicitly viewing declarations of neutrality as the flipside of the power to declare war.[216]

Notably, although both Hamilton and Madison discussed the division of foreign affairs powers between the political branches extensively, neither discussed Congress’s power to regulate foreign commerce as relevant to that division. Hamilton, for instance, did not list the power to regulate foreign commerce and impose tariffs as an exception to the executive power—although it is naturally encompassed in the “exceptions and qualifications” to, or limitations on, executive power in the Constitution. At the same time, there is no doubt that he understood that the Neutrality Proclamation affected a regulation of commerce. In one of his later essays, he justified the timing of the proclamation—coming only after Britain entered a war that Prussia and Austria were already fighting with France—as necessary given the threat the British (as well as the Dutch and Spanish) navies posed to U.S. commerce and the need of U.S. merchants to know how the United States would respond to the threat.[217] Likewise, Madison recognized the connection between the Neutrality Proclamation and commerce with France, but he limited his legal analysis primarily to powers over war, peace, and treaty interpretation.[218]

Prakash and Ramsey offer an explanation for why this debate overlooked the foreign commerce power. As they explain, the question Washington (and thus Hamilton and Madison) faced was whether it was constitutional for the President “to publicly pronounce the views and goals of the United States . . . on international matters, even though that policy might contradict or go beyond existing laws.”[219] Critically, though, “the President’s policy statement, of itself, was not binding” except on those working in the executive branch.[220] Debates over the constitutionality of the Neutrality Proclamation thus had little to do with the President’s ability to make law, including with respect to the regulation of commerce. Prakash and Ramsey argue that the Pacificus–Helvidius debate is properly understood as “turn[ing] on the (fairly academic) question of whether the President could by his declaration bind Congress’s subsequent ability to decide to enter the war.”[221] The President’s ability to enforce his Proclamation was limited to those constitutional and statutory tools otherwise available to him. Until Congress enacted the Neutrality Act of 1794, as well as two statutes imposing or authorizing embargoes in March and June of 1794, Washington only enforced his policy using diplomacy, persuasion, and prosecutions under existing criminal laws, the last of which were widely viewed as constitutionally suspect.[222]

Nor did these statutes ratify a broad view of presidential power. To be sure, with respect to foreign commerce, the June 1794 Embargo Authorization Act authorized the President to impose an embargo “whenever, in his opinion, the public safety shall so require.” [223] This was a broad predicate that foreshadowed later delegations predicated on executive branch policy determinations. But Congress limited the President’s power significantly in another way. It provided that “[t]he authority aforesaid shall not be exercised, while the Congress of the United States shall be in session” and that any presidential embargo would automatically expire fifteen days after Congress reconvened.[224]

In other words, the neutrality episode demonstrates a President who recognized that he could not create binding U.S. law, including regarding foreign commerce with belligerents, and a Congress that reserved for itself that power, delegating to the President the power to act only in its absence. As Prakash and Ramsey put it, “[T]he executive power did not extend to the creation of obligations under domestic law, and thus the domestic legal system could not be invoked to enforce a unilateral presidential policy having no other basis in law.”[225] Where lawmaking regarding foreign commerce was concerned, the President’s constitutional powers yielded to Congress’s.[226] Thus, the Neutrality Proclamation, which has long been a seminal artifact in thinking about the relationship between the allocation of foreign affairs powers across the branches,[227] is properly understood as endorsing congressional supremacy where domestic lawmaking is concerned. Washington did not assert that his proclamation had any independent legal force outside of the executive branch, and while Congress eventually endorsed Washington’s policy, it was Congress’s action that ultimately provided the legal basis for the policy.[228]

Still, one might wonder whether the Founding Generation thought the answer was the same where the President’s enumerated foreign affairs powers, such as the commander-in-chief power, were implicated. Which branch prevailed when the President wished to give an order to or defend the military and Congress had not legislated or legislated to the contrary? The Founding Generation’s answer was that Congress prevailed. At least three examples demonstrate congressional primacy. The first is the Supreme Court’s decision in Little v. Barreme (1804).[229] That decision arose during the context of the Quasi-War between France and the United States from 1798 to 1800. During the conflict, Congress annually passed statutes that aimed to cut off commercial relations with France.[230] The 1799 statute authorized the President to seize any vessel that, upon a search, was determined to be “bound or sailing to any port or place within the territory of the French Republic or her dependencies.”[231] The Secretary of the Navy, however, ordered naval captains to seize ships sailing to or from a French port.[232] Captain Little of the Boston seized the Flying Fish, a Danish vessel sailing from a French port and sailed it to Boston, where it was condemned as a prize. The question the Supreme Court faced was whether Captain Little could seek shelter from liability because he was following military orders, even though those orders conflicted with the terms of Congress’s statute. Writing for the Court, Chief Justice Marshall held that Captain Little could not escape liability. The terms imposed by Congress took precedence over military orders, even when those orders were followed in good faith.[233]

Understood as a case about the respective war powers of the political branches, Little v. Barreme has sometimes puzzled commentators. But as Sindak argues, the better understanding of the episode is that Congress prevailed because the question was one of foreign commerce.[234] The statute at issue regulated commerce with France, and while it authorized the President to use the military to enforce Congress’s commercial rules, the fact that Congress had acted pursuant to the Foreign Commerce Clause meant that it acted in an area in which the executive branch lacked constitutional authority.

Practice during the first decade of the eighteenth century further confirmed congressional control of commerce during hostilities. In particular, it was unclear whether the President had the authority to lay down an embargo as a matter of war and national security, or if Congress had the authority through its taxation power or through the regulation of commerce.[235] During the Napoleonic Wars, as Britain and France targeted U.S. commercial vessels, the Jefferson administration looked to Congress to impose embargoes. It was the Embargo Act of 1807 that finally brought the Foreign Commerce Clause squarely into the courts in United States v. The William (1808). The Embargo Act itself illustrates congressional primacy even where matters of foreign affairs were concerned. In the Act, Congress itself directly imposed an embargo “on all ships and vessels in the ports and places within the limits or jurisdiction of the United States.”[236] The President was charged with enforcing the embargo and given the ability to make exceptions, but the act leaves no doubt that Congress itself chose and enacted the United States’ policy.[237] Indeed, when the initial embargo proved liable to evasion, Congress itself passed two supplementary acts, rather than rely only on the enforcement authority delegated to the President.[238] The first act required fishing vessels to post a bond guaranteeing that they would not be conduits for evading the embargo, and the second prohibited export by sea or land.[239]

The William was a vessel seized by the government for allegedly participating in a smuggling operation in violation of the embargo. The ship owners challenged the constitutionality of the embargo on the grounds that the Constitution did not empower Congress “to enact laws, so general and so unlimited, relative to commercial intercourse with foreign nations, as those now under consideration.”[240] The District Court of Massachusetts rejected this argument and held that

the power to regulate commerce is not to be confined to the adoption of measures, exclusively beneficial to commerce itself, or tending to its advancement; but, in our national system, as in all modern sovereignties, it is also to be considered as an instrument for other purposes of general policy and interest.[241]

In other words, Congress’s commerce powers extended to regulating commerce during military conflicts.[242]

The final set of examples involves the imposition of embargoes on trade with Britain in the lead up to and during the War of 1812 under Presidents Jefferson and Madison. The Embargo Act of 1807 applied generally to all nations.[243] In the last days of his administration, President Jefferson signed the Nonintercourse Act of 1809.[244] The Nonintercourse Act did three notable things. First, it ostensibly narrowed the embargo to apply only to Britain and France and provided that it would expire at the end of Congress’s next session.[245] Second, it authorized the President to lift the embargoes on Britain and France if they stopped targeting U.S. vessels.[246] Third, it repealed the Embargo Act of 1807 in two stages: Beginning in March 1809, it permitted U.S.-owned vessels to resume exports to non-embargoed countries, and it repealed the general embargo on foreign-owned vessels effective from the next Congress.[247]

The Madison administration’s effort to stay within its delegated authority was evident from its reaction to the way Congress wrote the repeal provisions of the 1809 Act, even if it interpreted its discretion liberally. Upon taking office, President Madison, after receiving assurances that Britain would stop targeting U.S. vessels, was determined to lift the embargo against Britain, pursuant to his delegated authority under Section 11 of the Act.[248] This created an awkward situation. Although the 1809 Act repealed the 1807 general embargo, it delayed the effectiveness of the repeal with respect to foreign nations other than Britain and France until the next session of Congress.[249] Albert Gallatin, the Treasury Secretary, noted that this arrangement created a situation in which British shipping, the primary concern of the 1809 Act, would be treated better than the shipping from nations with whom Congress was not concerned but who remained subject to the 1807 embargo for two more years.[250] In a message to Congress, Gallatin described this state of affairs and noted that executive discretion to correct it would be “consistent with the presumed intention of the Legislature.”[251] While acknowledging congressional supremacy on the issue, he also “submit[ted] the propriety of acting immediately on the subject [i.e., of allowing ships from friendly nations to depart U.S. ports in trade], and without waiting for a decision on the more complex modifications of the non-intercourse act which the late change in our foreign relations may render necessary.”[252]

But if Gallatin pushed the limits of executive discretion in the administration of the Embargo Act, Madison ultimately acknowledged congressional primacy. In August 1809, after the British went back on a promise to cease targeting U.S. vessels, Madison revoked the repeal of the embargo that he had previously authorized under the 1809 Act.[253] In the years leading up to the War and after it began, Madison asked Congress to authorize embargoes when he felt they would be useful. The most significant such request came in July 1813, when Madison requested a general embargo of British goods, a measure that Congress ultimately enacted in December 1813.[254] When Napoleon was defeated in 1814, undermining what efficacy the embargo might have had, Madison again turned to Congress to repeal it. Madison remained, in other words, deferential to Congress’s ultimate authority to regulate commerce, even in times of war.

The lesson to be drawn from the embargo acts is that, even though it involved the suspension of commerce in response to predations against U.S. ships by France and Britain, the President acted purely as an agent of Congress. The embargoes were imposed by Congress and the power exercised by the President was purely delegated authority, notwithstanding the context. Nor did the President’s delegated authority involve policy determinations. Rather, Congress authorized the President to make exceptions to its congressionally determined policy, and Congress made contingent choices as to when the embargo should be lifted. The President was charged with determining the factual predicate upon which Congress’s will depended, but, as the Supreme Court recognized, it was Congress “exercis[ing] its discretion in reviving” a previously imposed embargo, notwithstanding the President’s fact-finding role.[255]

***

The original history of the foreign commerce power—both the power to tax imports and to regulate foreign commerce—is one of legislative supremacy. That legislative supremacy was hard won and well-established in eighteenth century Britain and applied even to commercial regulations linked to the resolution of major conflicts. Bookending the Founding Era, Jefferson, Madison, and John Marshall each acknowledged that executive power over foreign affairs, and in particular military conflicts, did not override Congress’s explicit regulation of foreign commerce. When the President’s foreign affairs powers intersected with Congress’s foreign commerce powers, the latter prevailed as a matter of constitutional law. As a matter of practice, though, the President enjoyed some degree of deference from Congress when he sought foreign commercial legislation, as demonstrated by the Jeffersonian and Madisonian embargoes. That degree of deference would, as we explain in the next Part, grow over time.

IV. The Rise of the Trader in Chief

The preceding Part established that the Constitution’s design—confirmed by both the records of the Constitutional Convention and the early practice of the national government—did not allocate power over foreign commerce to the executive, even when foreign commerce was intermingled with military or larger diplomatic issues. This Part examines the subsequent century, spanning roughly from 1830 to 1930, to assess whether that original allocation of authority endured in practice. As we show here, the historical record reveals a pattern of executive involvement in foreign commercial affairs. But the President did not regularly claim constitutional authority over those affairs.[256] To the contrary, executive actions in foreign commerce were invariably grounded in legislative instruction. The President’s participation in trade policymaking was neither autonomous nor constitutionally self-justifying; it was contingent on statutory delegation and remained subject to congressional oversight.[257] The result is a consistent historical practice—for the first half of the nation’s history—of congressional primacy in foreign commerce, and a marked lack of any constitutional claims of executive power to act independently or to interpret statutes broadly on constitutional grounds.

A. Delicate Delegations: Conditional Tariff Adjustments

During the early decades of the Republic and continuing throughout the nineteenth century, presidential involvement in regulating foreign commerce operated within a clear and consistent framework: The executive acted only pursuant to express statutory delegations of authority. This framework took root almost immediately. Between 1794 and 1830, Congress enacted at least a dozen statutes granting the President authority to execute commercial policies that Congress formulated.[258] These delegations were narrow in scope and highly conditional, empowering the President to amend, suspend, or apply particular rules of trade “in the interest of the United States,”[259] or “if in his judgment the public interest should require it.”[260] Acts passed in 1794, 1798, 1799, 1806, 1807, 1809, 1815, 1817, 1824, 1828, and 1830 uniformly reflected this fiduciary model.[261]

These statutes were not open-ended grants of foreign commercial discretion. Rather, they reflected Congress’s view that presidential action required a conditional, factual trigger tethered to circumstances specified ex ante by the legislature. This retail-level delegation preserved legislative supremacy while affording the executive a narrow margin for reactive implementation.[262] The limited duration of these delegations—many expired within a few years—underscored Congress’s continuing control over commercial policy and its refusal to yield that domain wholesale. In each of these cases, the President was simply authorized to determine when the circumstances contemplated by Congress had arisen and to issue a proclamation to put the congressionally prescribed action into effect.[263] That structure remains embedded in contemporary trade statutes, particularly those permitting tariff adjustments following formal findings by the executive. In this sense, the trustee-style model developed during this period did not fade with time but continues to structure the boundaries of lawful presidential trade action today.[264]

In the decades preceding the Civil War, debates over trade policy largely unfolded along partisan lines, with Presidents adhering closely to their respective party platforms.[265] Congress alternated between enacting protectionist and more liberal tariff regimes, and the executive branch consistently implemented the legislative will without substantial resistance. While Presidents occasionally shaped party rhetoric—most visibly through addresses to Congress or on the campaign trail[266]—interbranch conflict over tariff policy remained minimal.[267] As sectional tensions mounted in the years before the Civil War, Presidents remained disinclined to contest congressional prerogatives in this domain. For instance, President Lincoln viewed tariff policy primarily as a fiscal instrument to support wartime expenditures.[268] One of his successors, President Grover Cleveland, is known to have said to a supporter at the time of his election that he knew “nothing about the tariff.”[269] Years later, President Cleveland devoted a congressional address to the topic in which he implored Congress to reconsider tariff rates, leading to a congressional battle over tariffs in the 1888 election season.[270] These Presidents relied on Congress and worked with it to regulate foreign commerce and to make the necessary adjustments to tariff rates for the nation’s foreign affairs and economic priorities.

By the end of the nineteenth century, tariff-rate-adjustment delegations expanded the President’s discretion. The McKinley Tariff Act of 1890 allowed the President to remove individually listed products from various tariff classifications and apply different duties.[271] Even then, the President was not given wholesale discretion to set tariffs. The Act required the President to suspend certain duty-free provisions of the law where he found that a trading partner was imposing unequal and unreasonable conditions on those imports that did produce such goods with a view “to secure reciprocal trade with countries.”[272] The Supreme Court confirmed the constitutional integrity of this delegated discretion, famously finding that Congress may “delegate a power to determine some fact or state of things upon which the law makes, or intends to make, [Congress’s] action depend. To deny this would be to stop the wheels of government.”[273]

Over the next 25 years, Congress granted the President greater authority to adjust tariff rates.[274] The Payne–Aldrich Tariff Act of 1909 authorized the President to make certain adjustments to tariff rates by proclamation.[275] And Presidents did so, issuing 134 proclamations with broad reach across the commercial world.[276] But each proclamation was limited to the life of the Act and directed toward a particular country and good.[277] The Payne–Aldrich Act also permitted the President to create a Tariff Board to investigate certain trends in the application of tariffs, but Congress then terminated the Board when control of both houses changed in the 1912 election.[278]

In 1917, when the United States entered World War I, it became necessary once again to consider how to regulate foreign commerce with hostile nations. Congressional primacy over commerce during times of war remained uncontested. Congress passed the Trading with Enemy Act in October 1917,[279] just months after Congress issued a declaration of war. The statute closely resembled the Embargo Acts of a century earlier. It directly imposed restrictions on trading with nationals of countries on whom Congress had declared war and delegated some discretion to the President in the administration of those restrictions.[280]

Despite these delegations, both the President and the Congress were clear on the matter that Congress was responsible for tariff policy and foreign commercial policy more generally.[281] All these delegations came in the form of legislation passed by Congress—legislation that set tariff rates for all U.S. imports, leaving room only for small adjustments based on presidential factfinding, and even directly regulating commerce with enemies during times of war.

B. Discriminating Deals: Trade Executive Agreements

During this same period (1830–1930), the United States began to explore novel forms of foreign commercial regulation, increasingly turning to trade-related treaties with key partners that extended beyond the traditional friendship, commerce, and navigation agreements that characterized the Republic’s early diplomatic practice.[282] The U.S. Tariff Commission, an independent executive branch agency, described U.S. trade policy as “opportunist,” divorced from foreign policy activities, but motivated by the nation’s increased global leadership.[283] Once the United States committed to “far-reaching participation in world politics,” wrote the Commission, it could no longer treat its commercial negotiations as bilateral, isolated conversations in which the President would evaluate other nations’ foreign commercial practices and react accordingly.[284] A practice in which the President develops “special arrangements” would be “troublesome,” according to the Commission.[285]

These specialized trade arrangements also began to appear in treaties negotiated by the executive during the early nineteenth century. Notably, an 1831 treaty with France—intended to resolve outstanding claims arising from the Napoleonic Wars—included a U.S. commitment to reduce duties on French wines.[286] Although this concession was one of several reciprocal obligations, it drew the attention of other foreign governments who soon pressed for comparable treatment of their own exports to the United States in subsequent negotiations.[287]

A decade later, the Tyler administration negotiated the first so-called reciprocal trade agreement with what is now Germany.[288] German officials complained to the administration about the special treatment toward the French in the 1831 treaty and informed President Tyler’s envoy that they were unwilling to make any changes in the German tariff rules unless accompanied by corresponding reductions from the United States.[289] The Tyler administration proceeded with the negotiation of the treaty, but, as highlighted in the prior Section, the domestic politics of tariff rates were complex and the President was not empowered to dole out special reductions without congressional approval.[290]

The negotiation of the trade treaty with Germany was one of the earliest times that a President unilaterally departed from congressional will on trade, or at least sought to do so—and the result was decisive. The Senate refused to ratify the agreement and others like it. The failed treaty with Germany was one of seven that were negotiated but not brought into force due to resistance in the Senate.[291] The Senate Foreign Relations Committee released a critical report that advised rejection of the agreement on the basis that the President had no constitutional authority to regulate foreign commerce. The Committee asserted that “the control of trade and the function of taxing belong, without abridgment or participation, to Congress” and that “[r]epresentatives of the people . . . may better discern what true policy prescribes and rejects, than . . . the executive department.”[292] “The appropriate function of the executive, the committee said, was ‘to follow, not to lead; to fulfill, not to ordain, the law . . . not to go forward with [a] too ambitious enterprise.’”[293] In the following years, the Senate rejected almost every agreement that sought to regulate foreign commerce by lowering tariff rates.[294]

As congressional interests shifted in favor of more liberal trade policy, and in light of pressing economic and political issues facing the nation, policymakers entertained not just greater presidential discretion to adjust tariff rates but also new means of adopting reciprocal reductions in tariff rates and other barriers to trade. For example, the United States concluded reciprocal agreements with six Latin American countries, Austria-Hungary, Germany, Great Britain (for the West Indies), and Spain (for Cuba and Puerto Rico).[295] Although Congress did not formally approve these agreements, it did choose to terminate them in 1894 when political control shifted toward a more protectionist policy.[296] Partisan control of the Congress flipped again in 1896, prompting the more liberal Dingley Tariff Act of 1897. That Act invited the President to negotiate reciprocal trade treaties.[297] The executive branch negotiated six treaties, but none were ratified.[298] More restrictive trade agreements were concluded with nine countries under another directive in this Act, but they were terminated by Congress in 1909.[299]

The historical record of congressional-executive interaction over trade agreements during this period reflects a consistent institutional pattern: The President negotiated agreements, and Congress—principally through the Senate—evaluated the resulting texts, approving or rejecting them in full. In most cases, these negotiations were expressly authorized by statute, with Congress delineating the scope of the President’s authority in advance. When the executive exceeded those bounds, the legislature responded by refusing to ratify the resulting agreements. The substance of these arrangements—like tariff policy more generally—tracked the prevailing legislative sentiment, oscillating between protectionist and liberalizing impulses. Partisan dynamics shaped the broader discourse, as with domestic tariff debates. Yet unlike tariffs, where the President exercised some measure of discretion and policy leadership, his role in trade agreement negotiations was more narrowly circumscribed. The bounds of the President’s role were reinforced by the structural features of the treaty process itself: Trade agreements, unlike ordinary legislation, were indivisible instruments, subject to approval or rejection in full.

***

This Part illustrates how delegations of foreign commercial authority, while present from the early Republic, became a systematic feature of federal trade policy in the closing decades of the nineteenth century. Even then, such delegations were both demarcated and transactional: Congress regularly did initiate, direct, and revoke executive action. During this century-long period, no President claimed independent constitutional power over trade policy. Instead, Presidents regularly acknowledged Congress’s primacy, often soliciting legislative authorization before undertaking new foreign commercial initiatives. This pattern of interbranch exchange—characterized by negotiation, request, and response—suggests not a gloss-ratified expansion of executive power but a statutory framework built upon legislative supremacy.

The dynamic identified here gave rise to a “trader in chief” role for the President. But that role, to the extent it existed, was wholly derivative of Congress’s authority.[300] The President operated as a statutory agent, not a constitutional principal.[301] Authority to adjust tariff rates, to negotiate reciprocal trade arrangements, to coordinate with foreign governments on commercial matters, and even to suspend foreign commerce during times of war, was invariably rooted in express delegations or executed pursuant to the shared treaty power. No example from this period supports the proposition that the executive’s engagement in trade was predicated on an independent Article II mandate.

In sum, the distinctive feature of the 1830–1930 period is the institutionalization of legislative delegation and not the emergence of executive prerogative in foreign commerce. Over these decades, Congress began to vest the President with limited authority to alter tariff schedules and expanded his power to initiate trade negotiations, in some respects out of necessity for the operation of a global economy.[302] This Part has shown that these powers remained bounded by statute and policy, not by constitutional implication. The President remained a conduit for legislative design, not an architect of foreign commercial authority. The President’s supra-constitutional actions began to take shape in the second half of the twentieth century, as the next Part argues, and have now reached an inflection point.

V. The Statutory Bounds of Contemporary Commerce

From 1930 to the present, the federal government’s regulation of foreign trade has increased just as foreign commerce itself has expanded by leaps and bounds through innovation.[303] Advances in transportation and technology have increased the volume of cross-border trade and demanded greater oversight and facilitation by governments around the world. With those developments, Congress, the executive, and eventually the courts, have become more engaged in these matters than at the time of the Founding.

As this expansion began, so did the commentary that sought to shift the pendulum of foreign commerce powers toward the White House. Already in the early part of the twentieth century, commentators remarked that

[a]nother and very specific way in which the power of the President is exercised over commerce is through his special authority over foreign relations . . . . [T]he President . . . is obliged to respect the power of Congress over commerce, and must observe in commercial treaties the policies established by the laws of Congress . . . . [D]espite its importance, [Congress] is a changing mechanism and one which has little actual contact with commerce.[304]

From 1930 through the early 2000s, Congress delegated vast authority to the executive branch in foreign commerce.[305] Whether acting as Congress’s agent in executing U.S. trade policy or managing the entirety of the U.S. trade apparatus, the executive was charged with exercising considerable discretion in foreign commerce.[306] Although some of those delegations were challenged by aggrieved industries, no presidential administration tested the foundations of this shared system like the Trump and Biden administrations.

This Part considers the recent Presidents’ extraordinary views on the foreign commerce power. To be sure, the Trump administration’s take on an executive foreign commerce power is not unique nor limited to the Trump administration, though recent months have raised the stakes higher than ever before. We examine two tools used by the Trump administration: tariffs and trade agreements. Both tools have dominated the administration’s approach to global economic governance in the first year of Trump’s second term.[307] However, even if the administration were to abandon them tomorrow, the impact of their use will permeate both domestic and international law for the foreseeable future. In sum, by explaining how President Trump has sought to use trade agreements and tariffs to achieve national policy objectives, we show how a set of capacious delegations gave rise to strained constitutional claims over statutory territory.

A. Tariffs

Despite Congress’s constitutional ownership of tariffs, tariffs are at the center of President Trump’s “America First Trade Policy” in his second term.[308] The President has relied on a handful of statutes both to threaten and collect double- and triple-digit percent tariffs on imports from nearly every country in the world.[309] His aggressive anti-liberalization policy has upended global trade flows and supply chains to an unprecedented degree, far exceeding the scope and magnitude of the tariff program put in place during the President’s first term—which was already a notable departure from Congress’s longstanding trade posture. In this Section, we review the statutes on which the President has relied as well as his claims to foreign affairs authority as a means of extending those statutes’ reach.

The second Trump administration’s most impactful tariff action during its first year in office has been its imposition of what have been called “worldwide tariffs.”[310] On April 2, 2025, the President announced a 10 percent tariff on products from all countries and additional tariffs that vary by country of origin.[311] The President has claimed that these all-encompassing tariffs were authorized under the 1977 International Emergency Economic Powers Act (IEEPA). The IEEPA is a descendant of the 1917 Trading with the Enemy Act. In the IEEPA, Congress tried to cabin what had become the President’s routine use of the Trading with the Enemy Act to regulate foreign commerce. The IEEPA provides the President with certain extraordinary authorities that he may use to “deal with any unusual and extraordinary threat, which has its source in whole or substantial part outside the United States, to the national security, foreign policy, or economy of the United States, if the President declares a national emergency with respect to such threat.”[312] In this instance, the President declared an emergency regarding “a lack of reciprocity in our bilateral trade relationships, disparate tariff rates and non-tariff barriers, and U.S. trading partners’ economic policies that suppress domestic wages and consumption, as indicated by large and persistent annual U.S. goods trade deficits . . . .”[313] Having made an emergency declaration, the IEEPA empowers the President, in relevant part, to

investigate, block during the pendency of an investigation, regulate, direct and compel, nullify, void, prevent or prohibit, any acquisition, holding, withholding, use, transfer, withdrawal, transportation, importation or exportation of, or dealing in, or exercising any right, power, or privilege with respect to, or transactions involving, any property in which any foreign country or a national thereof has any interest by any person, or with respect to any property, subject to the jurisdiction of the United States.[314]

In addition to these global measures, the administration has implemented tariffs that are country- and product-specific. In February 2025, again relying on the IEEPA, President Trump announced the imposition of so-called trafficking tariffs on products from some of the United States’ biggest trading partners: Canada, China, and Mexico. He referred to an “unusual and extraordinary threat” from those governments’ failures to take sufficient actions to address unlawful migration to the United States and to prevent the entry of illicit drugs into the United States.[315] Similarly, in March 2025, the White House issued an executive order in which it authorized the Secretary of State to determine whether imports to the United States from any third country found importing oil from Venezuela ought to be subject to a 25 percent tariff.[316] The Trump administration also imposed 40 percent tariffs on goods from Brazil under the IEEPA due to Brazil’s prosecution of its former President Jair Bolsonaro,[317] as well as 25 percent tariffs on Indian imports in response to India’s continued purchase of Russian oil during the Russian war in Ukraine.[318] Although only the “trafficking” and “worldwide” tariffs were at issue in Learning Resources, the Supreme Court’s decision that IEEPA does not authorize tariffs would appear to foreclose all of these duties.[319]

But Learning Resources hardly means the end of extraordinary tariffs. In contrast to the tariffs imposed under the IEEPA, the special tariffs imposed during President Trump’s first term relied on delegations that first require an investigation and findings by an agency. These statutes enable the executive branch to raise tariffs when an agency identifies a threat to the U.S. economy from another government’s actions or a threat to U.S. national security arising from the imports of a particular product.[320] The administration has deployed these powers again in its second term. For example, the administration has relied on Section 232 of the Trade Expansion Act of 1962 to increase duties on imports of steel, aluminum, and their derivative products from all countries, as well as to impose new duties on products such as automobiles, certain automobile parts, copper, timber and lumber, and trucks.[321] Section 232 permits the President to “adjust imports” of an article where the Secretary of Commerce finds that the article is being imported “in such quantities or under such circumstances as to threaten to impair the national security.”[322] Additionally, in 2025, the administration has turned to Section 232 as the basis for investigations into imports of semiconductors, pharmaceuticals, critical minerals, commercial aircraft, jet engines, polysilicon, and unmanned aircraft systems.[323]

Relying on Section 301 of the Trade Act of 1974, Trump has imposed tariffs on products from Nicaragua,[324] imposed fees on certain foreign vessels,[325] and threatened additional tariffs on products from countries imposing digital services taxes[326] and on countries engaging in unfair trade practices regarding fishing.[327] Section 301 permits the Office of the U.S. Trade Representative (USTR) to impose duties or other import restrictions, where the USTR finds that an act, policy, or practice of a foreign country “is unreasonable or discriminatory and burdens or restricts United States commerce.”[328] The Trump administration relied on Section 301 during his first term to impose tariffs on most imports from China, tariffs that the Biden administration continued and even increased with respect to some products.[329] And, after the Supreme Court struck down the President’s IEEPA tariffs in Learning Resources, the administration opened new Section 301 investigations that it claimed could allow it to reinstate many of the IEEPA tariffs,[330] while in the meantime imposing a 10 percent duty on nearly all products from all countries under Section 122 of the Trade Act of 1974.[331]

These tariff orders have prompted many lawsuits, with plaintiffs arguing that President Trump exceeded his statutory authority.[332] In defending the President’s actions, the government has repeatedly argued that the President’s foreign affairs and national security powers require courts to construe those statutes broadly such that even explicit limits should not apply.[333] The Federal Circuit, which has exclusive appellate jurisdiction over appeals arising from laws authorizing tariffs, upheld the first Trump-term Section 232 tariffs against all challenges. Contrary to the historical understanding of the foreign commerce power, the court held in one case that “the President has some independent constitutional authority over national security and dealings with foreign nations” that applies to foreign commerce statutes, including when they are used to impose tariffs.[334] In 2021, after the Federal Circuit upheld Section 232 tariffs despite the fact that President Trump had imposed the challenged tariffs after a statutory deadline had passed, a dissenting judge took the majority to task for ignoring the commercial context of the statute:

Because the procedures set forth in § 232 are trade focused, and the relief provided is trade specific, the subject matter of § 232 flows directly Congress’s constitutional power over the Tariff. The majority decision, however, is untethered from the U.S. trade law context. As such, it answers the wrong question.[335]

Our analysis suggests three things with respect to future litigation over tariffs specifically, and foreign commerce in general. First, the government’s argument that any foreign commerce statute should be read in light of the President’s Article II powers is inconsistent with the original understanding of the Foreign Commerce Clause, the Tariff Clause, and their relationship to Article II. The Supreme Court’s holding in Learning Resources that the President lacks any inherent authority to impose tariffs (at least in peacetime)[336] should foreclose the type of separation of powers reasoning on which the Federal Circuit relied in rejecting challenges to the Section 232 tariffs.

This rejection of the relevance of Article II to the interpretation of tariff delegations is especially important given the government’s litigation strategy. In arguing V.O.S. Selections (which was later consolidated with Learning Resources) before the Federal Circuit, the government referred to IEEPA as a “supplement” to the President’s “Article II powers over foreign affairs” and argued that the Court ought to uphold the President’s actions because he had invoked his foreign relations power.[337] According to the government, “Presidents have exercised that authority across many administrations to impose tariffs that in their judgment will protect national security, foster economic prosperity, and facilitate negotiations with foreign counterparts.”[338] As we have shown in prior research, the government has consistently sought to conflate Article I and Article II powers to broaden the reading of foreign commerce statutes.[339] Despite this well-established position, the Court’s opinion in Learning Resources repeatedly emphasizes that the government disclaimed any reliance on Article II at oral argument before the Supreme Court.[340] In evaluating future foreign commerce delegations, it will be important for lower courts to hold the government to its concession in Learning Resources, now reflected in the Court’s holding, that the President lacks Article II powers over tariffs.

Second and relatedly, prior to Learning Resources, several Justices on the current Court had suggested that ordinary rules on statutory construction that instruct a court to read statutes narrowly, such as the major questions doctrine, may not apply “when a congressional statute confers wide discretion to the executive . . . if ‘the discretion is to be exercised over matters already within the scope of executive power.’”[341] Whatever the relevance of that type of foreign affairs exceptionalism in matters concerning shared constitutional powers (such as war powers), our analysis reveals that it has no place where foreign commerce statutes are concerned. The Supreme Court adopted this view in Learning Resources, holding that there is no foreign affairs exception applicable to the interpretation, including the application of the major questions doctrine, to tariff delegations because the President and Congress do not enjoy concurrent constitutional authority to impose tariffs during peacetime.[342] Indeed, the Supreme Court has historically treated foreign commerce statutes in the same manner as domestic economic statutes, including announcing the canonical “intelligible principle” test for ordinary nondelegation cases in a dispute over the 1922 Tariff Act.[343] And while these cases involve only tariffs, the history we have discussed suggests no reason to differentiate between tariffs and other foreign commerce issues that are also committed exclusively to Congress.

In short, courts should be attentive to whether the President actually does have any shared constitutional power over the constitutional matter in dispute. This Article has shown that where foreign commerce is involved, he does not. The interpretation of the scope of the President’s authority under foreign commerce statutes should thus proceed in the same manner as the interpretation of domestic commerce statutes. The fact that the Supreme Court has so held in Learning Resources will play an important role in future foreign commerce litigation, including the pending cases challenging the Section 122 tariffs imposed in February 2026, and may spur plaintiffs to think about renewing challenges to Federal Circuit decisions that rested on the erroneous view that the President’s Article II powers are relevant to the scope of tariff delegations.

Third, Learning Resources explicitly reserved the question of whether the President might be able to impose tariffs during wartime. This reservation also suggests that the extent to which the President can regulate foreign commerce more broadly during wartime did not fall within the scope of the Court’s decision. President Trump’s removal of the President of Venezuela through military action and his war against Iran, neither authorized by Congress and both with significant commercial implications, mean that the scope of the President’s inherent authority to regulate foreign commerce during militarized conflicts may well become the subject of litigation. But our analysis shows that peacetime versus wartime does not make a difference to whether the President enjoys inherent authority. In the early years of the Republic, Presidents still regularly sought congressional authorization for embargoes, during armed conflicts, in trade with countries against whom Congress had not issued a declaration of war. In short, when the President’s foreign affairs powers do overlap with Congress’s foreign commerce powers, Congress prevails.

B. Trade Agreements

Tariffs are not the only means through which recent administrations have relied on their foreign affairs powers to supplement or enhance their statutory authorities. President Trump’s unprecedented negotiation of what he calls trade “deals” during both his terms offers yet another illustration of how Congress’s powers over tariffs and commerce have been minimized, even when they have not been fully erased.[344]

The setting of the stage for this presidential power shift began much earlier than the modern tariff statutes. The Senate’s rejection of numerous trade treaties in the mid-nineteenth century prompted a shift in the relationship between the branches in the twentieth century when it came to the negotiation, conclusion, and implementation of deals addressing foreign commerce.[345] The primary institutional outcome of this shift was to abandon the treaty as the vehicle through which the country would bargain on trade. Instead, Congress delegated authority to the President to enter into trade agreements, which it would then approve and implement through legislation—at least so it began.[346] This Section explains how Presidents in several administrations have overstepped that arrangement and have mistakenly relied on an absent constitutional foreign affairs power to do so.

The Reciprocal Trade Agreements Act of 1934 was the first time that Congress granted the President substantial discretion to enter into international agreements regulating foreign commerce.[347] Most familiar to the trade policy world, though, is the detailed statutory scheme known as Trade Promotion Authority (TPA), first instituted in 1974.[348] Under this scheme, the executive can seek expedited legislative approval of a trade agreement so long as the agreement covers congressionally approved objectives, among other legislative requirements. The 2015 TPA legislation, for example, covers twenty-one issue areas in the negotiating objectives ranging from “trade in goods” to “anti-corruption.”[349] The executive, led by the USTR, may then negotiate an agreement that it brings back to Congress for approval and implementation.[350] For several decades under TPA, congressional consent to trade agreements has come from both houses of Congress via legislation, rather than through Senate advice and consent.

Scholars debated the constitutionality of the congressional-executive trade agreement in the 1990s.[351] The issue at that time was intra-congressional—did trade agreement approval require two-thirds of the Senate, or both houses of Congress? Despite these concerns, modern practice takes this debate as settled. All but one of the sixteen major free trade agreements into which the United States has entered have been implemented via the TPA process. In addition to TPA, certain additional statutes give the President latitude to make agreements to resolve investigations conducted pursuant to specific statutory rules.[352]

Today, the issue has shifted from whether the full Congress must approve trade agreements to whether and when Congress’s authorization or consent is required at all. In recent years, and indeed in recent months with considerable expediency, Presidents and their designees have entered into hundreds of trade-related agreements not authorized, approved by, or subsequently implemented by Congress in any way.[353]

The issue of the executive’s legal basis for trade agreements arose during the first Trump administration with respect to two deals negotiated between the United States and Japan.[354] Members of Congress also raised concerns about the lack of congressional approval with the Biden administration as the administration negotiated the Indo-Pacific Economic Framework (IPEF) and similar agreements.[355] Unusually, given current practice, Congress approved and implemented one such agreement with Taiwan via legislation. This legislation also required congressional consultation and authorization before bringing any future agreements into force.[356] In signing the Taiwan bill, President Biden indicated that he felt that congressionally imposed restrictions on his ability to negotiate trade agreements “would impermissibly infringe upon [the President’s] constitutional authority to negotiate with a foreign partner.”[357]

The second Trump administration has continued this path of entering into agreements that unquestionably regulate commerce with foreign nations but lack any statutory basis. In April 2025, President Trump demanded that countries commence negotiations with the United States in the hopes of avoiding higher tariff rates for their exporters. The President claimed that his administration would negotiate “90 deals in 90 days.”[358] According to the administration, such trade-related agreements would be necessary to eliminate the national emergency concerning the U.S. trade deficit by extracting tariff and nontariff concessions from these countries.[359] In June 2025, the United States and the United Kingdom reported the reciprocal implementation of certain trade-related guarantees.[360] Shortly thereafter, the Trump administration said that the United States and China agreed to a de-escalation of tariffs and other retaliatory measures,[361] and the United States and China seem poised to implement a more comprehensive resolution of their trade concerns.[362] As of the time of writing, the administration has also announced that is has reached agreements with about a dozen trading partners.[363]

When challenged, Presidents have defended these unlawful agreements with a handful of unsatisfactory arguments. First, the executive branch has pointed to tenuous claims of delegated authority. The USTR, for instance, argues that the agency’s organic statute—which provides that, inter alia, the Trade Representative shall “have lead responsibility for the conduct of, and shall be the chief representative of the United States for, international trade negotiations”[364]—constitutes authorization for actually entering into trade agreements. Although framed as a statutory argument, this claim rests on an implicit constitutional argument that only the most minimal congressional blessing is necessary to authorize executive branch action given its constitutional role in foreign affairs.

Second, drawing from the historical gloss approach, some have maintained that Congress has acquiesced to this executive branch practice.[365] These proponents note that Congress is not on record as having objected; ergo, it has consented.[366] Congress does not know about many of these agreements because the executive branch does not report them, so it is unclear that congressional acquiescence can be inferred even under a historical gloss approach.[367] More to the point here, these historical arguments rest on practices over the course of the last century or even just the last several decades. However, those practices are inconsistent with the original understanding of the Constitution and the historical practices from the Founding through the early twentieth century.[368] The government’s arguments thus seek to use recent historical practices to vary the original understanding and well-established practice of congressional control over trade agreements.[369]

A third response is that Congress must approve only those agreements that change U.S. law.[370] Under this view, Congress does not need to consent to agreements that do not require a change to federal statutes. U.S. Trade Representative Jamieson Greer has said that these 2025 agreements would not be provided to Congress for approval and implementation; rather, they would be implemented as executive agreements. Ambassador Greer justified this position by asserting that the agreements would include few, if any, commitments by the United States.[371] This is a variation on the acquiescence argument: Congress typically has not approved agreements that do not purport to change U.S. law, so why should it start now? For example, the Obama administration took the position that it did not need congressional consent for some international agreements, including those that covered areas outside of trade, as long as it had the authority under general statutory delegations to implement U.S. commitments.[372]

But if Congress has the power “to regulate commerce with foreign nations” and to “make all Laws which shall be necessary and proper for carrying into Execution” that power,[373] then surely Congress has the right to choose the instrument through which it regulates foreign commerce, including by insisting on its consent to the creation of international commitments related to foreign commerce. Congress made this point as early as 1974 in its passage of the TPA. Regardless of the form of agreement and whether the agreement requires changes to U.S. law, the executive branch does not have independent power to enter into a trade-related agreement without that authorization. Moreover, the deals negotiated by the Trump administration do appear to involve changes to U.S. law that require congressional authorization. Specifically, they contemplate that the United States will raise duties on partner countries permanently, a change that conflicts with trade agreements implemented by statute.[374]

These tenuous claims of broader delegated authority also extend to the modification and withdrawal from trade agreements. Given that it is clear, as a constitutional matter, that the President may not repeal legislation, it ought to be clear that the President lacks the authority to withdraw from trade agreements implemented through legislation as well, at least to the extent that the legislation presumes the agreement’s continuation in force or even explicitly restricts withdrawal.[375] Nevertheless, a 2001 study by the Congressional Research Service concluded that neither policymakers nor scholars have “seriously challenged” that Presidents may unilaterally withdraw from such agreements.[376] During his first term, President Trump claimed the authority to withdraw from the NAFTA.[377] Prominent commentators agreed with Trump, arguing that “the suggestion that the commerce authority is exclusive in a way that would distinguish congressional-executive agreements from Article II treaties is unpersuasive.”[378] The OLC has likewise concluded that “there is no good reason to believe that the Constitution preserves any greater role for Congress in the termination of a congressional-executive agreement on international trade than on any other subject matter.”[379] The OLC based this determination on the principle that the President is the “constitutional actor with ‘exclusive prerogatives in conducting the Nation’s diplomatic relations’” and as “the agent of the United States in conducting diplomacy, coupled with his constitutional responsibility to execute the laws.”[380] The OLC invoked Youngstown, and concluded that withdrawing from an agreement related to foreign commerce is, therefore, “supported by both the President’s own independent foreign-affairs authority and Congress’s approval [of the agreement]—putting the President’s authority at its constitutional zenith.”[381]

But as we have noted, Congress regularly maintained and exercised its authority to control modification and withdrawal from trade agreements during the nineteenth century.[382] Until the twentieth century, there was no consistent congressional practice that supported a presidential power to withdraw, at least when such withdrawal implicated congressional legislation in the foreign commerce context. In other words, whatever may be true for noncommercial agreements, the original understanding and historical practice, once we focus on the correct timeframe, does not support a unilateral presidential power of withdrawal from trade agreements.

Conclusion

Since returning to office, President Trump has invoked an executive foreign affairs power to implement an array of cross-border commercial moves, among them the imposition of duties and related restrictions as well as the formulation of so-called trade deals. Our examination of those exemplary policy tools, and like tools maintained by President Biden, has laid plain the problems with the recent reliance on a vision of Article II authority wrongly superimposed over Article I. But these are not the only spaces that exemplify this approach. Our analysis has lessons for other areas of foreign commercial lawmaking, and beyond, that we touch on here by way of conclusion: nonenforcement of foreign commercial restrictions; the administrative law foreign affairs exception; and international regulatory cooperation.

While President Trump has been conspicuous in asserting expansive executive authority in foreign commercial regulation, his administration has also invoked the President’s Article II powers as a basis for refusing to enforce foreign commercial statutes. Most notably, the President has declined to pursue enforcement actions under the Foreign Corrupt Practices Act and failed to implement statutory mandates concerning the prohibition of TikTok.[383] These statutes, duly enacted pursuant to Congress’s foreign commerce power, articulate a regulatory framework that the executive is bound to administer. The refusal to enforce such provisions reflects not mere discretion, but a broader challenge to the constitutional architecture of separated powers: Congress legislates and the executive executes. When the President declines to enforce statutes designed to govern foreign commerce, he not only disregards legislative command but also unsettles the institutional balance envisioned by the Framers.

In yet another example of the administration running roughshod over the separation of trade law powers, in March 2025, Secretary of State Marco Rubio published a notice in the Federal Register in which he purported to determine that “all efforts, conducted by any agency of the federal government, to control the status, entry, and exit of people, and the transfer of goods, services, data, technology, and other items across the borders of the United States, constitute a foreign affairs function of the United States under the Administrative Procedure Act [(APA)], 5 U.S.C. 553, 554.”[384] This notice reflects still another move by the administration to exert an extraordinarily broad understanding of statutory terms in favor of a general foreign affairs “function” that attempts to swallow the foreign commerce power, not to mention immigration issues. Secretary Rubio characterizes his own “primary foreign affairs duty: the duty to protect the people of the United States from any threats originating from foreign actors or from foreign soil.”[385] But he claims that “the scope of a foreign affairs function of the United States is much broader.”[386] Apart from its inconsistency with judicial interpretations of the foreign affairs exception in the APA, as a practical matter, excepting all regulatory activity that governs the movement of goods and services across borders would give the executive a free hand in areas in which Congress clearly did not intend to eliminate the deliberative processes required by the APA.

Finally, a substantial portion of modern executive branch activity operates within the domain of transnational regulatory cooperation. Nearly every federal agency now engages in regulatory initiatives with cross-border dimensions encouraged by prior administrations and expressly authorized by Congress.[387] The suggestion that international engagement exempts agency action from standard mechanisms of oversight and democratic accountability would have sweeping implications across the regulatory landscape from agricultural imports to advanced manufacturing to digital infrastructure. Again, the assertion that Article II alone empowers the President to displace congressional instruction in this sphere represents a sharp departure from constitutional design. The Founders placed the foreign commerce power squarely in congressional hands, with the expectation that executive implementation would remain subject to law.

In defending both its actions and omissions, the executive branch has not limited itself to conventional arguments of statutory interpretation. It has advanced sweeping claims rooted in an asserted constitutional authority over foreign commerce—claims that draw on a capacious conception of the foreign affairs power and appeal to a pattern of twentieth-century executive practice. Such a construction, if further endorsed by the courts, would mark a profound and unwarranted expansion of presidential power. These arguments find no foundation in the constitutional text or in the understandings of the Founding generation. The Constitution contemplates executive authority over foreign commerce only as delegated by statute; it does not permit the executive to augment those statutory delegations by invoking a generalized foreign affairs power. As novel disputes arise over presidential control of cross-border trade, this Article has sought to chart a corrective path—one that reaffirms the primacy of congressional authority and restores the constitutional balance.

Copyright © 2026 Kathleen Claussen* and Timothy Meyer**

*  Anne Fleming Research Professor & Professor of Law, Georgetown University Law Center.

**  Richard Allen/Cravath Distinguished Professor in International Business Law, Duke University School of Law. Thanks to the research librarian team at Georgetown Law for their contributions, especially David Isom. Thanks also to Lea Frenkel, Daniel Kim, and Jacob Koelsch for excellent research assistance.

       [1]. See, e.g., Eli Stokols, Doug Palmer, Myah Ward & Phelim Kine, Trump Got What He Wanted with Colombia. But His Tactics Could Come Back to Bite Him, Politico (Jan. 28, 2025), https://www.politico.com/news/2025/01/28/trump-colombia-trade-tariffs-007093 [https://perma.cc/RNU2-SCCJ] (discussing how a political skirmish about Colombian migrants to the United States prompted President Trump to threaten a 25 percent tariff on all Colombian goods).

       [2]. These totals reflect the actions as of August 1, 2025, when primary work on this Article was completed. Presidential Actions, White House, https://www.whitehouse.gov/presidential-actions/ [https://perma.cc/X923-QR44]; see also Chad P. Bown, Trump’s Trade War Timeline 2.0: An Up-to-Date Guide, PIIE: RealTime Econ. (Dec. 31, 2025), https://www.piie.com/blogs/realtime-economics/2025/trumps-trade-war-timeline-20-date-guide [https://perma.cc/Z3Y7-2HN5]. As one commentator has noted, “[t]his certainly gives the appearance that the president has sweeping powers over the setting of tariff rates.” Alan Wm. Wolff, Trump Tariffs and the Courts: Round 2, PIIE: RealTime Econ. (July 31, 2025), https://www.piie.com/blogs/realtime-economics/2025/trump-tariffs-and-courts-round-2 [https://perma.cc/Y76C-S4NN].

       [3]. See, e.g., Opening Brief for Appellants at 31, V.O.S. Selections, Inc. v. Trump, 149 F.4th 1312 (Fed. Cir. 2025) (Nos. 2025-1812, -1813) (“IEEPA’s text, history, and precedent confirm that it empowers the President to impose tariffs to address declared emergencies.”), aff’d sub nom., Learning Resources, Inc. v. Trump, 146 S. Ct. 628 (2026).

       [4]. We will necessarily address how the constitutional and statutory bounds intersect in Part V. Our prior work has considered the rise of the statutory authorities. See, e.g., Kathleen Claussen & Timothy Meyer, Economic Security and the Separation of Powers, 172 U. Pa. L. Rev. 1955, 1964–68 (2024). Historians have also charted that rise and how it has affected trade policy. See, e.g., Douglas A. Irwin, Clashing over Commerce: A History of U.S. Trade Policy (2017). But neither we nor they have investigated the Framers’ intent with respect to presidential authority.

       [5]. U.S. Const. art. I, § 8.

       [6]. See infra Part V.

       [7]. See infra Part V.

       [8]. See infra Part IV.

       [9]. See infra Part III.

     [10]. Opening Brief for Appellants, supra note 3, at 1 (emphasis added). The Supreme Court ultimately affirmed the Federal Circuit’s holding that IEEPA does not authorize the President to impose tariffs. Learning Resources, Inc. v. Trump, 146 S. Ct. 628 (2026).

     [11]. See Opening Brief for the Respondents in No. 24-1287 and the Petitioners in No. 25-250 at 22, Learning Resources, 146 S. Ct. 628 (Nos. 24-1287 & 25-250) (arguing that the nondelegation doctrine “has little or no force in the foreign-affairs context, where the President enjoys inherent Article II authority and Congress ‘must of necessity [sic] paint with’ a ‘broader’ brush.” (quoting Zemel v. Rusk, 381 U.S. 1, 17 (1965))). This framing is more subtle than that used before the Court of Appeals by suggesting that there is no problem with reading delegations broadly rather than framing statutes as mere supplements to constitutional authority. Ultimately, though, it has the same aim: to use the President’s purported constitutional power over foreign affairs to persuade courts to read statutory delegations broadly.

     [12]. See Kathleen Claussen, Trade Administration, 107 Va. L. Rev. 845, 857–62 (2021); Timothy Meyer & Ganesh Sitaraman, Trade and the Separation of Powers, 107 Calif. L. Rev. 583, 598 (2019) (“Congress believed there were policy benefits in delegating trade authority and insulating tariff rate-setting from congressional logrolling.”).

     [13]. See, e.g., Trade Act of 1974, 19 U.S.C. § 2411 (empowering the Trade Representative to take actions against foreign countries that restrict U.S. commerce); Trade Expansion Act of 1962, 19 U.S.C. § 1862(c)(1)(A)(ii) (empowering the President to “adjust the imports of the article and its derivatives so that such imports will not threaten to impair the national security”).

     [14]. See infra Part II.

     [15]. For a very short list of qualifying contributions, see, for example, Randy E. Barnett, The Original Meaning of the Commerce Clause, 68 U. Chi. L. Rev. 101 (2001); Richard A. Epstein, Constitutional Faith and the Commerce Clause, 71 Notre Dame L. Rev. 167 (1996); Mark Tushnet, Rethinking the Dormant Commerce Clause, 1979 Wis. L. Rev. 125. This topic is also discussed at length by the various essays featured in the Arkansas Law Review’s volume 55 symposium issue. See, e.g., Mark R. Killenbeck, Introduction: The Hand Maid of Liberty?, 55 Ark. L. Rev. 711 (2003).

     [16]. See, e.g., Curtis A. Bradley, Historical Gloss and Foreign Affairs: Constitutional Authority in Practice (2024); Harold Hongju Koh, The National Security Constitution in the 21st Century 21 (2024); Curtis Bradley & Jack Goldsmith, Foreign Affairs, Nondelegation, and the Major Questions Doctrine, 172 U. Pa. L. Rev. 1743, 1770 (2024); Saikrishna B. Prakash & Michael D. Ramsey, The Executive Power over Foreign Affairs, 111 Yale L.J. 231, 243 (2001).

     [17]. Exceptionally, see Jide Nzelibe, Presidential Constraint of Agencies in Foreign Affairs, 77 Admin. L. Rev. 371 (2025). In a related vein, Nicholas Parrillo has argued that Congress’s sweeping delegation of the power to impose embargoes in a 1794 statute “undermines the idea that there existed a foreign-affairs exception [to the nondelegation doctrine] to cover it.” Nicholas R. Parrillo, Foreign Affairs, Nondelegation, and Original Meaning: Congress’s Delegation of Power to Lay Embargoes in 1794, 172 U. PA. L. Rev. 1803, 1809 (2024). His analysis rests in large part on the importance of foreign commerce to the exercise of liberties protected by the nondelegation doctrine. Id.

     [18]. See infra Part III.A.

     [19]. See infra Parts III.B & III.C.

     [20]. See infra Part III.D.

     [21]. See infra Part III.D.

     [22]. Learning Resources, Inc. v. Trump, 146 S. Ct. 628, 638 (2026).

     [23]. Id.

     [24]. Complaint, Oregon v. Trump, No. 1:26-cv-01472-3JP (Ct. Int’l Trade Mar. 5, 2026), https://ag.ny.gov/sites/default/files/court-filings/state-of-oregon-et-al-v-donald-j-trump-et-al-complaint-2026.pdf; Complaint, Burlap & Barrel, Inc. v. Trump, No. 1:26-cv-01606 (Ct. Int’l Trade Mar. 9, 2026), https://libertyjusticecenter.org/wp-content/uploads/011-Burlap-and-Barrel-v.-Trump-Mot-PI-and-SJ-2026.03.13.pdf [https://perma.cc/MH89-BTSW].

     [25]. Jason Horowtiz, Trump Threatens to End Trade with Spain, N.Y. Times (Mar. 3, 2026), https://www.nytimes.com/2026/03/03/world/middleeast/trump-spain-trade.html.

     [26]. Patricia Cohen, Oil Shocks Send Tremors Through World Economy: ‘This Really is the Big One’, N.Y. Times (Mar. 12, 2026), https://www.nytimes.com/2026/03/12/business/economy/iran-oil-shock-economy-global-impact.html [https://perma.cc/K7MW-7RT9].

     [27]. See, e.g., South Dakota v. Wayfair, Inc., 138 S. Ct. 2080 (2018); Barclays Bank PLC v. Franchise Tax Bd., 512 U.S. 298 (1994); Franchise Tax Bd. v. Alcan Aluminium Ltd., 493 U.S. 331 (1990); Wardair Canada, Inc. v. Fla. Dep’t of Revenue, 477 U.S. 1 (1986); United States v. 12 200-Ft. Reels of Super 8mm. Film, 413 U.S. 123 (1973); Bd. of Trs. of Univ. of Ill. v. United States, 289 U.S. 48, 59 (1933); Brolan v. United States, 236 U.S. 216, 218–19 (1915); The Abby Dodge, 223 U.S. 166, 176–77 (1912); Buttfield v. Stranahan, 192 U.S. 470, 493 (1904); Gibbons v. Ogden, 22 U.S. 1, 195 (1824).

     [28]. See infra Part I.B. There is also some work by originalist scholars unpacking the meaning of these words to which we will return in Part III. We do not consider that work in detail because scholars have addressed the meaning of the text with respect to the Interstate Commerce Clause and only incidentally the Foreign Commerce Clause, as we explain below.

     [29]. Claussen & Meyer, supra note 4, at 1970–74. Exceptionally, see also Alan Wm. Wolff, Evolution of the Executive-Legislative Relationship in the Trade Act of 1974, 19 SAIS Rev. 16 (1975).

     [30]. This remains true even in the decades that follow—a theme to which we will return. See Michael A. Zuckerman, The Offshoring of American Government, 94 Corn. L. Rev. 165, 180 (2008) (“Lower courts borrow the dormant Interstate Commerce Clause standard to adjudicate challenges to state regulation of foreign commerce because the Supreme Court’s dormant Foreign Commerce Clause jurisprudence is relatively undeveloped.”); see also Antilles Cement Corp. v. Acevedo Vilá, 408 F.3d 41, 46 (1st Cir. 2005) (“Although the language of dormant Commerce Clause jurisprudence most often concerns interstate commerce, essentially the same doctrine applies to international commerce.”).

     [31]. See Conrad J. Weiler, Jr., How “Commerce Among the Several States” Became “Interstate Commerce,” and Why it Matters, 34 Const. Comment. 329, 375–76 (2019) (“[T]he legal community had convinced itself within a few years of the introduction of ‘the interstate commerce power’ that it was the constitutional power.”).

     [32]. Gibbons v. Ogden, 22 U.S. 1, 195 (1824).

     [33]. Id. at 195, 199 (instructing that “when a State proceeds to regulate commerce with foreign nations, or among the several States, it is exercising the very power that is granted to Congress”).

     [34]. Louis M. Greeley, What Is the Test of a Regulation of Foreign or Interstate Commerce?, 1 Harv. L. Rev. 159, 159 (1887).

     [35]. Id.

     [36]. See, e.g., Scott Sullivan, The Future of the Foreign Commerce Clause, 83 Fordham L. Rev. 1955, 1969 (2015) (explaining that Congress’s power to regulate commerce “‘among the several states’ has been overlooked in favor of an all-encompassing ‘Commerce Clause’ doctrine”); Michael S. Knoll & Ruth Mason, The Dormant Foreign Commerce Clause After Wynne, 39 Va. Tax Rev. 357, 358 (2020) (surveying dormant Foreign Commerce Clause doctrine).

     [37]. See, e.g., Bd. of Trs. of Univ. of Ill. v. United States, 289 U.S. 48, 59 (1933); Brolan v. United States, 236 U.S. 216, 218–19 (1915); The Abby Dodge, 223 U.S. 166, 176–77 (1912); Buttfield v. Stranahan, 192 U.S. 470, 493 (1904).

     [38]. See, e.g., supra note 37 (citing cases).

     [39]. Bd. of Trs. of Univ. of Ill., 289 U.S. at 59.

     [40]. Id. at 56–57.

     [41]. See Anthony J. Colangelo, The Foreign Commerce Clause, 96 Va. L. Rev. 949, 953 (2010) (defining “the ‘inward-looking’ foreign commerce power” as “the power to regulate domestically under the Foreign Commerce Clause”); see also S.-Cent. Timber Dev., Inc. v. Wunnicke, 467 U.S. 82, 100 (1984) (“It is a well-accepted rule that state restrictions burdening foreign commerce are subjected to a more rigorous and searching scrutiny.”).

     [42]. In Japan Line, Ltd. v. Los Angeles, the Court invalidated a California tax on Japanese shipping containers partly on the ground that the federal government had less ability to ensure fair apportionment when dealing with foreign sovereigns. 441 U.S. 434, 442, 451 (1979) (“[T]he ‘home port doctrine,’ as a rule for taxation of moving equipment, has yielded to a rule of fair apportionment among the States.”). Interestingly, this is the first case to use the phrase “Foreign Commerce Clause.” Id. at 449 n.14.

     [43]. Michelin Tire Corp. v. Wages, 423 U.S. 276, 285 (1976) (emphasis added) (footnotes omitted). Michelin Tire’s first component mirrors the Japan Line, Ltd. “one voice” requirement. Further, in Japan Line, Ltd., the Court reiterated that foreign commerce is an area where the federal government must “speak with one voice,” limiting state interference. Japan Line, Ltd., 441 U.S. at 449; see also Barclays Bank PLC v. Franchise Tax Bd., 512 U.S. 298, 308 (1994) (applying the Japan Line, Ltd. three-part test, the third part of which is whether the measure impedes the federal government from speaking with one voice when regulating commercial regulations with foreign governments); Itel Containers Int’l Corp. v. Huddleston, 507 U.S. 60, 77 (1993); Kraft Gen. Foods, Inc. v. Iowa Dep’t of Revenue & Fin., 505 U.S. 71, 79 (1992) (“[A] State’s preference for domestic commerce over foreign commerce is inconsistent with the Commerce Clause even if the State’s own economy is not a direct beneficiary of the discrimination.”); Wardair Canada, Inc. v. Fla. Dep’t of Revenue, 477 U.S. 1, 8 (1986) (“[T]he concern in these Foreign Commerce Clause cases is not with an actual conflict between state and federal law, but rather with the policy of uniformity, embodied in the Commerce Clause, which presumptively prevails when the Federal Government has remained silent.”); Container Corp. of Am. v. Franchise Tax Bd., 463 U.S. 159, 170 (1983); Mobil Oil Corp. v. Comm’r of Taxes, 445 U.S. 425, 443–44 (1980); La. Pub. Serv. Comm’n v. Tx. & N.O.R. Co., 284 U.S. 125, 132 (1931); South Dakota v. Wayfair, Inc., 138 S. Ct. 2080, 2089 (2018).

     [44]. By the twenty-first century, the use of the “one voice” doctrine became a rallying cry for executive authority, rather than federal authority, and proliferated outside of foreign commerce as well. See, e.g., Zivotofsky v. Kerry, 576 U.S. 1, 14 (2015) (“Recognition is a topic on which the Nation must ‘speak . . . with one voice.’. . . That voice must be the President’s. Between the two political branches, only the Executive has the characteristic of unity at all times. And with unity comes the ability to exercise, to a greater degree, ‘[d]ecision, activity, secrecy, and dispatch.’” (first quoting Am. Ins. Ass’n v. Garamendi, 539 U. S. 396, 424 (2003), then quoting The Federalist No. 70 (Alexander Hamilton))). But see David Moore, Beyond One Voice, 98 Minn. L. Rev. 953 (2014) (citing evidence to the contrary).

     [45]. Accordingly, the laws discussed in this Section are perhaps most appropriately viewed as “foreign commerce” rather than “commerce with foreign nations,” but they have been treated through the lens of the Foreign Commerce Clause, so we likewise examine them here.

     [46]. Colangelo, supra note 41, at 953 (defining “the ‘outward-looking’ foreign commerce power” as “the power to regulate extraterritorially”).

     [47]. Id. at 969.

     [48]. Prosecutorial Remedies and Other Tools to end the Exploitation of Children Today (PROTECT) Act of 2003, 18 U.S.C. § 2423(b).

     [49]. Id.

     [50]. See, e.g., Maritime Drug Law Enforcement Act (MDLEA), 46 U.S.C. §§ 70501–70508. But see United States v. Davila-Mendoza, 972 F.3d 1264, 1277 (11th Cir. 2020) (invalidating MDLEA as applied).

     [51]. See, e.g., Abitron Austria GMBH v. Hetronic Int’l, Inc., 600 U.S. 412, 428 (2023) (holding that the Lanham Act sections prohibiting the unauthorized use in commerce of a protected trademark when that use is likely to cause confusion do not apply extraterritorially).

     [52]. See Veazie v. Moor, 55 U.S. 568, 573 (1852) (“Commerce with foreign nations, must signify commerce which in some sense is necessarily connected with these nations, transactions which either immediately, or at some stage of their progress, must be extraterritorial.”); Lord v. S.S. Co., 102 U.S. 541, 544 (1880) (quoting Veazie, 55 U.S. at 573).

     [53]. Baston v. United States, 137 S. Ct. 850 (mem.), 851 (2017) (Thomas, J., dissenting from denial of certiorari) (“The few decisions from this Court addressing the scope of the Clause have generally been confined to laws regulating conduct with a significant connection to the United States.”).

     [54]. Id. at 850; see also In re Sealed Case, 936 F.3d 582, 586 (D.C. Cir. 2019).

     [55]. See generally Meyer & Sitaraman, supra note 12; Kathleen Claussen, The Other Trade War, 103 Minn. L. Rev. Headnotes 1 (2018); Matthew Schaefer, Self-Executing International Agreements and Private Rights of Action: Revisiting the 4th Restatement of Foreign Relations Law in the Context of International Trade and Investment Agreements, 45 U. Pa. J. Int’l L. 743 (2024); Daniel K. Tarullo, Law and Politics in Twentieth Century Tariff History, 34 UCLA L. Rev. 285, 286 (1986); Kenneth M. Casebeer, The Power to Regulate “Commerce with Foreign Nations” in a Global Economy and the Future of American Democracy: An Essay, 56 U. Mia. L. Rev. 25, 26 (2001); Naomi Harlin Goodno, When the Commerce Clause Goes International: A Proposed Legal Framework for the Foreign Commerce Clause, 65 Fla. L. Rev. 1139, 1140 (2013).

     [56]. See, e.g., Anderson L. Cao, Limiting States’ Roles in Foreign Commerce: Teaching Old-World Dogs New-World Tricks, 23 Hous. J. Int’l L. 349 (2001); Daniel M. Price & John P. Hannah, The Constitutionality of United States State & Local Sanctions, 39 Harv. Int’l L.J. 443 (1998); Daniel M. Price, John P. Hannah & Marinn F. Carlson, Crosby v. NFTC and the Future of State and Local Sanctions, 32 Law & Pol’y Int’l Bus. 37 (2000); Howard N. Fenton, III, The Fallacy of Federalism in Foreign Affairs: State and Local Foreign Policy Trade Restrictions, 13 Nw. J. Int’l L. & Bus. 563 (1993).

     [57]. See, e.g., William J. Aceves, Lost Sovereignty? The Implications of the Uruguay Round Agreements, 19 Fordham Int’l L.J. 427 (1995); Matt Schaefer & Thomas Singer, Multilateral Trade Agreements and U.S. States, 26 J. World Trade 31 (1992); David A. Gantz, A Post-Uruguay Round Introduction to International Trade Law in the United States, 12 Ariz. J. Int’l & Compar. L. 1 (1995).

     [58]. See, e.g., Bruce Ackerman & David Golove, Is NAFTA Constitutional?, 108 Harv. L. Rev. 799, 919 (1995); Laurence H. Tribe, Taking Text and Structure Seriously: Reflections on Free Form Method in Constitutional Interpretation, 108 Harv. L. Rev. 1221 (1995).

     [59]. Many of these were symposia or student notes. See, e.g., Kathleen Claussen, Old Wine in New Bottles? The Trade Rule of Law, 44 Yale J. Int’l. L. Online 61, 64 (2019); Harold Hongju Koh, Trump Change: Unilateralism and the “Disruption Myth” in International Trade, 44 Yale J. Int’l. L. Online 96, 102 (2019).

     [60]. See Jacob Gladysz, Note, The National Security Exception in WTO Law: Emerging Jurisprudence and Future Direction, 52 Geo. J. Int’l L. 835, 842–49 (2021).

     [61]. See Claussen & Meyer, supra note 4, at 1960–64.

     [62]. See Colangelo, supra note 41, at 952 (“Other commonly used enumerated sources of extraterritorial legislation do contemplate a degree of foreign consent for their authorization, which in turn limits the subject matter of the resultant law.”).

     [63]. 343 U.S. 579, 642, 644 (1952) (Jackson, J., concurring) (“That military powers of the Commander-in-Chief were not to supersede representative government of internal affairs seems obvious from the Constitution and from elementary American history.”); id. at 610–11 (Frankfurter, J., concurring) (“[A] systematic, unbroken, executive practice, long pursued to the knowledge of the Congress and never before questioned, engaged in by Presidents who have also sworn to uphold the Constitution, making as it were such exercise of power part of the structure of our government, may be treated as a gloss on ‘executive Power’ vested in the President by § 1 of Art. II.”).

     [64]. This is most clear in the Roberts Court’s jurisprudence on the President’s appointment and removal power, in which the Court has struck down a range of congressional statutes creating removal protections for certain officials that have a long twentieth-century pedigree but are, in the Court’s view, inconsistent with the original meaning of the separation of powers. Similar issues have also come up in the Roberts Court’s rights jurisprudence, especially in regard to the Second Amendment. See infra notes 108–18 and accompanying text.

     [65]. Michael D. Ramsey, The Vesting Clauses and Foreign Affairs, 91 Geo. Wash. L. Rev. 1513, 1514 (2023).

     [66]. U.S. Const. art. I, § 8, cls. 1, 3, 10.

     [67]. Id. cl. 11.

     [68]. Id. cls. 12, 13.

     [69]. Id. cl. 14.

     [70]. Id. art. II, § 2, cl. 1.

     [71]. Id. cl. 2.

     [72]. Id. art. II, § 3, cl. 1.

     [73]. See Prakash & Ramsey, supra note 16, at 243.

     [74]. See Zivotofsky v. Kerry, 576 U.S. 1, 13, 22 (2015) (holding that the President has preclusive power to recognize foreign nations and their boundaries).

     [75]. Curtis A. Bradley, Jack Goldsmith & Oona A. Hathaway, The Rise of Nonbinding International Agreements: An Empirical, Comparative, and Normative Analysis, 90 U. Chi. L. Rev. 1281, 1285 (2023).

     [76]. See Rebecca Ingber, Congressional Administration of Foreign Affairs, 106 Va. L. Rev. 395, 397 (2020).

     [77]. Prakash & Ramsey, supra note 16, at 234, 254.

     [78]. See id. at 265. But see Julian Davis Mortenson, The Executive Power Clause, U. Pa. L. Rev. 1269, 1271 (2020) (“For the founders, ‘the executive power’ meant the power to execute the law. Nothing more. And nothing less.”).

     [79]. Prakash & Ramsey, supra note 16, at 262.

     [80]. Id.

     [81]. Koh, supra note 16, at 3.

     [82]. See Oona A. Hathaway, National Security Lawyering in the Post-War Era: Can Law Constrain Power?, 68 UCLA L. Rev. 2, 84 (2021) (“To effectively counterbalance the executive branch, Congress not only needs more lawyers who understand national security law matters and are cleared into classified programs, but most importantly, it also needs an institutional counterweight to the Department of Justice’s [Office of Legal Counsel].”).

     [83]. See id. at 89.

     [84]. See, e.g., War Powers Resolution, 50 U.S.C. §§ 1541–1548; United States-Taiwan Initiative on 21st Century Trade First Agreement Implementation Act, 19 U.S.C. § 2112 note.

     [85]. Meyer & Sitaraman, supra note 12, at 638.

     [86]. INS v. Chadha, 462 U.S. 919, 958–59 (1983) (“There is unmistakable expression of a determination that legislation by the national Congress be a step-by-step, deliberate and deliberative process.”).

     [87]. Harlan G. Cohen, Formalism and Distrust: Foreign Affairs Law in the Roberts Court, 83 Geo. Wash. L. Rev. 380, 384–85 (2015); Ganesh Sitaraman & Ingrid Wuerth, The Normalization of Foreign Relations Law, 128 Harv. L. Rev. 1897, 1919 (2015).

     [88]. But see Ramsey, supra note 65, at 1514 (“[M]uch of the apparent difficulty in foreign affairs constitutionalism can be mitigated by giving up thinking about foreign affairs as a meaningful constitutional category.”).

     [89]. 299 U.S. 304 (1936).

     [90]. Id. at 320.

     [91]. Id. at 312; Zivotofsky v. Kerry, 576 U.S. 1, 21 (2015) (“This description of the President’s exclusive power was not necessary to the holding of Curtiss-Wright—which, after all, dealt with congressionally authorized action, not a unilateral Presidential determination.”).

     [92]. Jean Galbraith, The Runaway Presidential Power Over Diplomacy, 108 Va. L. Rev. 81, 83 (2022).

     [93]. Id. at 86, 113 (quoting Prohibition on Spending for Engagement of the Office of Science and Technology Policy with China, 35 Op. O.L.C. 116, 121–22 (Sep. 19, 2011)).

     [94]. See generally id.

     [95]. There are, of course, exceptions. See, e.g., Ackerman & Golove, supra note 58, at 929 (“Only effective frameworks will discourage Presidents . . . from supposing that the Constitution grants the executive branch a vast prerogative for unilateral foreign adventurism.”); Prakash & Ramsey, supra note 16, at 346–54.

     [96]. See Youngstown Sheet & Tube Co. v. Sawyer, 343 U.S. 579, 610–11 (1952) (Frankfurter, J., concurring).

     [97]. Id. at 637 (Jackson, J., concurring).

     [98]. Id.

     [99]. Id.

   [100]. See Dames & Moore v. Regan, 453 U.S. 654, 686–88 (1981); Haig v. Agee, 453 U.S. 280, 291 (1981) (cautioning that in foreign affairs, “congressional silence is not to be equated with congressional disapproval”).

   [101]. Curtis A. Bradley & Trevor W. Morrison, Historical Gloss and the Separation of Powers, 126 Harv. L. Rev. 411, 414 (2012); Curtis A. Bradley & Neil S. Siegel, Historical Gloss, Constitutional Convention, and the Judicial Separation of Powers, 105 Geo. L.J. 255, 258 (2017); Bradley, supra note 16.

   [102]. Bradley & Morrison, supra note 101, at 417.

   [103]. Id. at 418–20.

   [104]. William Baude, Constitutional Liquidation, 71 Stan. L. Rev. 1, 4 (2019) (“Liquidation was a specific way of looking at post-Founding practice to settle constitutional disputes . . . .”).

   [105]. Curtis A. Bradley & Neil S. Siegel, Historical Gloss, Madisonian Liquidation, and the Originalism Debate, 106 Va. L. Rev. 1, 2, 41 (2020) (contrasting the narrow view of liquidation that settles the meaning of ambiguous terms based on Founding-era practices with a broader view that might allow “reliquidation” based on later practices).

   [106]. See infra Part IV; Claussen & Meyer, supra note 4 (discussing the evolution of executive dominance of foreign commerce in the twentieth century).

   [107]. See Restatement (Fourth) of the Foreign Rels. L. of the U.S. § 313(a) & cmt. C (A.L.I. 2019) (providing in the black letter text that “[a]ccording to established practice, the President has the authority to act on behalf of the United States” when withdrawing from agreements, but acknowledging that eighteenth- and nineteenth-century practice does not provide any consistent state practice supporting that view).

   [108]. See Galbraith, supra note 92, at 85–87.

   [109]. Leah M. Litman, Debunking Antinovelty, 66 Duke L.J. 1407, 1410–12 (2017) (explaining that “antinovelty rhetoric” views a practice’s novelty as “a strong indication that it [is] unconstitutional” and providing examples of its use by the Supreme Court).

   [110]. Id. at 1411–12.

   [111]. Id. at 1425.

   [112]. 561 U.S. 477, 510 (2010).

   [113]. Id. at 505 (Kavanaugh, J., dissenting) (quoting Free Enter. Fund v. Pub. Co. Acct. Oversight Bd., 537 F.3d 667, 699 (D.C. Cir. 2008)).

   [114]. See Seila Law LLC v. CFPB, 591 U.S. 197, 222 (2020) (holding that the CFPB’s single-director structure is unconstitutional and had “no foothold in history or tradition”); N.Y. State Rifle & Pistol Ass’n, Inc. v. Bruen, 597 U.S. 1, 33 n.8 (2022) (“[I]n light of the text of the Second Amendment, along with the Nation’s history of firearm regulation, we conclude below that a State may not prevent law-abiding citizens from publicly carrying handguns because they have not demonstrated a special need for self-defense.”); Jarkesy v. SEC, 34 F.4th 446, 464–65 (5th Cir. 2022).

   [115]. See Trump v. Slaughter, 146 S. Ct. 18 (2025) (granting certiorari before judgment to decide the constitutionality of for-cause removal protections).

   [116]. 145 S. Ct. 1415, 1415 (2025) (concluding that the Trump administration is likely, on appeal, to be able to show that for-cause removal protections for members of the National Labor Relations Board and Merit Systems Protections Board are unconstitutional); id. at 1419 (Kagan, J., joined by Sotomayor & Jackson, J.J., dissenting) (characterizing the majority’s holding as “allow[ing] the President to overrule Humphrey’s by fiat”).

   [117]. Wilcox, 145 S. Ct. at 1415 (“The Federal Reserve is a uniquely structured, quasi-private entity that follows in the distinct historical tradition of the First and Second Banks of the United States.”).

   [118]. 576 U.S. 1, 30 (2015). To be fair, the Court’s opinion in Zivotofsky does not rely on originalist analysis. Several members of the Court in the majority there have now been replaced by members that considerably more originalist in their outlook.

   [119]. Id. at 23–24 (internal citations and quotations omitted).

   [120]. Higher Education Relief Opportunities for Students Act of 2003, 20 U.S.C. § 1098bb(a)(1).

   [121]. Biden v. Nebraska, 600 U.S. 477, 485, 501 (2023).

   [122]. See discussion infra Part III.A.

   [123]. See, e.g., Prakash & Ramsey, supra note 16; see also Michael W. McConnell, The President Who Would Not Be King: Executive Power under the Constitution 235–62 (2020); Gary Lawson & Guy Seidman, The Jeffersonian Treaty Clause, 2006 U. Ill. L. Rev. 1, 41; Michael Stokes Paulsen, Youngstown Goes to War, 19 Const. Comment. 215, 237–38 (2002); John C. Yoo, War and the Constitutional Text, 69 U. Chi. L. Rev. 1639, 1677 (2002).

   [124]. Prakash & Ramsey, supra note 16, at 253–54; McConnell, supra note 123, at 11.

   [125]. See Mortenson, supra note 78, at 1271–72; Julian Davis Mortenson, Article II Vests the Executive Power, Not the Royal Prerogative, 119 Colum. L. Rev. 1169, 1172 n.8 (2019); David Gray Adler, The Framers and Executive Prerogative: A Constitutional and Historical Rebuke, 42 Presidential Stud. Q. 376 (2012); Jack N. Rakove, Taking the Prerogative out of the Presidency: An Originalist Perspective, 37 Presidential Stud. Q. 85, 85 (2007); Curtis A. Bradley & Martin S. Flaherty, Executive Power Essentialism and Foreign Affairs, 102 Mich. L. Rev. 545, 568–70 (2004).

   [126]. Alfred McCoy, As the “Tariff Man,” Trump is creating the Second Gilded Age, Salon (July 30, 2025), https://www.salon.com/2025/07/30/tariff-man-is-bringing-back-the-gilded-age/ [https://perma.cc/2TMP-DQJU]. In his Learning Resources dissent, Justice Thomas takes the opposite view, arguing that the power to impose tariffs was “a prerogative right of the King.” Learning Resources v. Trump, 146 S. Ct. 628, 685 (Thomas, J., dissenting) (internal quotation marks omitted). He makes this argument to show that the power to impose tariffs was not a “legislative” power, even though the Constitution grants it to Congress, and thus that Congress could delegate the power without implicating nondelegation concerns. Id. at 684. In response, Justice Gorsuch noted that while “[t]ariffs may have been among the King’s prerogative powers during the reign of Edward I,” that was no longer true as early as 1400, and, as described in the text that follows, by the 1688 Glorious Revolution Parliament had “secured supremacy in fiscal matters.” Learning Resources, 146 S. Ct. at 670 (Gorsuch, J., concurring) (internal citation omitted).

   [127]. Learning Resources, 146 S. Ct. at 638 (“A tariff,’ after all, is a tax levied on imported goods and services.” (internal quotation marks and citation omitted)).

   [128]. G.L. Harriss, King, Parliament, and Public Finance in Medieval England to 1369, at 36 (1975) (discussing dates).

   [129]. Linda S. Popofsky, The Crisis Over Tonnage and Poundage in Parliament in 1629, 126 Past & Present 44, 49 (1990).

   [130]. Id. at 49–50.

   [131]. Conrad Russell, Parliaments and English Politics 1621–1629, at 225–38 (1979).

   [132]. 19 Proceedings of the Short Parliament of 1640, at 152–153 (Esther S. Cope & Willson H. Coates eds., 1977).

   [133]. Tonnage and Poundage Act 1640, 16 Car. 1 c. 8, § 6 (Eng.).

   [134]. Bill of Rights 1689, 1 W. & M. Sess. 2 c. 2 (Eng.) (“That levying money for or to the Use of the Crowne by pretense of Prerogative without Grant of Parlyament for longer time or in other manner then the same is or shall be granted is Illegall.”).

   [135]. Josh Chafetz, Congress’s Constitution: Legislative Authority and the Separation of Powers 47 (2017).

   [136]. Andrew Kent, Executive Power, the Royal Prerogative, and the Founders’ Presidency, 2 J. Am. Const. Hist. 403, 410 (2024).

   [137]. See 1713 Treaty of Utrecht, Apr. 11, 1713/Mar. 31, 1713, in 1 A Collection of All the Treaties of Peace, Alliance, and Commerce Between Great-Britain and Other Powers, from the Revolution in 1688, to the Present Time 287 (1772).

   [138]. John H. Jackson, Status of Treaties in Domestic Legal Systems: A Policy Analysis, 86 Am. J. Int’l L. 310, 319 (1992) (“The United Kingdom is generally considered the prime example of a dualist system.”).

   [139]. Id.; Arabella Lang, House of Commons Libr., Briefing Paper No. 5855: Parliament’s Role in Ratifying Treaties 5 (2017).

   [140]. Today, even the royal prerogative over treaties is somewhat diminished. In the twentieth century, the Ponsonby Rule emerged, under which the government would lay treaties before Parliament before ratifying them. The 2010 Constitutional Reform and Governance Act codified that obligation and granted Parliament the right to indefinitely block the ratification of treaties. Lang, supra note 139, at 10–13.

   [141]. Kevin Douglas Tufnell, ‘A Safe and Honourable Peace’: British Political Discourse, Politics and Policy Formation in the Making of the Treaty of Utrecht, 1708 to 1713, at 213 (Mar. 21, 2022) (Ph.D. thesis, University College London), https://discovery.ucl.ac.uk/id/eprint/10151123/ [https://perma.cc/2X2A-6GJ6].

   [142]. W.O. Henderson, The Anglo-French Commercial Treaty of 1786, 10 Econ. Hist. Rev. 104, 104 n.5 (1957).

   [143]. Id.; Doohwan Ahn, The Anglo-French Treaty of Commerce of 1713: Tory Trade Politics and the Question of Dutch Decline, 36 Hist. Eur. Ideas 167, 169 (2010).

   [144]. Ahn, supra note 143, at 168; Henderson, supra note 142, at 104–05.

   [145]. Henderson, supra note 142, at 104 n.5; Tufnell, supra note 141, at 232.

   [146]. Ahn, supra note 143, at 169 (“It was Bolingbroke’s firm belief that the Whig continental policy of containing France by fortifying the Dutch barrier was responsible for the continued Dutch dominance in world trade.”).

   [147]. Articles of Confederation of 1781, art. IX, para. 1.

   [148]. Id. art. VI, para. 3.

   [149]. Id. art. IX, para. 6.

   [150]. Albert Anthony Giesecke, American Commercial Legislation Before 1789, at 140 (1910) (“The states possessed sovereign powers in the matter of commercial legislation; moreover, they jealously guarded this power. Hence, in the Articles of Confederation, . . . the powers of Congress over commercial legislation were very restricted in scope.”).

   [151]. Irwin, supra note 4, at 55 (quoting Madison in 1787).

   [152]. 1 Joseph Story, Commentaries on the Constitution of the United States § 261 (Melville M. Bigelow ed., Little, Brown & Co., 5th ed. 1905) (1833).

   [153]. The Federalist No. 5, at 25 (John Jay) (Ian Shapiro ed., 2009).

   [154]. 2 The Records of the Federal Convention of 1787, at 625 (Max Farrand ed., 1911).

   [155]. Irwin, supra note 4, at 54; Giesecke, supra note 150, at 144 (“One commercial power Congress did have . . . it appointed commercial agents abroad and secured several treaties of commerce and friendship with foreign nations.”).

   [156]. Irwin, supra note 4, at 56–57.

   [157]. Robert J. Delahunty, Federalism Beyond the Water’s Edge: State Procurement Sanctions and Foreign Affairs, 37 Stan. J. Int’l L. 1, 18 (2001).

   [158]. Id. at 19.

   [159]. Id.

   [160]. St. George Tucker, View of the Constitution of the United States (1803), reprinted in View of the Constitution of the United States with Selected Writings 188 (Clyde N. Wilson ed., 1999); see also Delahunty, supra note 157, at 19.

   [161]. Irwin, supra note 4, at 56–57.

   [162]. Id. at 58; see also Giesecke, supra note 150, at 140 & n.69.

   [163]. Irwin, supra note 4, at 56–57.

   [164]. 3 The Records of the Federal Convention of 1787, supra note 154, at 521.

   [165]. Hugh Williamson, Speech at Edenton (Nov. 8, 1787), in The Documentary History of the Ratification of the Constitution Digital Edition (John P. Kaminski, Gaspare J. Saladino, Richard Leffler, Charles H. Schoenleber & Margaret A. Hogan eds., 2009), https://rotunda-upress-virginia-edu.libproxy.berkeley.edu/founders/RNCN-03-16-02-0071 [https://perma.cc/HV87-2DY5].

   [166]. The Address and Reasons of Dissent of the Minority of the Convention of the State of Pennsylvania to their Constituents (Dec. 12, 1787), in The Documentary History of the Ratification of the Constitution Digital Edition (John P. Kaminski, Gaspare J. Saladino, Richard Leffler, Charles H. Schoenleber & Margaret A. Hogan eds., 2009), https://rotunda-upress-virginia-edu.libproxy.berkeley.edu/founders/RNCN-02-02-02-0003-0003 [https://perma.cc/2HTA-BP9V].

   [167]. Tucker, supra note 160, at 188.

   [168]. 1 The Records of the Federal Convention of 1787, supra note 154, at 263.

   [169]. Id. at 133, 142.

   [170]. The Federalist No. 23, at 115 (Alexander Hamilton) (Ian Shapiro ed., 2009).

   [171]. Letter from William Bingham to Tench Coxe (June 12, 1788), in The Documentary History of the Ratification of the Constitution Digital Edition (John P. Kaminski, Gaspare J. Saladino, Richard Leffler, Charles H. Schoenleber & Margaret A. Hogan eds., 2009), https://rotunda-upress-virginia-edu.libproxy.berkeley.edu/founders/RNCN-02-20-02-0004-0199 [https://perma.cc/HQJ6-HDR3].

   [172]. 3 Story, supra note 152, at § 1096.

   [173]. See Jack N. Rakove, Original Meanings: Politics and Ideas in the Making of the Constitution 26 (1997) (citing letter to Jefferson).

   [174]. Id.

   [175]. 2 The Records of the Federal Convention of 1787, supra note 154, at 449.

   [176]. Michael J. Klarman, The Framers’ Coup: The Making of the United States Constitution 277 (2016).

   [177]. 2 The Records of the Federal Convention of 1787, supra note 154, at 625; 3 The Records of the Federal Convention of 1787, supra note 154, at 333; see also, Klarman, supra note 176, at 283.

   [178]. U.S. Const. art. II, § 2, cl. 2.

   [179]. Delahunty, supra note 157, at 25.

   [180]. 2 The Records of the Federal Convention of 1787, supra note 154, at 135.

   [181]. 1 The Records of the Federal Convention of 1787, supra note 154, at 242–43.

   [182]. Id. at 243.

   [183]. 2 The Records of the Federal Convention of 1787, supra note 154, at 308, 493.

   [184]. Id.

   [185]. Id. The “and” was added in September after the clause about the Indian tribes had also been added.

   [186]. We will later return to treaties, so we will not take up the language of the Treaty Clause here. For a more thorough evaluation of the meaning of the constitutional phrase, see, e.g., Lawson & Seidman, supra note 123.

   [187]. Albert S. Abel, The Commerce Clause in the Constitutional Convention and in Contemporary Comment, 25 Minn. L. Rev. 432 (1941); Bd. Of Trs. Of Univ. of Ill. V. United States, 289 U.S. 48, 53 (1933).

   [188]. Barnett, supra note 15, at 114–16.

   [189]. Id.

   [190]. Id. at 124–25.

   [191]. Abel, supra note 187, at 448 (“The identification of commercial regulation with customs stands out plainly in Wilson’s exclamation, when the provision forbidding export duties by the federal government was being debated, ‘To deny this power is to take from the common govt. half the regulation of trade,’ the other half being, as the context shows, the corresponding control over imports. Between them, the implication plainly is, they constituted the whole of commercial regulation.”).

   [192]. Id.

   [193]. Irwin, supra note 4.

   [194]. An Old Whig VI (Nov. 24, 1787), in The Documentary History of the Ratification of the Constitution Digital Edition (John P. Kaminski, Gaspare J. Saladino, Richard Leffler, Charles H. Schoenleber & Margaret A. Hogan eds., 2009), https://rotunda-upress-virginia-edu.libproxy.berkeley.edu/founders/RNCN-03-14-02-0056 [https://perma.cc/SC7H-49GJ].

   [195]. 2 Story, supra note 152, at § 259; Irwin, supra note 4.

   [196]. 3 The Records of the Federal Convention of 1787, supra note 154, at 520–21.

   [197]. To use Abel’s term. See Abel, supra note 187, at 465.

   [198]. Claussen, supra note 12, at 857.

   [199]. In July 1789, the Second Act of Congress established a system of tariffs on imported “goods and merchandises” while the Third Act established tariffs on the tonnage of ships. Already in Congress’s earliest days there was a debate about the proper objectives of a tariff, but most salient was the need for revenue. Act of Jan. 29, 1795, ch. 20, 1 Stat. 414; Act of June 13, 1798, ch. 53, § 5, 1 Stat. 565, 566; Act of July 31, 1789, ch. 5, 1 Stat. 29; Act of Aug. 7, 1789, ch. 9, 1 Stat. 53; Act of Aug. 4, 1790, ch. 35, §§ 62–64, 1 Stat. 145, 175; Act of May 27, 1796, ch. 31, 1 Stat. 474; Act of July 16, 1798, ch. 77, 1 Stat. 605; Act of Sep. 2, 1789, ch. 12, 1 Stat. 65.

   [200]. John F. Coyle, The Treaty of Friendship, Commerce and Navigation in the Modern Era, 51 Colum. J. Transnat’l L. 302, 307 (2013).

   [201]. Id. at 314–15.

   [202]. Id.

   [203]. Craig VanGrasstek, Trade and American Leadership 89 (2019).

   [204]. The President could also, of course, use his veto power to shape the legislation coming from Congress.

   [205]. 5 Annals of Cong. 596 (1796).

   [206]. Id. at 603.

   [207]. 29 Annals of Cong. 1019 (1816); Thomas Sergeant, Constitutional Law 401 (1822); 3 Story, supra note 152, at § 1841.

   [208]. George Washington, Proclamation of Neutrality (Apr. 22, 1793), reprinted by Nat’l Archives: Founders Online, https://founders.archives.gov/documents/Washington/05-12-02-0371 [https://perma.cc/S2NN-YPBR].

   [209]. Jeff Broadwater, Madison, Hamilton, and the Neutrality Proclamation of 1793: debating presidential power and foreign affairs, 83 Historian 171, 175 (2021).

   [210]. Id.

   [211]. Alexander Hamilton, Pacificus No. 1 (June 29, 1793), reprinted by Nat’l Archives: Founders Online, https://founders.archives.gov/documents/Hamilton/01-15-02-0038 [https://perma.cc/3ULS-AMDT].

   [212]. Id. (emphasis removed) (alteration in original).

   [213]. Id.

   [214]. Id.

   [215]. James Madison, “Helvidius” Number 1 (Aug. 24, 1793), reprinted by Nat’l Archives: Founders Online, https://founders.archives.gov/documents/Madison/01-15-02-0056 [https://perma.cc/HMH6-FJ8R]. Indeed, Madison pointed out that Hamilton himself had argued that treaty-making is legislative in nature in the Federalist. Id.

   [216]. James Madison, “Helvidius” Number 2 (Aug. 31, 1793), reprinted by Nat’l Archives: Founders Online, https://founders.archives.gov/documents/Madison/01-15-02-0061 [https://perma.cc/P6TA-2GUG].

   [217]. Alexander Hamilton, Americanus No. 1 (Jan. 31, 1794), reprinted by Nat’l Archives: Founders Online, https://founders.archives.gov/documents/Hamilton/01-15-02-0510 [https://perma.cc/3HD2-ZBLY].

   [218]. Madison, supra note 215.

   [219]. Prakash & Ramsey, supra note 16, at 327.

   [220]. Id. at 327 n.415.

   [221]. Id. at 335.

   [222]. Id. at 341–43; Parrillo, supra note 17, at 1808, 1816.

   [223]. Act of June 4, 1794, ch. 41, 1 Stat. 372 (1794).

   [224]. Id.

   [225]. Prakash & Ramsey, supra note 16, at 346.

   [226]. Similarly, in the context of the nondelegation doctrine, Parrillo writes that “setting aside the foreign aspect of commerce as somehow exceptional, leaving the nondelegation doctrine (and its putative protection of private liberty and property) strongly applicable only to the portion of commerce that happened domestically, would not have made sense in the founding era.” Parrillo, supra note 17, at 1822.

   [227]. See, e.g., id.; Koh, supra note 16, at 21; Bradley & Goldsmith, supra note 16, at 1770; David J. Barron & Martin S. Lederman, The Commander in Chief at the Lowest Ebb—A Constitutional History, 121 Harv. L. Rev. 941, 964 (2008).

   [228]. Prakash and Ramsey also note that

there was no discussion of the President imposing an embargo (or other regulation of commerce) during the Washington Administration; these matters were handled in Congress. In particular, Congress obviously thought the President lacked the ability to impose an embargo on his own authority, for in 1794 it delegated to the President the power to impose an embargo during the legislative recess “whenever, in his opinion, the public safety shall so require.” This further confirms the general understanding that foreign affairs powers conveyed to Congress by the Constitution were conveyed away from the President, even where these powers had previously been traditional executive powers.

Prakash & Ramsey, supra note 16, at 349 (quoting Act of June 4, 1794, ch. 41, 1 Stat. 372). We discuss embargoes at greater length below.

   [229]. 6 U.S. 170 (1804).

   [230]. Act of June 13, 1798, ch. 53, 1 Stat. 565; Act of Feb. 9, 1799, ch. 2, § 5, 1 Stat. 613, 615.

   [231]. Act of Feb. 9, 1799, ch. 2, § 5, 1 Stat. 613.

   [232]. Little, 6 U.S. at 178.

   [233]. Id. at 179.

   [234]. J. Gregory Sindak, The Quasi-War Cases—And Their Relevance to Whether “Letters of Marque and Reprisal” Constrain Presidential War Powers, 28 Harv. J.L. & Pub. Pol’y 465, 493 (2005).

   [235]. 3 Story, supra note 152, at § 1290.

   [236]. Embargo Act of 1807, ch. 5, 2 Stat. 451, 452.

   [237]. Id. at 452.

   [238]. Act of Jan. 9, 1808, ch. 8, 2 Stat. 453; Act of Mar. 12, 1808, ch. 33, 2 Stat. 473.

   [239]. Act of Jan. 9, 1808, ch. 8, 2 Stat. 453; Act of Mar. 12, 1808, ch. 33, 2 Stat. 473.

   [240]. United States v. The William, 28 F. Cas. 614, 620 (D. Mass. 1808) (No. 16,700).

   [241]. Id. at 621.

   [242]. However, the court also ruled that this power was restricted by the “treaty making power of the president and senate.” Id.

   [243]. Embargo Act of 1807, ch. 5, 2 Stat. 451, 452 (applying the embargo to “all ships and vessels”).

   [244]. See Nonintercourse Act of 1809, ch. 24, 2 Stat. 528.

   [245]. Id. §§ 1, 19.

   [246]. Id. § 11.

   [247]. Id. §§ 12, 19.

   [248]. James Madison, Presidential Proclamation (Aug. 9, 1809), reprinted by Nat’l Archives: Founders Online, https://founders.archives.gov/documents/Madison/03-01-02-0353 [https://perma.cc/PX87-NE52].

   [249]. Nonintercourse Act of 1809, ch. 24, § 19, 2 Stat. 533.

   [250]. Letter from Albert Gallatin to the House of Representatives, No. 143: Embargo (May 25, 1809), in 1 American State Papers: Commerce and Navigation 778 (Walter Lowrie & Walter S. Franklin eds.,1834).

   [251]. Id.

   [252]. Id.

   [253]. Madison, supra note 248.

   [254]. Letter from James Madison to Congress, No. 182: Prohibition of Exports (July 20, 1813), in 1 American State Papers: Commerce and Navigation 968 (Walter Lowrie & Walter S. Franklin eds., 1834).

   [255]. Cargo of the Brig Aurora v. United States, 11 U.S. 382, 388 (1813).

   [256]. See Bradley & Morrison, supra note 101, at 422–23; see also Youngstown Sheet & Tube Co. v. Sawyer, 343 U.S. 579, 610–11 (1952) (Frankfurter, J., concurring) (discussing the importance of “systematic, unbroken, executive practice” in constitutional interpretation).

   [257]. See Charles Evans Hughes, The Supreme Court of the United States: Its Foundations, Methods, and Achievements, an Interpretation 138–42 (1928) (describing the limited, delegated role of the President in tariff matters); see also Mortenson, supra note 125, at 1231–36 (“[T]he executive power didn’t encompass other authorities . . . . [T]he subordinacy of ‘executive power’ was one of its constitutive features.”).

   [258]. Ackerman & Golove, supra note 58, at 820–29, 821 (“Early statutes imposed this duty on the President, typically requiring him to issue a proclamation giving each complying country a clean bill of health. We call these “proclamation statutes,” and they have been very common.”).

   [259]. Act of Feb. 9, 1799, ch. 2, § 4, 1 Stat. 613, 615.

   [260]. Act of Dec. 19, 1806, ch. 1, § 3, 2 Stat. 411, 411.

   [261]. Act of June 4, 1794, ch. 41, § 1, 1 Stat. 372, 372 (Congress authorized the President to lay an embargo on ships as necessary “whenever, in his opinion, the public safety shall so require”); Act of Feb. 9, 1799 § 4, ch. 2, § 4, 1 Stat. 613 (Congress again made it lawful for the President to draw back restrictions on trade that Congress enacted “if he shall deem it expedient and consistent with the interest of the United States” or “whenever, in his opinion, the interest of the United States shall require”); Act of Dec. 19, 1806, ch. 1, § 3, 2 Stat. 411; Act of Mar. 3, 1815, ch. 77, 3 Stat. 224; Act of Mar. 3, 1817, ch. 39, 3 Stat. 361; Act of Jan. 7, 1824, ch. 4, § 4, 4 Stat. 2, 3; Act of May 31, 1830, ch. 220, § 1, 4 Stat. 425, 425–26; Act of June 26, 1884, ch. 121, § 14, 23 Stat. 53, 57.

   [262]. The scope of this retail/wholesale distinction was tested in Field v. Clark, 143 U.S. 649 (1892), in detail, and even earlier in Cargo of the Brig Aurora v. United States, 11 U.S. 382 (1813) (finding that the legislature may make the revival of an act depend upon a future event and direct that event to be made known by proclamation).

   [263]. See Buttfield v. Stranahan, 192 U.S. 470 (1904); Weber v. Freed, 239 U.S. 325 (1915); see also Shalev Roisman, Presidential Factfinding, 72 Vand. L. Rev. 825 (2019) (discussing the President’s role as a factfinder in a variety of statutory schemes).

   [264]. See S. Rep. No. 73-871, at 1–2 (1934) (“The committee has inserted the words ‘as a fact’ following the words in subsection (a) ‘the President, whenever he finds’. This is to make clear that Congress under the proposed bill is establishing a policy and directing the Executive to act in accordance with the congressional policy only when he finds as a fact that existing duties or other import restrictions are unduly burdening and restricting the foreign trade of the United States. In the same provision, to the words ‘existing duties or other import restrictions’ the words ‘of the United States or any foreign country’ have been added to clarify the meaning.”).

   [265]. See, e.g., Irwin, supra note 4, at 185–89, 191; VanGrasstek, supra note 203, at 40.

   [266]. In President Taft’s inauguration in 1909, he called for Congress to give him still greater authority but also noted that any such action was a congressional prerogative. See William Howard Taft, Inaugural Address (Mar. 4, 1909), in 1 Presidential Addresses and State Papers of William Howard Taft 53, 55 (1910) (“It is imperatively necessary, therefore, that a tariff bill be drawn . . . and as promptly passed as due consideration will permit . . . . I venture this as a suggestion only, for the course to be taken by Congress, upon the call of the Executive, is wholly within its discretion.”); see also Warren G. Harding, Inaugural Address (Mar. 4, 1921), in Inaugural Addresses of the Presidents of the United States, S. Doc. No. 101–10, at 237, 243–44 (1989).

   [267]. See e.g., 2 James K. Polk, The Diary of James K. Polk: During His Presidency, 1845 to 1849, at 55 (1910); Zachary Taylor, Inaugural Address (Mar. 5, 1849), in Inaugural Addresses of the Presidents of the United States, S. Doc. No. 101–10, at 111 (1989). Exceptionally, perhaps, Treasury Secretary Robert Walker was active in promoting free trade on behalf of the Polk administration, but the President still relied on Congress to pass legislation with new tariff rates rather than intervene himself in setting tariff rates for products. John M. Dobson, Two Centuries of Tariffs: The Background and Emergence of the U.S. International Trade Commission 13 (1976).

   [268]. Irwin, supra note 4, at 214.

   [269]. Marc-William Palen, Foreign Relations in the Gilded Age: A British Free-Trade Conspiracy?, 37 Diplomatic Hist. 217, 229 (2013); Allen Nevins, Grover Cleveland: A Study in Courage (1932).

   [270]. Irwin, supra note 4, at 259–61.

   [271]. McKinley Tariff Act of 1890, ch. 1244, 26 Stat. 567.

   [272]. Id. § 3, at 612; see also Field v. Clark, 143 U.S. 649, 680 (1892) (noting that Congress lawfully authorized the President to suspend tariff exemptions when foreign nations imposed “unequal and unreasonable” trade conditions, as part of a reciprocal trade framework); Irwin, supra note 4, at 304–10; H.R. Rep. No. 73-1000 (1934); Francis B. Sayre, The Constitutionality of the Trade Agreements Act, 39 Colum. L. Rev. 751 (1939).

   [273]. Field, 143 U.S. at 694 (note also the dissent commenting that this act ought to be distinguished as it is clearly lawmaking).

   [274]. See, e.g., Dingley Tariff Act, ch. 11, 30 Stat. 151, 151 (1897); Fordney-McCumber Act of 1922, ch. 356, § 315, 42 Stat. 858, 941–43.

   [275]. Payne–Aldrich Tariff Act of 1909, ch. 6, § 2, 36 Stat. 11, 82.

   [276]. H.R. Rep. No. 73-1000, at 10 (1934).

   [277]. See, e.g., Proclamation: Tariff on British Products, 36 Stat. 2505, 2505–06 (1910); Proclamation: Tariff on Swiss Products, 36 Stat. 2507, 2507–08 (1910); Proclamation: Tariff on Turkish Products, 36 Stat. 2509, 2509–10 (1910) (with respect to the Ottoman Empire, Great Britain, and Swiss Confederation).

   [278]. Payne–Aldrich Tariff Act of 1909, ch. 6, § 2, 36 Stat. 11, 82–83; Rebecca Gomez Betancourt & Stephen Meardon, The Scientific Tariff: From Origins to the Travails of F.W. Taussig, 32 Euro. J. Hist. Econ. Thought 596, 607 (2025) (“The election of Democrat Woodrow Wilson and Democrats’ gain of unified control of Congress in 1912 spelled the demise of the Tariff Board.”).

   [279]. Trading with the Enemy Act, ch. 106, 40 Stat. 411 (1917).

   [280]. Id. §§ 2, 3 (defining an “enemy” as any resident “of any nation with which the United States is at war” and imposing sanctions on such enemies, subject to the discretion of the President to lift those sanctions).

   [281]. See U.S. Int’l Trade Comm’n, Investigation No. 332–335, The Economic Effects of Significant U.S. Import Restraints 65 (2009) (“Prior to the 1930 act, tariff changes were viewed as entirely the domain of Congress.”); see also Hal Shapiro & Lael Brainard, Trade Promotion Authority Formerly Known As Fast Track: Building Common Ground on Trade Demands More Than a Name Change, 35 Geo. Wash. Int’l L. Rev. 1, 6 (2003) (“Prior to the twentieth century U.S. regulation of foreign commerce was almost exclusively a congressional prerogative.”); Ian F. Fergusson, Cong. Rsch. Serv., RL33743, Trade Promotion Authority (TPA) and the Role of Congress in Trade Policy 2–3 (2015).

   [282]. See Giesecke, supra note 150, at 144.

   [283]. U.S. Tariff Comm’n, Reciprocity and Commercial Treaties 10 (1919).

   [284]. Id.

   [285]. Id.

   [286]. Claims and Duties on Wines and Cotton (Franco-American Treaty of 1831), July 4, 1831, 8 Stat. 430 (1831).

   [287]. Elizabeth Feaster Baker, Henry Wheaton, 1785-1848, at 142–43 (1937).

   [288]. See John Tyler, Dec. 3, 1844: Fourth Annual Message, reprinted by Miller Ctr., https://millercenter.org/the-presidency/presidential-speeches/december-3-1844-fourth-annual-message [https://perma.cc/FV4W-AJXZ].

   [289]. Baker, supra note 287, at 143, 227.

   [290]. Id. at 143.

   [291]. U.S. Tariff Comm’n, supra note 283, at 21.

   [292]. S. Journal, 43d Cong., 2d Sess. 857 (1874); see also VanGrasstek, supra note 203, at 94 (“[T]he Senate rejected it because legislators believed that the president had no constitutional authority to deal with commerce. Representatives reiterated that view in their aforementioned refusal to pass implementing legislation for the 1884 treaty with Mexico.”).

   [293]. Alfred E. Eckes, Jr., Opening America’s Market: U.S. Foreign Trade Policy Since 1776, at 65 (1995) (quoting S. Journal, 43d Cong., 2d Sess. 857 (1874)).

   [294]. More than half never even got a vote. See VanGrasstek, supra note 203, at 89, 94.

   [295]. These agreements were terminated following the passage of the Tariff Act of 1894, which removed duty-free access for many products.

   [296]. This is one of the few moments in the congressional-executive relationship before 1930 when the President interpreted his mandate broadly and arguably beyond the plain language of the 1890 Act. Congress intervened very quickly. Wilson–Gorman Tariff Act of 1894, ch. 349, § 71, 28 Stat. 509, 569.

   [297]. See Dingley Tariff Act of 1897, ch. 11, § 4, 30 Stat. 151, 204.

   [298]. U.S. Tariff Comm’n, supra note 283, at 21.

   [299]. Payne–Aldrich Tariff Act of 1909, ch. 6, § 3, 36 Stat. 11, 83. Beginning in 1904, a special arrangement allowed for preferential treatment of U.S. imports into Brazil, and in 1903, a treaty was made with Cuba. U.S. Tariff Comm’n, supra note 283, at 21.

   [300]. See Claussen, supra note 12, at 857–917.

   [301]. See id. at 859–60.

   [302]. See generally FCC v. Consumers’ Rsch., 606 U.S. 656, 701 (2025) (Kavanaugh, J., concurring) (commenting that Congress delegates at least in part because it must adapt legislation to “complex conditions involving a host of details with which the national legislature cannot deal directly” (citing A.L.A. Schechter Poultry Corp. v. United States, 295 U. S. 495, 530 (1935) and Panama Refin. Co. v. Ryan, 293 U.S. 388, 421 (1935))).

   [303]. See Claussen, supra note 12, at 868, 890; Esteban Ortiz-Ospina, Diana Beltekian & Max Roser, Trade and Globalization, Our World in Data (2014), https://ourworldindata.org/trade-and-globalization [https://perma.cc/M2EB-55GZ]; Yi Wen & Brian Reinbold, The Evolution of Total Trade in the U.S., Fed. Rsrv. Bank of Saint Louis (Mar. 2, 2020), https://www.stlouisfed.org/on-the-economy/2020/march/evolution-total-trade-us [https://perma.cc/UB6X-ARL5].

   [304]. 2 Emory R. Johnson, T.W. Van Metre, G.G. Huebner & D.S. Hanchett, History of Domestic and Foreign Commerce of the United States 243, 265 (1915).

   [305]. See Claussen, supra note 12, at 857.

   [306]. See id. at 865, 887, 910; Kathleen Claussen, Trade’s Security Exceptionalism, 72 Stan. L. Rev. 1097, 1146, 1161 (2020); Meyer & Sitaraman, supra note 12, at 648–49.

   [307]. See Alexandra Butler, Trump Tariffs: Full List of Countries Hit by U.S. President’s New Trade War Levies, Independent (Aug. 7, 2025) https://www.the-independent.com/news/world/americas/us-politics/trump-tariffs-list-india-canada-china-us-trade-b2803470.html [https://perma.cc/YKK6-R4RP].

   [308]. See Memorandum of Jan. 20, 2025: America First Trade Policy, 90 Fed. Reg. 8471 (Jan. 30, 2025).

   [309]. See Bown, supra note 2.

   [310]. V.O.S. Selections, Inc. v. United States, 772 F. Supp. 3d 1350, 1364 (Ct. Int’l Trade 2025) (per curiam) (noting that “the worldwide tariffs remain[ed] in place at 10 percent” at the time of writing), aff’d in part, vacated in part sub nom., V.O.S. Selections, Inc. v. Trump, 149 F.4th 1312 (Fed. Cir. 2025), aff’d sub nom., Learning Resources, Inc. v. Trump, 146 S. Ct. 628 (2026).

   [311]. Exec. Order No. 14,257, 90 Fed. Reg. 15041 (Apr. 7, 2025) (excepting articles encompassed by 50 U.S.C. 1702(b): steel, aluminum, automobiles, and automotive parts subject to current or future duties imposed pursuant to section 232 of the Trade Expansion Act of 1962; and other products enumerated in Annex II of the order, including: copper, pharmaceuticals, semiconductors, lumber, critical minerals, energy products, and articles from a trading partner subject to the rates set forth in Column 2 of the Harmonized Tariff Schedule of the United States).

   [312]. International Emergency Economic Powers Act, 50 U.S.C. § 1701(a).

   [313]. Exec. Order No. 14,257, 90 Fed. Reg. 15041, 15041 (Apr. 7, 2025).

   [314]. 50 U.S.C. § 1702(a)(1)(B).

   [315]. 50 U.S.C. § 1701; Proclamation No. 10896, 90 Fed. Reg. 9817 (Feb. 18, 2025). The tariffs announced in Proclamation No. 10896 were suspended until March 2025. Fact Sheet: President Donald J. Trump Imposes Tariffs on Imports from Canada, Mexico, and China, White House (Feb. 1, 2025), https://www.whitehouse.gov/fact-sheets/2025/02/fact-sheet-president-donald-j-trump-imposes-tariffs-on-imports-from-canada-mexico-and-china/ [https://perma.cc/4FQH-WRVC].

   [316]. Exec. Order No. 14,245, 90 Fed. Reg. 13829, 13830 (Mar. 27, 2025).

   [317]. Addressing Threats to the United States by the Government of Brazil, Exec. Order No. 14,323, 90 Fed. Reg. 37739 (Aug. 5, 2025); see also Initiation of Section 301 Investigation: Brazil’s Acts, Policies, and Practices, 90 Fed. Reg. 34069, 34070 (July 18, 2025).

   [318]. Addressing Threats to the United States by the Government of the Russian Federation, Exec. Order 14,329, 90 Fed. Reg. 38701 (Aug. 11, 2025).

   [319]. Learning Resources, Inc. v. Trump, 146 S. Ct. 628, 643–44 (2026) (“Our task today is to decide only whether the power to ‘regulate . . . importation,’ as granted to the President in IEEPA, embraces the power to impose tariffs. It does not.”).

   [320]. See Claussen, supra note 306, at 1117–23.

   [321]. See Proclamation No. 10896, 90 Fed. Reg. 9817 (Mar. 5, 2025); Proclamation No. 10908, 90 Fed. Reg. 14705 (Apr. 3, 2025); Proclamation No. 10962, 90 Fed. Reg. 37727 (Aug. 5, 2025); Proclamation 10976, 90 Fed. Reg. 48127 (Oct. 6, 2025); Proclamation No. 10984, 90 Fed. Reg. 48451 (Oct. 22, 2025).

   [322]. See Trade Expansion Act of 1962, 19 U.S.C. § 1862(3)(A). The steel, aluminum, and auto duties are based on Commerce Department investigations from President Trump’s term that the Biden administration elected not to close, thus allowing President Trump to immediately impose these duties upon resuming office. See Julian Arato, Kathleen Claussen & Timothy Meyer, The “America First Trade Policy” in Practice, 119 Am. J. Int’l L. 668, 671 (2025).

   [323]. See, e.g., Exec. Order No. 14,223, 90 Fed. Reg. 11359 (Mar. 6, 2025).

   [324]. Notice of Action: Nicaragua’s Acts, Policies, and Practices Related to Labor Rights, Human Rights and Fundamental Freedoms, and the Rule of Law, 90 Fed. Reg. 57807 (Dec. 12, 2025).

   [325]. Notice of Action and Proposed Action in Section 301 Investigation of China’s Targeting the Maritime, Logistics, and Shipbuilding Sectors for Dominance, Request for Comments, 90 Fed. Reg. 17114 (Apr. 23, 2025).

   [326]. Memorandum of Feb. 21, 2025: Defending American Companies and Innovators from Overseas Extortion and Unfair Fines and Penalties, 90 Fed. Reg. 10685 (Feb. 26, 2025).

   [327]. Exec. Order No. 14,276, 90 Fed. Reg. 16993 (Apr. 22, 2025); see also Notice of Action and Proposed Action in Section 301 Investigation of China’s Targeting the Maritime, Logistics, and Shipbuilding Sectors for Dominance, 90 Fed. Reg. 17114, 17116 (Apr. 23, 2025) (instituting fees on Chinese maritime transport vessels); Colin Grabow, New Shipping Fees and Requirements Pose Fresh Threat to US Economy, Cato Inst. (May 19, 2025), https://www.cato.org/blog/new-shipping-fees-requirements-pose-fresh-threat-us-economy [https://perma.cc/B5DQ-W4E9].

   [328]. 19 U.S.C. § 2411(b)–(c).

   [329]. See Notice of Modification of Section 301 Action: China’s Acts, Policies, and Practices Related to Technology Transfer, Intellectual Property, and Innovation, 83 Fed. Reg. 47974 (Sep. 21, 2018) (increasing duties on select imports from China by 10 percent, with a future increase to 25 percent for the relevant imports) (Trump administration); Notice of Modification: China’s Acts, Policies and Practices Related to Technology Transfer, Intellectual Property and Innovation, 89 Fed. Reg. 76581 (Sep. 18, 2024) (increasing duties on select imports from China, including electric vehicles, battery parts, solar modules, and critical minerals) (Biden administration).

   [330]. Office of the U.S. Trade Representative, Initiation of Section 301 Investigations: Acts, Policies, and Practices of Certain Economies Relating to Structural Excess Capacity and Production in

Manufacturing Sectors (Mar. 11, 2026), https://ustr.gov/sites/default/files/files/Press/Releases/2026/USTR%20301%20FRN%20Industrial%20Excess%20Capacity%203-11-26.pdf [https://perma.cc/GRK9-5AER].

   [331]. Proclamation No. 11012, 91 Fed. Reg. 9339 (Feb. 20, 2026).

   [332]. To determine the status of trade-related lawsuits currently pending against the Trump administration, see Alex Lemonides, Seamus Hughes, Mattathias Schwartz, Lazaro Gamio & Camille Baker, Tracking the Lawsuits Against Trump’s Agenda, N.Y. Times (Jan. 21, 2026), https://www.nytimes.com/interactive/2025/us/trump-administration-lawsuits.html [https://perma.cc/C235-497J].

   [333]. See Claussen & Meyer, supra note 4, at 1971–73 (citing and discussing cases).

   [334]. Am. Inst. Int’l Steel, Inc. v. United States, 806 F. App’x 982, 990 (Fed. Cir. 2020).

   [335]. TransPacific Steel LLC v. United States, 4 F.4th 1306, 1338 (Fed. Cir. 2021) (Reyna, J., dissenting).

   [336]. Learning Resources, Inc. v. Trump, 146 S. Ct. 628, 637–38 (2026).

   [337]. Opening Brief for Appellants, supra note 3, at 1; Oral Argument, Trump v. V.O.S. Selections, Inc., 146 S. Ct. 73 (2025) (No. 2025-1812) (on file with the California Law Review).

   [338]. Opening Brief for Appellants, supra note 3, at 1.

   [339]. Claussen & Meyer, supra note 4.

   [340]. Learning Resources, 146 S. Ct. at 638, 642.

   [341]. Gundy v. United States, 588 U.S. 128, 159 (2019) (Gorsuch, J., dissenting); see also FCC v. Consumers’ Rsch., 606 U.S. 656, 700–01 (2025) (Kavanaugh, J., concurring).

   [342]. Learning Resources, 146 S. Ct. at 642; see also id. at 663 (Gorsuch, J., concurring) (rejecting the claim that the imposition of tariffs involves a matter of overlapping constitutional powers).

   [343]. J.W. Hampton, Jr. & Co. v. United States, 276 U.S. 394, 409 (1928).

   [344]. See, e.g., Fact Sheet: President Donald J. Trump Further Modifies the Reciprocal Tariff Rates, White House (July 31, 2025), https://www.whitehouse.gov/fact-sheets/2025/07/fact-sheet-president-donald-j-trump-further-modifies-the-reciprocal-tariff-rates/ [https://perma.cc/65C4-VFXY].

   [345]. See supra Part IV.B.

   [346]. See Oona A. Hathaway, Curtis A. Bradley & Jack L. Goldsmith, The Failed Transparency Regime for Executive Agreements: An Empirical and Normative Analysis, 134 Harv. L. Rev. 629, 632–33 (2020).

   [347]. See Claussen, supra note 12, at 866.

   [348]. Trade Act of 1974, Pub. L. No. 93-618, § 102(b)(c), 88 Stat. 1978, 1982–83 (1975) (codified as amended at 19 U.S.C. § 2112(b)(1), (c)).

   [349]. Bipartisan Congressional Trade Priorities and Accountability Act of 2015, 19 U.S.C. §§ 4201–4210.

   [350]. See id. at § 4203.

   [351]. See, e.g., Ackerman & Golove, supra note 58.

   [352]. Section 301 of the Trade Act of 1974 authorizes the executive branch to enter into an agreement in the context of a Section 301 investigation to eliminate any burden or restriction on U.S. commerce that results from a trading partner’s actions. See Trade Act of 1974, 19 U.S.C. § 2411(a)(2)(B)(ii)(II) (authorizing the Trade Representative to circumvent mandatory action under subsection (a) upon finding that the foreign country has “agreed to an imminent solution to the burden or restriction on United States commerce that is satisfactory to the Trade Representative”). Section 232 likewise authorizes the President to negotiate an agreement that limits or restricts certain imports. Trade Expansion Act of 1962, 19 U.S.C. § 1862 (c)(1)(A)(ii) (noting that the President may negotiate “an agreement which limits or restricts the importation into, or the exportation to, the United States of the article that threatens to impair national security”).

   [353]. Kathleen Claussen, Trade’s Mini-Deals, 62 Va. J. Int’l L. 315, 326 (2022) (“[T]here are hundreds of trade-related agreements not approved by Congress post-negotiation regardless of their legislative or regulatory effects.”).

   [354]. See Kyla H. Kitamura, Cong. Rsch. Serv., U.S.-Japan Trade Agreements and Tariff Negotiations 1–2 (2025).

   [355]. See Press Release, Pat Toomey, Toomey Statement on Taiwan Trade Discussions (June 2, 2022), https://www.toomey.senate.gov/newsroom/press-releases/toomey-statement-on-taiwan-trade-discussions [https://perma.cc/3S3C-H8B7]; Hans Nichols, Biden to Offer New Economic Framework for Latin America, Axios (June 6, 2022), https://www.axios.com/2022/06/06/biden-to-offer-new-economic-framework-for-latin-america [https://perma.cc/7526-3EER].

   [356]. United States-Taiwan Initiative on 21st-Century Trade First Agreement Implementation Act, Pub. L. No. 118-13, § 7(3)(b), (e), 137 Stat. 63, 66–67 (2023) (codified at 19 U.S.C. § 2112 note).

   [357]. See Joe Biden, Statement from President Joe Biden on H.R. 4004, the United States-Taiwan Initiative on 21st-Century Trade First Agreement Implementation Act, White House (Aug. 7, 2023), https://bidenwhitehouse.archives.gov/briefing-room/statements-releases/2023/08/07/statement-from-president-joe-biden-on-h-r-4004-the-united-states-taiwan-initiative-on-21st-century-trade-first-agreement-implementation-act/ [https://perma.cc/4CP6-VD68].

   [358]. Megan Messerly, Daniel Desrochers & Ari Hawkins, Trump Wanted ‘90 Deals in 90 Days.’ Instead, He’s Finding Wins Where He Can, Politico (June 12, 2025), https://www.politico.com/news/2025/06/12/trump-wanted-90-deals-in-90-days-instead-hes-finding-wins-where-he-can-00403638 [https://perma.cc/78KF-E868].

   [359]. See Jamieson Greer, U.S. Trade Representative, Written Responses to Questions for the Record from Mike Crapo, Chairman of the Senate Comm. on Fin., The President’s 2025 Trade Policy Agenda (Apr. 8, 2025), https://www.finance.senate.gov/download/responses-to-questions-for-the-record-to-jamieson-greer [https://perma.cc/SZ93-TNPV].

   [360]. See Fact Sheet: Implementing the General Terms of the U.S.-U.K. Economic Prosperity Deal, White House (June 17, 2025), https://www.whitehouse.gov/fact-sheets/2025/06/fact-sheet-implementing-the-general-terms-of-the-u-s-uk-economic-prosperity-deal/ [https://perma.cc/5BLM-AVUK]. At the time of writing, no written binding international agreement had been completed.

   [361]. See Kevin Breuninger, U.S.-China Trade Talks End Without Extension of Tariff Truce, As Trump Weighs Options, CNBC (July 29, 2025), https://www.cnbc.com/2025/07/29/trump-china-trade-tariffs-bessent.html [https://perma.cc/HUD2-ZMT9]; Arendse Huld, Trump Raises Tariffs on China to 145%—Overview and Trade Implications, China Briefing (Apr. 11, 2025), https://www.china-briefing.com/news/trump-raises-tariffs-on-china-to-145-overview-and-trade-implications/ [https://perma.cc/S5VU-UV5U]; Lewis Jackson & Amy Lv, China’s Export Controls Are Curbing Critical Mineral Shipments to the World, Reuters (Apr. 21, 2025), https://www.reuters.com/world/china/chinas-export-controls-are-curbing-critical-mineral-shipments-world-2025-04-20/ [https://perma.cc/65LM-VRVN].

   [362]. Hannah Miao & Chun Han Wong, U.S., China Sound Confident Note After Trade Talks, Wall St. J. (Oct. 26, 2025), https://www.wsj.com/world/china/bessent-sounds-confident-note-after-trade-talks-with-china-f09d310d [https://perma.cc/C2GJ-CHW7].

   [363]. Fact Sheet, supra note 344.

   [364]. 19 U.S.C. § 2171(c)(1)(C); see also Ari Hawkins & Doug Palmer, Greer Defends USTR authority over trade deals, Politico (Feb. 6, 2025), https://subscriber.politicopro.com/article/2025/02/greer-defends-ustr-authority-over-trade-deals-002-02874 [https://perma.cc/LF9A-7QRG] (quoting U.S. Trade Representative Jamieson Greer as saying “With respect to trade authority in the United States government, my statute is the statute. It designates the Office of U.S. Trade Representative as the chief trade negotiator, negotiating trade deals”).

   [365]. See Claussen, supra note 353, at 353.

   [366]. See id.

   [367]. Kathleen Claussen, Trade Transparency: A Call for Surfacing Unseen Deals, 122 Colum. L. Rev. F. 1, 2, 3–7 (2022).

   [368]. See supra Parts III & IV.

   [369]. See supra Part II.B.

   [370]. A third response to these critiques advanced by the prior two administrations is that Congress has consented to the executive branch’s entrance into trade agreements based on delegated authorities, such as the statute that created the Office of the U.S. Trade Representative, 19 U.S.C. § 2171, originally part of the Trade Act of 1974, or Section 103(a), also of the Trade Act of 1974. Relying on these statutes to justify such an approach concedes that Congress must consent to trade agreements. But rather than refer to Congress’s silence, proponents point to the statutes as an implicit delegation not only to negotiate but also to conclude agreements. Given that Congress has expressly given and taken away trade agreement authority to and from the executive for more than a century, relying on such an implicit delegation seems to go too far.

   [371]. Greer, supra note 359, at 23.

   [372]. Harold Hongju Koh, Triptych’s End: A Better Framework to Evaluate 21st Century International Lawmaking, 126 Yale L.J.F. 337, 342 (2017).

   [373]. U.S. Const. art. I, § 8.

   [374]. See, e.g., 19 U.S.C. § 3521(c) (limiting duty increases to countries that are not members of the World Trade Organization and limiting the increases to the base or bound tariff rates listed in Schedule XX).

   [375]. See Clinton v. New York, 524 U.S. 417, 448–49 (1998) (holding the line-item veto unconstitutional because it repeals legislation without going through bicameralism and presentment).

   [376]. Cong. Rsch. Serv., 106th Cong., Treaties and Other International Agreements: The Role of the United States Senate 199 (Comm. Print 2001).

   [377]. Andrew Restuccia, Doug Palmer & Adam Behsudi, Trump Says He Will Withdraw from NAFTA, Pressuring Congress to Approve New Trade Deal, Politico (Dec. 2, 2018), https://www.politico.com/story/2018/12/02/trump-trade-canada-mexico-1006164 [https://perma.cc/JS2Y-UF3M].

   [378]. Curtis A. Bradley, Exiting Congressional-Executive Agreements, 67 Duke L.J. 1615, 1629 (2018). Other commentators disagreed. See, e.g., John C. Yoo, Laws as Treaties?: The Constitutionality of Congressional-Executive Agreements, 99 Mich. L. Rev. 757, 815 (2001) (“[Interchangeability] would be tantamount to granting the President a direct share of the legislative power—a result . . . that is at odds with our understanding of the executive power.”); Joel P. Trachtman, Power to Terminate U.S. Trade Agreements, 51 Int’l Law. 445, 447 (2018) (“[U]nilateral Presidential power to terminate these regimes will upset the balance of powers between the executive and legislative branches that exists today, and that was envisaged by the framers of the Constitution.”). Although the Supreme Court has never held that the President may withdraw from Article II treaties without Congressional consent, since the twentieth century the executive branch has presumed the President possesses such a power, and the majority of commentators have agreed.

   [379]. Authority to Withdraw from the North American Free Trade Agreement, 42 Op. O.L.C. 133, 147 (2018).

   [380]. Id. at 151 (quoting Prohibition of Spending for Engagement of the Office of Science and Technology Policy with China, 35 Op. O.L.C. 116, 120 (2011)).

   [381]. Id.

   [382]. See supra Part IV.B.

   [383]. Exec. Order No. 14,209, 90 Fed. Reg. 9587 (Feb. 14, 2025); Exec. Order No. 14,166, 90 Fed. Reg. 8611 (Jan. 30, 2025).

   [384]. U.S. Dep’t of State, Public Notice 12682, 90 Fed. Reg. 12200, 12200 (Mar. 14, 2025).

   [385]. Id.

   [386]. Id.

   [387]. See Elena Chachko, Kathleen Claussen & David Zaring, Off. of the Chair, Admin. Conf. of the U.S., International Regulatory Cooperation in Agency Practice: Assessment and Best Practices (2025), https://www.acus.gov/sites/default/files/documents/IRC-Final-Report-2025.pdf [https://perma.cc/7MQA-99NT].

Table of Contents Show

Introduction

How do you transition an economy from fossil fuels to zero-carbon energy? At a conceptual level, there are four major approaches reflecting two axes of difference. Decarbonization policy can operate through carrots (incentives) or through sticks (penalties), and it can be universal or selective. The distinction between carrots and sticks is relatively familiar: Any decarbonization policy faces a choice between penalizing the production of excessive carbon emissions or rewarding clean alternatives.[1] The distinction between universal and selective policy is less obvious, but equally important. Carrots or sticks can both be selective, meaning that they apply only to energy generation or usage in specific industries, or in specific firms singled out for inclusion in the policy. Alternatively, carrots or sticks can both be universal, meaning that they apply to all sources of carbon emissions and/or clean energy. Universal policies are said to have the appeal of being “technology neutral,” meaning that they are indifferent between any two technologies that reduce carbon emissions by the same degree.[2]

Consider examples of each of the four approaches. Starting with sticks, the selective sort are those policies that penalize or restrict some (but not all) carbon emissions, like the Environmental Protection Agency (EPA)’s rules on power plant emissions, promulgated pursuant to the Clean Air Act.[3] Those rules are selective because they pertain only to emissions in a particular industry (the power sector). An example of a universal stick would be a carbon tax, insofar as it would be imposed on all carbon emissions.[4] Selective carrots, on the other hand, include grants offered on a discretionary basis to renewable energy companies, as well as traditional technology-specific tax credits such as a tax credit for solar energy production. Finally, an example of a universal carrot would be a technology-neutral tax credit—a subsidy offered to any recipient that can displace carbon emissions with clean energy.

This Article explores the appeal—and unintended consequences—of embracing universal carrots, the option Congress has largely chosen in recent years. Universal carrots are theoretically well suited to encouraging innovation in clean energy technology because they are technology neutral. But they carry a distinctive risk: that in their universality they will overflow beyond their targeted recipients and instead end up in the wrong hands, enriching the very carbon emitters they were meant to exclude.[5]

In order to deal with that concern, a policy of universal carrots needs rules and/or standards that define the universe within which carrots can be handed out freely. In the case of decarbonization policy, that is the task of carbon accounting. Carbon accounting is the discipline of measuring the greenhouse gas (GHG) emissions that result from an industrial process.[6] Over the last few years, carbon accounting has emerged as a critical measurement paradigm within environmental law.[7] In an influential article, Shelley Welton identified conceptual flaws in carbon accounting as one of several critical risks to the “net-zero” paradigm of reducing carbon emissions.[8] More recently, Leehi Yona explained how GHG undercounting has the potential to undermine various bodies of regulatory law.[9] In general, scholars who have addressed emissions accounting have taken for granted that if it is to become part of our law, it will be part of environmental law.[10] But in fact, Congress has assigned this baroque task of scientific measurement and accounting not to any of the science-focused agencies, but to the Department of Treasury (Treasury) and the Internal Revenue Service (IRS), because those agencies administer the universal carrots. For better or worse, real-world carbon accounting will be a product of tax law.

Is tax law an appropriate home for gatekeeping access to the universal carrots of clean energy policy? David Weisbach and Jacob Nussim reason that it makes sense to channel spending through tax law only when “the coordination benefits between the tax system and the other program are high and the specialization benefits of separate programs are low.”[11] So, for example, welfare programs are a good candidate for tax integration because the tax system and welfare system make determinations along similar dimensions, like applicants’ incomes.[12] But integrating carbon accounting into tax law does not only mean that the IRS will become responsible for administering it; it also means that tax experts at the Treasury will come up with rules, tax accounting firms will recommend avoidance strategies under those rules, tax lawyers will litigate disputes, and tax academics will propose principles for law reform. In this sense, integrating a policy domain into tax law means integrating it with a professional community eager to apply its own internally legitimate principles and standards.[13] Whether this integration is a good thing should depend on whether the other domain (here, carbon accounting) has much to learn from tax law.

This Article argues that the principles, rules, doctrines, and administrative strategies of tax law can usefully inform how the government conducts carbon accounting. Fundamentally, income tax accounting and carbon accounting are both projects involving tracking down a measure of interest (income, emissions) and deciding how to allocate it among responsible parties. At the level of theory and goal setting, concepts from tax theory have already influenced the prevailing conception of what carbon accounting is meant to achieve. These concepts include a comprehensive tax base, technology neutrality, and additionality, the principle that one should only be able to take credit for effects that would not have occurred without one’s actions. At the level of rules and doctrine, tax law’s experience in determining when to split income among taxpayers and when to blend disparate tax rates together can usefully inform questions of splitting and blending within emissions accounting. And at the level of administration, tax law’s familiarity with third-party verification—including verification from other administrative agencies—suggests a path for overcoming the technical challenges of carbon accounting that the IRS could not manage alone. In sum, tax law is surprisingly well positioned to house the law and administration of carbon accounting.

Yet even tax law cannot easily dispatch with the risks of universal carrots. Universal carrots are supposed to be available to anyone who produces energy without GHG emissions. That standard, however, gives taxpayers complete discretion in how they achieve zero emissions. Like other performance standards, it opens up boundless degrees of freedom for clever pseudo-compliance that violates the purpose of the law.[14] This is a familiar conundrum in tax law, where taxpayers invent “tax shelters” that technically comply with the rules but substantively avoid tax. This Article argues that energy firms will likewise develop “carbon shelters,” transactions that qualify for green subsidies while making use of high-emission energy.

In fact, the current set of green-subsidy rules actually create carbon shelters by granting taxpayers “safe harbors,” fact patterns that are treated as compliant with a zero-emissions standard even though they likely fall short of it.[15] Safe harbors are administratively simpler than pure performance-based regulation, but in practice, many of the safe harbors designed by the Treasury and the IRS achieve that simplification by giving taxpayers a surefire defense for sheltering emissions.[16] This Article diagnoses logical flaws in many of the safe harbors proposed in green-subsidy rulemaking and cautions against relying on safe harbors as the solution to the rigors and uncertainty of performance-based regulation.[17]

Instead, I argue that tax law’s green-subsidy regulations should evolve in the opposite direction: Rather than granting safe harbors for potentially high emission practices, the law should deny subsidies under fact patterns where positive emissions are likely but hard to measure. In other words, when administrative complexity makes it impossible to stick with a pure performance standard, the law should default to unsafe harbors (or prohibitions) rather than safe harbors.

This cautious approach is the appropriate response to the asymmetric downside risk posed by carbon shelters. The history of carbon accounting has been a continual story of underestimating the full scope of emissions before updating the standard methodology to account for emissions previously ignored.[18] Especially when new energy technologies become eligible for subsidy—the very point of technology neutrality—regulators will probably underestimate their consequential emissions in the first instance. And because most green tax credits are uncapped, there is no limit to the amount of money the government might become obligated to spend on carbon shelters—the very opposite of the statutory intent to reward green energy.

To counteract this risk, energy tax law should adopt broad anti-abuse rules that rule out subsidy eligibility for certain classes of transactions that, while not necessarily abusive in every instance, are particularly likely to serve as shelters. This approach borrows its logic from a prominent strategy used to combat tax shelters. There, Congress has written broad rules targeting risky fact patterns. The passive loss rules reflect this approach: While the concept of a passive loss is entirely coherent within income tax accounting, experience teaches that taxpayers opportunistically generate passive losses to offset nonpassive income.[19] Similarly, carbon offsets and negative emissions technology are theoretically valid constructs within carbon accounting, but the risks of taxpayers exploiting those concepts to shelter emissions are immense. For that reason, this Article proposes a broad prohibition on including negative emissions in green-subsidy carbon accounting.[20]

This Article proceeds as follows. Part I recounts the intellectual and policy history by which selective carrots and universal sticks gradually transformed into the near-universal carrots that define federal energy policy today. Part II examines the challenge of making carrots universal by comparing all potential subsidy recipients on the basis of their carbon emissions. This Part introduces the conceptual challenge of measuring emissions as well as the administrative challenge of instantiating emissions accounting inside the IRS. Part III then argues that the federal government’s current approach to emissions accounting will spur the proliferation of carbon shelters, transactions that win green subsidies while increasing emissions. This Article proposes to head off carbon shelters by developing broad anti-abuse rules, modeled on prophylactic rules from tax law.

I. The Rise of Universal Carrots: A Brief History

This Part presents the intellectual and policy history of how carbon accounting has become part of tax law. I argue that carbon accounting is a necessary ingredient for any regime of technology-neutral, universal energy incentives. Therefore, understanding the emergence of carbon accounting requires understanding the move toward these universal carrots. This Part begins by depicting a longtime consensus in energy policy around addressing climate change through sticks, including the universal stick of a carbon tax. It then explains why Congress instead deployed selective carrots for several decades. Finally, it narrates the recent shift from selective to (mostly) universal carrots—carrots completely reliant on carbon accounting for their basic functioning.

A. From Sticks to Carrots

For decades, the conventional wisdom was that any effort to tackle climate change through the tax code would take the form of a carbon tax.[21] The intellectual architecture undergirding the idea of carbon taxation can be traced back to the foundations of environmental economics.[22] A.C. Pigou, the Cambridge economist, identified what are today known as externalities: the effects of an economic transaction on third parties.[23] Today, taxes aimed at discouraging the imposition of harms onto others are commonly called “Pigouvian.”[24] In his famous 1977 paper, William Nordhaus advocated for the uniform Pigouvian taxation of carbon emissions across the economy.[25] Nordhaus reasoned that taxes would be superior to regulating the allowable quantity of emissions because the cost of abating a ton of carbon varies from sector to sector, and so the approach of regulating quantities would have to be customized for every application, whereas a single tax rate per ton of carbon could be applied uniformly to the entire economy.[26] This principle of uniformity is seen as one of the main virtues of carbon pricing: It does not force anyone in particular to reduce their emissions but allows the market to decide where it is most cost-effective to decarbonize and where continuing to emit is worth the high cost.

The carbon tax idea, brought to maturity by Nordhaus, soon became one of the most popular ideas within academic economics. The “Largest Public Statement of Economists in History” is an open letter endorsing a carbon tax and dividend, originally published in The Wall Street Journal in 2019 and now signed by over 3,500 U.S. economists.[27] Former Treasury Secretary Larry Summers called it “one of the few ideas of economic policy that commands broad, bipartisan support.”[28] The academic popularity of carbon taxation has brought modest real-world success. Finland, Sweden, Norway, Denmark, and Poland adopted carbon taxes in the early 1990s.[29] Switzerland, New Zealand, Australia, the United Kingdom, and France have since adopted versions of carbon taxes, although France suspended its tax after one year of operation in 2018.[30] Cap-and-trade markets, an alternative approach that sets a pollution quota and allows participants to sell emissions permits, have been more politically successful, emerging in the European Union, the northeastern United States, and California.[31]

Despite its immense academic popularity, carbon taxation has its own shortcomings.[32] Its most significant weakness is one of political economy: Carbon taxes are difficult to achieve in a democracy because they are experienced as a cost to all but a concentrated benefit to none.[33] From a political durability perspective, it is easier to pursue clean energy policies that “create sticky interests by incentivizing multidecade investments in clean energy infrastructure.”[34] Beyond the political challenge, a concern about efficacy is that a carbon price threatens emitting firms with higher operating expenses but does not directly subsidize the large capital expenses that will be necessary to substantially decarbonize the economy.[35] In principle, the promise of lower operating expenses can convince firms to make capital investments since, over a long enough time, the savings will recoup the initial investment. But firms will only do this once their current assets exhaust their useful life, which will be several decades hence for many gas and coal plants.[36] Furthermore, due to status quo bias and risk aversion, it often takes an extreme incentive to encourage firms to make a risky change.[37] In light of these critiques, some scholars have concluded that decarbonization policies should offer concentrated incentives (i.e., subsidies) rather than the diffuse incentives of a carbon price.[38]

Indeed, irrespective of the decades of academic discourse on carbon taxation, subsidies for low- and zero-emissions energy technology have long found more success at the federal level. Subsidies can be justified on the same Pigouvian grounds as taxes: If there are positive social externalities to some behavior, a subsidy allows anyone to “fully internalize the benefits to society of his or her behavior,” which leads to an optimal supply of the subsidized behavior by reducing its price.[39]

Starting in the 1960s, Congress enacted a string of tax subsidies for green-energy technologies. The first step was the initial investment tax credit (ITC), enacted in 1962.[40] The investment credit permitted taxpayers to claim a credit for 7 percent of the amount invested in new capital equipment, with no limitation by type of equipment or industry.[41] In 1978, motivated by reducing dependence on Middle Eastern oil, Congress moved to supplement the investment credit with a more targeted version exclusively for investments made in alternative energy equipment.[42] Congress defined six categories of credit-worthy technology: alternative energy property, solar or wind energy property, specially defined energy property, recycling equipment, shale oil equipment, and equipment for producing natural gas from geopressured brine.[43] Each category reflected a distinct subsector of the burgeoning renewable energy industry.

In 1992, Congress created the newest form of energy subsidy: a production tax credit (PTC) for the production of electricity from renewable sources.[44] Whereas the ITC was distributed in proportion to money invested in renewable energy equipment, the PTC would be distributed in proportion to kilowatt-hours of electricity generated and sold.[45] In this way, the logic of the PTC was more closely tied to the ultimate outcome of interest: the quantity of renewable energy consumed and its displacement of fossil alternatives. The PTC was also somewhat more targeted than the ITC had been in terms of technological scope: It initially only applied to electricity produced from wind or closed-loop biomass.[46] Additional methods of electricity generation were gradually included; most notably, Congress added solar in 2004.[47]

Solar and wind have been the two most widespread sources of renewable energy generation since the beginning of the renewable industry in the 1970s. Although frequently grouped together, solar and wind have different use cases, market structures, and cost constraints.[48] The original design of the PTC was much more favorable to wind than to solar: The subsidy level was enough to make wind cost competitive with fossil generation but not enough to do the same for solar.[49] The ITC was equally one-sized, if better suited to solar than to wind. Paired together, they were adequately targeted to an infant renewables sector composed almost entirely of those two technologies.[50] But as new green energy technologies came onto the scene, it became increasingly difficult to design subsidies customized to the needs of each one. In that context, this Article next turns to the development of universal carrots, subsidies meant to apply equally across different technologies.

B. From Selective to Universal

The most influential principle in the evolution of modern green-subsidy policy has been the conviction that subsidies should be neutral among technologies. According to this view, subsidies should not be made available only for specific technologies (e.g., solar and wind energy) that policymakers may favor but should be open to any technology that can achieve a certain goal (e.g., electricity generation or fuel combustion) with low or zero emissions. This conviction grew directly out of the academic consensus behind carbon taxes, which have a virtue of making all sources of carbon emissions more expensive and therefore equally benefiting all technologies that might mitigate emissions. As David Schizer explains, “covering all sources of harm (with taxes or permits) is easier than reaching all options for abating this harm (with subsidies).”[51] In other words, it is easier to identify the relevant harm—carbon emissions—than to identify all possible solutions for that harm.

If it were not possible to have a carbon tax, Schizer and likeminded peers argue that any subsidy-based alternative should at least be broadened to include any low-emissions technology rather than attempting to “pick winners.”[52] Zach Liscow and Quentin Karpilow summarize the “conventional view” among economists and legal scholars that the “government should not attempt to directly influence the direction of technological development” but should instead “tax pollution, provide technology-neutral innovation support, and then let the cards lay where they fall, trusting the private sector to respond in the cheapest, most efficient way.”[53]

And yet, some scholars who cheer technology neutrality in the case of a carbon tax have warned that it could not be replicated in a world of green subsidies. David Weisbach explains that “we cannot design a technology-neutral subsidy” in the sense of being neutral between fossil energy sources and the green technologies that might replace them.[54] This is because the emissions impact of a subsidy depends on which dirty technology it helps replace, and so a subsidy for wind cannot simultaneously be neutral between wind and coal and between wind and gas.[55] If the subsidy value for wind is equivalent to an efficient carbon tax on gas (i.e., one that accounts for the negative externalities of an equivalent amount of gas generation), then it must be too low in relation to coal, because coal produces more emissions than gas.[56] If the subsidy for wind is efficient in relation to coal, then it must be too high in relation to gas. The only way around this problem would be to customize subsidies to account for which fossil energy source the subsidized green energy is likely to replace—a horrendously complicated undertaking.[57]

A second limitation to the idea of technology neutrality is that, from an innovation policy perspective, the goal of spreading government support among all equally performing technologies may miss the power of “innovation snowballing.”[58] In certain cases, there are increasing returns to concentrating investment within a select few technological paradigms. Encouraging investment in technologies related to solar energy (for one example) will reduce the costs of solar and attract additional entrepreneurs to the solar sector, potentially leading to an innovation snowball that would be delayed if the initial investment were spread among many technologies.[59] The innovation snowballing perspective is the best justification for investments that Congress has directed the Department of Energy (DOE) to coordinate in selected clean energy paradigms like hydrogen and carbon capture.[60]

Nonetheless, the ideal of technology neutrality and the closely associated method of performance-based regulation have been central to the evolution of green subsidies over the past fifteen years. Senator Max Baucus encapsulated the appeal of technology neutrality in the opening remarks of a 2009 hearing: “[T]he government might just set a performance standard, regardless of the technology employed. We could encourage things like reductions in greenhouse gas emissions, improvement in efficiency, or increased energy content. And then we would leave the job of picking the best technology to the competition of the market.” [61] The first notable proposal came in 2013, when a draft bill overhauling the energy tax credit landscape would have reduced forty-two separate credits into two technology-neutral credits, one for all possible means of generating electricity and the other for all possible transportation fuels.[62] That particular bill did not make it onto the President’s desk, but the basic principles of that discussion draft reappeared in the Inflation Reduction Act (IRA), which finally enshrined technology neutrality in law.[63]

C. Universal Carrots Require Carbon Accounting

In 2021 and 2022, Congress enacted sweeping energy-related tax credits. The Congressional Budget Office estimated that the new credits would cost $400 billion, while private sector analysts concluded that the figure might actually double to $800 billion.[64] The uncertainty between the two sums arose because Congress made most of the credits available to as many firms as qualified, in as large a volume as those firms wished to pursue the incentivized activities.[65] What distinguished these green subsidies from prior ones was not just their size but the centrality of GHG emissions accounting principles found within their rules. Post-IRA, six major tax credits relied heavily on emissions accounting.[66]

The practice of GHG emissions accounting first emerged in response to the United Nations’ Kyoto Protocol, which asked participating nations to compile national emissions inventories.[67] The Kyoto Protocol and its successor, the Paris Agreement, are concerned not just with measuring emissions but with assigning responsibility to specific parties—in that case, nations.[68] For non-state actors, the emissions accounting protocol of choice has been the GHG Protocol, the source of the now-famous tripartite distinction between scopes of emission accounting.[69] Under this approach, Scope 1 emissions include an actor’s “direct” GHG emissions, Scope 2 emissions include indirect emissions resulting from the actor’s electricity use, and Scope 3 emissions include all other indirect emissions within the actor’s supply chain.[70]

At the federal level, emissions accounting first appeared in a 2007 amendment to the Clean Air Act authorizing the EPA’s Greenhouse Gas Reporting Program, which requires emissions reports from facilities that emit at least 25,000 megatons of CO2 annually.[71] The EPA estimates that this program covers 85 to 90 percent of total U.S. greenhouse gas emissions from over eight thousand facilities.[72]

Emissions accounting is also critical to the EPA’s Renewable Fuel Standard program, where the EPA must approve putatively renewable fuels before refiners and importers of gasoline and diesel fuel can claim credit for incorporating those fuels into their mix. When companies petition for a new “fuel pathway” (a type of fuel made from a specific feedstock, e.g., jet fuel made from soybean oil), the EPA conducts an analysis of the total “life-cycle” emissions of the fuel pathway.[73] This form of evaluation, known as life-cycle assessment (LCA), is the core exercise of emissions accounting. As of 2013, the EPA had approved ten of the forty-two petitions it had received for new fuel pathways.[74] These fuel pathway evaluations have served as the cutting edge of government-supervised emissions accounting.[75] Emissions accounting has also made its way into federal securities law under an SEC rule that requires firms to include emissions-related disclosures in their annual reports.[76]

The first appearance of emissions accounting in the tax code came in 2018. Prior to 2018, there was already a tax credit for capturing and disposing of carbon oxides. When carbon oxide is captured and disposed of in secure geologic storage, there is no need for complicated emissions accounting: Every ton of carbon stored is one fewer ton in the atmosphere, simple as that.[77] But in 2018, Congress amended the preexisting tax credit for carbon oxide sequestration to add a new subcredit for the utilization (i.e., rather than just sequestration and storage) of carbon oxides.[78] Carbon oxide can be “utilized” by being sold for use in oil wells, concrete production, chemical manufacturing, or carbonated beverage bottling. In those scenarios, the amount of carbon emissions actually displaced is highly fact specific.[79] Congress therefore made the carbon capture and utilization credit dependent on the amount of carbon oxide “permanently isolated” or “displaced from being emitted” into the atmosphere.[80] Furthermore, Congress directed that those terms be measured based on “an analysis of lifecycle greenhouse gas emissions”—an instruction to engage in emissions accounting.[81]

The remaining instances of emissions accounting in the tax code all appeared in August 2022 in the IRA.[82] The credit for sustainable aviation fuel paid at a rate indexed to the “lifecycle greenhouse gas emissions reduction percentage,” or the extent to which the fuel’s lifecycle greenhouse gas emissions are lower than that of petroleum-based jet fuel.[83] The credit for clean fuel production was set to replace the sustainable aviation fuel credit and all other fuel-specific credits in 2025, and it too subsidized fuels based on their emissions rates.[84] The IRA offered credit for clean hydrogen at four tiers, where the more valuable tiers depended on the hydrogen being produced through a process that results in a lifecycle greenhouse gas emissions rate below certain thresholds.[85] Congress created the “Clean Electricity Credits” to replace the longstanding electricity Production Tax Credit and Investment Tax Credit in 2025. The primary difference between these new credits and their predecessors was that, while the old ones were made available for specifically enumerated electricity generation technologies (as discussed above), the new ones were available for any electricity produced at a facility with a “greenhouse gas emissions rate not greater than zero.”[86] In just a few years, “lifecycle greenhouse gas emissions” and “greenhouse gas emissions rate” had become critical terms to the operation of the new era of universal carrots.

D. Universal Carrots Get Less Universal

In 2025, Congress passed and President Trump signed the One Big Beautiful Bill Act (OBBBA), a partial rebuke to the suite of green subsidies enacted under President Biden. The most important change was the acceleration of the termination date for solar and wind ITCs and PTCs, and hydrogen PTCs: Credits will no longer be available for solar, wind, and hydrogen facilities placed in service after 2027.[87] Removing solar, wind, and hydrogen from the scope of green energy credits is obviously a step away from universality. But it is not quite a return to the world of selective carrots that predated the IRA. Prior to the IRA, Congress had a whitelist of approved energy technologies. Post-OBBBA, Congress has a blacklist of disallowed technologies but still relies on an open-ended carbon accounting standard to determine which non-blacklisted technologies can qualify for subsidy.

The impending disappearance of certain green technologies from credit eligibility moderately softens some of the concerns raised in this Article. On the one hand, solar and wind never would have posed the carbon accounting challenges discussed throughout this Article. There is no debate that they are zero-emissions technologies, and the Treasury’s regulations have provided an expedited process for confirming that they are entitled to full credit under the technology-neutral ITC and PTC.[88] The hydrogen credit, on the other hand, would have been a major locus of accounting contestation and will remain so through 2027 when its credit expires. I discuss issues with implementing the hydrogen credit at various points herein, both because tax credit expiration dates are rarely set in stone and because of general lessons exemplified by the hydrogen credit and its regulations.

Moreover, other provisions of the OBBBA added to the concerns about measurement and carbon shelters raised below. As discussed in the next Part, the IRA wrote a maximally ambitious version of carbon accounting into statute and left the Treasury with the job of making pragmatic compromises in regulations. The OBBBA took a different approach, adding various statutory rules about what factors regulators may and may not consider in carbon accounting. And so, while the OBBBA validates the continued centrality of carbon accounting to energy tax law, it also expands opportunities for carbon sheltering, which I delve into in future sections.[89]

II. Administering Universal Carrots

In this Part, I explore the challenge of making carrots universal, which is fundamentally the challenge of distinguishing low-carbon recipients from high-carbon alternatives. The former are meant to be included, and the latter kept out. This Part breaks down the task of administering universal carrots into three components: what emissions get counted, who does the counting, and how to count.

The first challenge of administering universal carrots is the question of which emissions count. I begin by articulating the principles that carbon accounting is meant to uphold, according to Congress’s mandate. Congress has instructed the tax agencies to count allemissions “related to” subsidized projects.[90] I argue that this directive requires conducting both attributional and consequential carbon accounting. The former is concerned with tracking down emissions in a product’s supply chain and the latter with modeling indirect emissions caused by a production process, even when they occur at a distance elsewhere in the economy. I show that the central normative principles behind both attributional and consequential accounting have either traveled over from, or have informative parallels in, the world of tax law.

The second challenge is the question of which government officials oversee taxpayers’ carbon accounting. The IRS will rely on peer agencies like the EPA and DOE, both to develop standardized carbon accounting models, and to evaluate more bespoke models submitted by taxpayers. But even with the greater expertise of those agencies, the difficulty of evaluating carbon accounting is that, according to statute, each taxpayer’s submission is subject to a broad standard—whether the taxpayer achieves zero emissions—rather than a clear-cut set of rules. Regulators face the prospect of having to double-check each taxpayer’s modeling choices, which threatens to be administratively unworkable, no matter which agency is given the job.

This leads to the third challenge: If an open-ended performance standard is not administratively viable, what can replace it? The Treasury and IRS have answered this question by creating safe harbors, rules that simplify adjudication by treating certain fact patterns as presumptively zero-emission, even if they might actually not be. In effect, the Treasury and IRS are sacrificing accuracy for administrability. But sacrificing too much accuracy risks producing self-defeating outcomes and undermining the task of reducing GHG emissions.

A. What Gets Counted

Once Congress has committed to technology neutrality, and once it has determined that the test subsidy seekers must pass is achieving a low GHG emissions rate, the next question must be: Which emissions count?

In tax law, the first step in designing any new tax is to define the base to which it applies, i.e., to identify the set of economic activities subject to tax. For example, the federal income tax has a base of “all income from whatever source derived,”[91] while the excise tax on coal has a base of “[the value of] coal from mines located in the United States sold by the producer.”[92] Perhaps the most consequential step in defining the scope of green subsidies is to name which emissions are inside and outside the base.

In tax theory, there is a normative ideal of a “comprehensive tax base.”[93] The idea is that the base should be as broad as possible to avoid distorting economic decision-making. If the base is less than fully broad, actors will have an incentive to move their activities outside the base. The comprehensive base can also be justified on fairness or neutrality grounds: It is neutral among all activities because all are included within the base.[94] The comprehensive base concept can be applied to an income tax, a consumption tax, or even a carbon tax: In each instance, the idea is that all cases of the target phenomenon (income, consumption, carbon emissions) should be included.

But there is something inherently more difficult about applying the comprehensive base ideal in the context of a subsidy compared to a tax. Assuming that every person and entity who might potentially conduct the taxable activity is subject to a tax, nothing should escape the base. If everyone who might potentially earn income is subject to an income tax, the base has a chance to be comprehensive (assuming no exemptions, etc.). But consider a subsidy that, by design, will only flow to a select cohort of recipients. If the logic of the subsidy is “pay everyone who produces energy without carbon emissions,” but not everyone involved in producing energy is examined, it is possible that the industry will arrange itself such that subsidy seekers will not generate any emissions directly but instead outsource their emissions-generating activities to other firms that do not apply for the subsidy. The firms that cannot hide their connection to carbon emissions will not apply for the subsidy but will nonetheless benefit from it by continuing to do business with the subsidy winners and possibly bargaining for some share of the subsidy spoils.

To guard against the risk of mistakenly subsidizing firms with hidden emissions, Congress has adopted an expansive understanding of which emissions should count. As described in Part I, each of the major green-subsidy statutes calls for an analysis of “lifecycle greenhouse gas emissions,” and in each statute that term is given the meaning as described in subparagraph (H) of section 211(o)(1) of the Clean Air Act (42 U.S.C. § 7545(o)(1)).[95] Congress had added this definition of lifecycle emissions for the purpose of the EPA’s Renewable Fuel Standard program.[96] The definition is:

The aggregate quantity of greenhouse gas emissions (including direct emissions and significant indirect emissions such as significant emissions from land use changes), as determined by the Administrator, related to the full fuel lifecycle, including all stages of fuel and feedstock production and distribution, from feedstock generation or extraction through the distribution and delivery and use of the finished fuel to the ultimate consumer, where the mass values for all greenhouse gases are adjusted to account for their relative global warming potential.[97]

The definition draws attention to two distinct ways carbon accounting could go wrong. First, one might fail to count emissions at every relevant stage in the production process, or “the full fuel lifecycle.”[98] As explained below, this is the challenge of attributional accounting: the task of attributing all relevant emissions, and only the relevant emissions, in a multistage, multiproduct supply chain to the product at hand. Second, one might fail to count “significant indirect emissions.” As explained below, this is the challenge of consequential accounting: the task of determining which emissions are causally attributable to a production process, even if they do not occur anywhere in the direct supply chain. Attributional and consequential emissions are both necessary to count in order to fulfill Congress’s mandate, but they implicate completely different justifications and require different accounting strategies as the rest of this Section details.

1.   Attributional Accounting

The task of attributional carbon accounting is to identify which emissions a given product should be held responsible for.[99] At a high level, this process has two steps: first, identify all the emissions connected to the product under evaluation (i.e., the emissions in its supply chain), and then determine which ones should be assigned to the product. The first step is about making sure to include all emissions, encompassing those upstream and downstream of the product. The second step is about deciding whether any of those emissions should be allocated to other products to avoid double counting. The need to count all emissions both upstream and downstream of the product under analysis follows directly from Congress’s directive to count emissions at “all stages of fuel and feedstock production and distribution, from feedstock generation or extraction through the distribution and delivery and use of the finished fuel to the ultimate consumer.”[100] Analysts (and Treasury regulations) refer to that expansive scope of analysis as “cradle to grave.”[101]

After one has identified all emissions from cradle to grave, the next step of attributional accounting is to decide whether any of those emissions should nonetheless be assigned to some other product rather than the one under evaluation. Depending on the purpose at hand, carbon accounting can be applied in either a mutually exclusive or an overlapping manner. In an overlapping approach, emissions are assigned to every actor with the capability of reducing those emissions (e.g., assign the same supply chain emissions to a producer, transporter, and customer).[102] The overlapping approach allows double counting, which can be appropriate if the goal is to assign responsibility to all actors who might be able to reduce emissions.[103]

If double counting is unavailable, the subsidy applicant will be tempted to split emissions among multiple products so as to reduce the emissions associated with the product under investigation. Often, a single industrial process yields multiple products. From a subsidy claimant’s perspective, the incentive is to attribute some of the emissions associated with the product covered by the law to other products not covered by the law. For example, the production of hydrogen often generates steam as a byproduct, and a taxpayer might like to allocate any emissions from this process to the steam rather than to the hydrogen (because the steam is not subject to any tax or subsidy). Depending on how emissions are allocated among products—e.g., by volume or sales value—the producer might try to increase the prominence of secondary products to deflate the emissions share assigned to the primary product. Perverse incentives for generating waste have already plagued carbon offset markets; here, there would be an incentive to produce what was previously classified as waste and sell it (no matter how cheaply) as a byproduct to soak up the primary product’s emissions.[104]

The prospect of producers strategically allocating emissions resembles a familiar problem in tax law of taxpayers allocating income to low-tax related parties. In general, the tax law does not allow taxpayers to offload income to related parties while continuing to benefit from it. Parents might wish to gift income-generating assets to their lower-bracket children but federal law taxes children’s investment income above a threshold at the parents’ marginal rate.[105] U.S. companies might wish to license intellectual property or sell services to their controlled foreign affiliates at below-market rates so as to allocate relatively more taxable income to low-tax jurisdictions and relatively less to the United States, but section 482 (the transfer pricing rule) allows the IRS to reallocate tax items between the parties in line with an arm’s-length standard, i.e., what an unrelated third party would have paid.[106] And a high-tax partner might wish to contribute an appreciated asset to a partnership and then distribute the asset to a low-tax partner, but the “mixing bowl” rules force the contributing partner to recognize the taxable gain.[107]

Among these tax law analogies, the transfer pricing rule is the best analogy for emissions splitting. Just as the transfer pricing rule necessitates a complex, fact-dependent inquiry into the proper arm’s-length price between related parties, so too does emission splitting require a complex, fact-dependent inquiry into the share of emissions properly attributable to the main product versus byproducts. Instead of looking to arm’s-length prices as the relevant baseline, here the IRS might limit eligible byproducts to those the taxpayer produced before a certain date, such as the date the IRA was enacted.[108] Similarly, the IRS could limit the allowable quantity of byproducts based on prevailing technical efficiency standards.[109] Relying on these heuristics is one way to protect against the abuse of allocating excessive emissions to byproducts.

2. Consequential Accounting

The purpose of consequential carbon accounting is to measure indirect emissions. Whereas attributional accounting starts with a known total universe of emissions and works to allocate them among products and responsible parties, consequential accounting starts with an industrial process and seeks to estimate the emissions it will cause elsewhere in the economy. The motivation for consequential accounting is the understanding that focusing only on the direct emissions of a given project “omits partial and general equilibrium responses.”[110] Much of the scientific literature GHG accounting relies on has been produced based on small-scale, localized studies. The concern is that these studies might underestimate the aggregate and systemic effect of many actors pursuing the same incentive at once.[111] So-called “indirect” emissions refer to emissions caused elsewhere in the economy by the industrial process under evaluation, most commonly via supply and demand effects. Indirect emissions are substantial: Based on evidence from existing carbon taxes in Europe, indirect carbon leakage through electricity and fuel markets is quantitatively larger than direct carbon leakage through supply chains.[112]

IRS’s legal directive to conduct consequential accounting can be found in the same clause of the Clean Air Act discussed above and incorporated into every green-subsidy statute.[113] The key language is “including direct emissions and significant indirect emissions such as significant emissions from land use changes . . . related to the full fuel lifecycle.”[114] “Indirect emissions” require consequential modeling to estimate. The example of indirect emissions offered by the statute—“such as significant emissions from land use changes”—provides the most useful guidance as to what emissions should be considered.[115]

The reference to land use changes invokes the category of indirect emissions that has received the closest study to date, even as it is just one example of a broader set of indirect effects that consequential carbon accounting should theoretically track. The requirement to measure indirect land-use emissions was first created under the EPA’s Renewable Fuel Standard rule.[116] There, the concern was that by turning crops into biofuels, the global economy would need to cultivate new agricultural land to replace the lost food supply, and that the resulting deforestation would cause substantial emissions.[117] Indeed, the Department of Energy has acknowledged that diverting a large fraction of global corn production for ethanol has caused enough land use changes (including nitrate leaching, phosphorus runoff, and soil erosion) to undermine the fuel emissions benefits.[118] In response, Congress instructed the EPA to account for land use changes, and the EPA interpreted this mandate to include such changes wherever in the world they might occur.[119] In the OBBBA, however, Congress reversed course and explicitly excluded land use effects from the calculation of emissions for the Clean Fuel Credit.[120] Excluding consideration of indirect land-use emissions goes much further than merely failing to require them; it is an example of an emissions-sheltering safe harbor, discussed below in Part III.

But land use changes are just one example of a broader category of what some analysts call “market-mediated emissions.”[121] Market-mediated emissions arise when producers use some scarce input of production (e.g., land, or solar energy), which reduces supply of that input for the rest of the economy, which causes other people or firms to substitute for the scarce input with some higher-emissions replacement. For example, consider a hydrogen production plant that runs by electrolysis, the process of splitting water into hydrogen and oxygen. Suppose the plant powers its electrolyzers with electricity purchased from a solar farm. Solar energy is zero-emissions in its own right, but suppose that a residential community had previously been buying the very same solar energy, and now that the hydrogen plant has stepped in, the community will have to buy electricity from a gas power plant. The resulting emissions are market-mediated: The hydrogen plant is indirectly responsible for the gas-fired emissions because it used the scarce solar resource without bringing any new solar (or other renewable energy) into existence. To extrapolate out from this example, whenever the subsidized activity relies on some input that other buyers might otherwise wish to procure, an indirect emissions analysis requires asking about the consequence of those buyers needing to find a replacement.

3. Additionality in Emissions Accounting

The requirement to count indirect and market-mediated emissions, discussed above, implicates the principle of additionality. Additionality is the principle that one should only be able to take credit for effects that occur on the margin, i.e., that would not have taken place absent the policy or action under evaluation. In carbon accounting, whenever someone claims that a production process uses only green energy, it is important to ask whether that green energy is additional to the overall energy system or whether it has been displaced from some preexisting green energy application. If the green energy would have existed anyway, it does not make sense to reward a subsidy claimant for using it. The additionality principle should also prevent taking credit for actions that were otherwise already required by law.[122]

Commentators have long recognized the additionality problem as a potentially severe flaw in tax subsidy programs.[123] When Congress subsidizes some behavior, there is often a possibility that most subsidy recipients would have conducted the behavior in the absence of a subsidy, and so the payments they receive are arguably wasteful.[124] For example, models produced by the National Renewable Energy Laboratory suggest that solar energy deployment would continue to expand and remain cost competitive even if tax credits were to expire, suggesting that those credits do not achieve additionality.[125] Tax policy analysts have often noted that it would be desirable to distinguish between activities that would have been conducted anyway and those that were triggered by a subsidy, and only pay for the latter group.[126] One relatively simple way to implement such a rule might be to subsidize only activities that exceed the taxpayer’s historical baseline. Congress designed the Research and Development Credit (R&D Credit) to work in just that fashion, indexed to increases in research expenditure over a firm-specific historical baseline.[127] Measuring increases over a firm-specific baseline remains a key operation in calculating the credit.[128] Multiple studies have found that each dollar of R&D Credit stimulates about a dollar in research investment. In other words, the credit achieves additionality.[129]

By making additionality an implicit requirement of consequential accounting, Congress, the Treasury, and the IRS are attempting to make firms responsible for indirect emissions that few previous climate policies have managed to internalize. The regulations under the Clean Electricity Credits interpret “significant indirect emissions” to include “market-mediated changes in related commodity markets.”[130] Similarly, the regulations under the Clean Hydrogen Production Tax Credit aim to incorporate “significant indirect emissions associated with electricity,” i.e., the market-mediated effect of creating new demand for electricity.[131] The regulations make clear that the Treasury interpreted Congress’s instruction as encompassing an additionality requirement, specifically that hydrogen plants must be powered by incremental, new clean energy sources, not preexisting clean energy.[132]

It is striking that the additionality-related safeguards in the Clean Hydrogen Credit and Clean Electricity Credits are much more administratively complex than the R&D Credit or any other additionality-related rules Congress or the IRS have attempted in the tax code to date. James Salzman and David Weisbach notice this asymmetry and ask why there is a “double standard” in the extent to which commentators worry about additionality in the world of carbon accounting and offsetting compared to the world of tax policy more generally.[133] They convincingly dispatch any distinction grounded in non-additional carbon offsetters being bad actors, in the difference between offsets preventing a bad versus subsidies achieving a good, or in offsets being quantity-based versus subsidies being price-based.[134] But Salzman and Weisbach do not address one important reason for taking additionality more seriously in the context of carbon accounting: the possibility that the intervention will not just fail to be additional but will actually subtract from the intended outcome.[135] Subsidies for green manufacturing (e.g., hydrogen production) often carry this risk: Not just that they will displace existing green energy sources but that dirty energy will fill the gap left by the displaced green energy, leading to higher net carbon emissions than in the absence of a subsidy. Green subsidies don’t merely risk wasting money; they risk perverse, self-defeating outcomes.

B. Who Does the Counting

For observers skeptical that tax law can accommodate carbon accounting, one major reason for doubt is the issue of administrative competence. How will the IRS, an agency with limited experience in environmental science or energy systems, enforce highly technical carbon accounting rules? Rather than administer domain-specific programs by itself, the IRS increasingly functions as a hub for gathering information and assessments from external authorities both within the executive branch and outside government. By relying on outside experts to make specific technical determinations, the IRS need not be limited by in-house capacity. Furthermore, just as it does in tax administration, the IRS relies wherever possible on information provided by independent third parties to cross-reference subsidy seekers’ claims.

Of course, climate policy is not alone in turning to tax law for its implementation. The tax code is also home to the federal government’s most significant undertakings in low-income housing policy, anti-poverty policy, and technological research and development policy. Congress relies so heavily on tax law for multiple reasons. For one, taxing and spending legislation enjoys a procedural advantage in overcoming the filibuster in a tightly divided Senate.[136] Second, as courts have become more skeptical of novel regulatory uses of the Commerce Clause, Congress has increasingly sought to wrap its regulatory efforts in the Spending Clause, where courts have remained relatively accepting.[137] Third, when it comes to disbursing money, it is superficially simpler for Congress to write an eligibility formula into the tax code rather than create a new program office and hire a workforce to deliver discretionary grants. We can observe the administrative difference between tax subsidies and nontax grants in the CHIPS and Science Act, which created parallel tax credits and grants for semiconductor manufacturing. The tax credit included no appropriation for additional program staff, whereas the Commerce Department received funds to hire approximately 140 people to administer semiconductor-related grants.[138] Of course, imposing an unfunded mandate on the IRS is not necessarily administratively superior in the long run. Doing so forces the IRS to rely on its peer agencies for technical support, as detailed in the rest of this Section.

Administrative law scholars have long recognized the theoretical appeal and empirical reality of interagency coordination.[139] Congress often directs agencies to coordinate with one another to share expertise. In tax rulemaking, Congress sometimes directs the Treasury to follow some other agency’s lead in defining a statutory term. For example, the semiconductor manufacturing tax credit (§ 48D) defines the critical term “legacy semiconductors” by reference to a simultaneously enacted provision that includes an equivalent concept, and Commerce Department regulations thereunder.[140] Likewise in adjudications, Congress sometimes directs the IRS to cross-reference data collected by other agencies as part of determining eligibility for a tax expenditure. For example, in administering the health insurance premium tax credit, the IRS coordinates with the Centers for Medicare & Medicaid Services in combining eligibility criteria from each agency’s respective databases.[141] And in guidance for prevailing wage requirements applicable to various tax credits, Congress directed the IRS to follow prevailing wage determinations published by the Department of Labor.[142] The widespread reality of interagency coordination challenges the notion that the IRS can only administer traditional tax laws.[143]

The IRS’s reliance on peer agencies is essential for adjudicating carbon accounting. The agencies most relied upon are the Department of Agriculture, the EPA, and the DOE, as detailed in the following. To start, in guidance under the Sustainable Aviation Fuel credit, the IRS announced that biofuel producers who use crops certified under the Department of Agriculture’s Climate Smart Agriculture pilot program will receive a reduction from the lifecycle emissions value otherwise calculated under the regulations.[144] The Climate Smart Agriculture program does not audit or monitor crops for any preexisting regulatory purpose, so an independent third-party verifier, rather than the Department of Agriculture, must certify compliance with the program.[145]

Relying on peer agencies is even more useful when the peer agency already collects relevant data under a preexisting program. For example, the IRS relies on the EPA for verifying the sequestration of carbon under the Carbon Sequestration Credit. In the course of developing its own Underground Injection Control program, the EPA had established a monitoring, reporting, and verification program to track carbon dioxide sequestered in underground wells.[146] The newest Treasury regulations require taxpayers to comply with either the EPA’s monitoring standard or a different, third-party monitoring standard.[147] This approach involving the EPA was created in response to a scandal where the Treasury Inspector General found that of over $1 billion in carbon sequestration credits claimed between 2010 and 2019, 87 percent of credits were claimed by taxpayers that did not have an approved EPA monitoring plan in place at the time the credit was claimed.[148] By setting eligibility for the tax credit based on demonstrated compliance with the EPA monitoring program, the IRS aims to avoid future discrepancies of this sort.

But among all the agencies, reliance on the Department of Energy is most critical to administering carbon accounting. The DOE has the technical expertise to build and evaluate life-cycle assessment (LCA) models, which play a key role in taxpayers’ claims for technology-neutral credits. For several of the green energy credits, the DOE has released a customized LCA model that satisfies the statutory requirement to measure lifecycle emissions. For example, the Sustainable Aviation Fuels Credit requires that LCA modeling be in accordance with either the Carbon Offsetting and Reduction Scheme for International Aviation, a model adopted by the International Civil Aviation Organization, or “any similar methodology” that satisfies the Clean Air Act definition of lifecycle emissions.[149] In response, the DOE developed an LCA model customized for modeling aviation fuel emissions.[150] This model modified the general-purpose GREET model, the DOE’s longtime standard-bearer emissions model. Likewise, the Clean Hydrogen Credit requires that lifecycle emissions be measured according to GREET “or a successor model (as determined by the Secretary).”[151] The DOE responded by developing a hydrogen-specific GREET model, 45VH2-GREET.[152] Treasury regulations provide that if the taxpayer’s feedstock and production pathway are included in the latest GREET, the taxpayer must use that model to determine its emissions rate.[153] The DOE has not yet developed a GREET model for electricity generation—likely the most daunting of these undertakings because of the vast range of possible technologies. But Treasury regulations provide a process for the Secretary to designate a suitable LCA model and for taxpayers to rely on emissions rates determined using such a model, suggesting that 45Y-GREET may be forthcoming.[154]

The DOE’s most significant role comes in adjudicating taxpayers’ petitions for emissions rates in circumstances not covered by the pre-approved GREET models. Both the Clean Hydrogen Credit and Clean Electricity Credits provide a process to petition the DOE for a “provisional emissions rate,” i.e., an emissions rate determined by the taxpayer’s own bespoke emissions modeling.[155] The petition process follows a template established by the Carbon Utilization Credit, which requires submitting an LCA report to the DOE for technical review.[156]

The DOE’s technical review of LCA reports is essentially the last line of defense in the distribution of green subsidies. Of course, the IRS can reject a credit application even after DOE approval, but it is hard to expect it would do so on technical, LCA-related grounds.[157] The DOE’s review is also the most opaque stage of the administrative process.[158] Taxpayers have been dissatisfied with the lack of visibility into the DOE’s review and have requested more information on emissions factors for comparison technologies (necessary for establishing baselines in LCA), materiality thresholds for possible deviation between taxpayer-conducted LCA and DOE models, and examples of approved LCA submissions.[159]

C.   How to Count: Performance Standards and Safe Harbors

The ambiguity and discretion that taxpayers decry in the DOE’s technical review process is no accident: It is the only way to adjudicate a scientifically contested performance standard like a GHG emissions rate. Performance-based regulatory standards work by defining a required outcome but leaving “the means of achieving that outcome to the discretion of the regulated entity.”[160] Put simply, performance standards are standards that can be achieved by unlimited potential means.

Performance standards differ from standards that specify exactly how a regulated entity must act—known as means standards, prescriptive standards, or design standards.[161] For example, the EPA’s regulation of diesel engine emissions obligated manufactures to pass an emissions test “while leaving it to each company’s discretion to determine how to attain that required outcome.”[162] Performance standards are praised for giving regulated parties flexibility to choose the least costly means of achieving the stated outcome.[163] In theory, this open-endedness should allow firms to innovate and develop new means of compliance over time without waiting for the law to change and accommodate them. Cary Coglianese refers to this feature of performance standards as allowing “longitudinal flexibility” (i.e., evolving methods of compliance over time), in addition to “cross-sectional flexibility” (i.e., enabling different firms to comply in different ways).[164] Performance-based regulation has been endorsed by Presidents of both parties and enshrined within the guiding documents of regulatory review.[165] At the 2009 Senate hearing that laid the intellectual groundwork for technology-neutral energy tax subsidies, an environmental economist testified that “[b]ecause of our limited ability to foresee technological solutions that are possible but do not yet exist, it is almost always more effective and economical to specify the energy or environmental objective rather than a specific means of achieving it.”[166]

Despite its justifiable appeal, performance-based regulation has distinct downsides, as often noted by administrative law scholars. For several reasons, performance-based standards are more difficult to administer than design standards.[167] To start, performance standards put pressure on legislators to choose the right measure of performance, because emphasizing performance on one metric will tend to undermine the importance of performance on other, not-explicitly relevant metrics. Next, performance standards invite conflict over adjudicating performance, whereas design standards tend to rely on cut-and-dry rules that are more easily adjudicated.[168] As a result, performance standards offer regulated parties less certainty and predictability about whether they will be considered compliant.[169] Performance standards will generally require monitoring, especially to the extent the standard invites parties to choose their own method of compliance.[170] Finally, performance standards may invite gaming or evasive behavior intended to narrowly satisfy the measurable standard while departing from regulators’ underlying goals. This is sometimes referred to as “teaching to the test.”[171] The end result is to increase the cumbersomeness and complexity of regulation.[172]

A GHG emissions rate is a particularly challenging performance standard to implement because there is no universally applicable formula for determining the emissions rate of an industrial process. There are agreed-upon modeling frameworks in the form of ISO standards and GREET models, but operating a model still requires countless judgment calls, customizations, and ad-hoc modifications. Modeling is relatively foreign to the tax law and to the IRS, which is why the key modeling responsibilities within emissions tax law are outsourced to the DOE and to modeling tools it publishes. Whoever conducts LCA must make decisions about the upstream and downstream boundaries of the system, allocating emissions among coproducts, establishing a counterfactual for the emissions that would have been caused by input products, and so on.[173] If the IRS and DOE are committed to treating GHG emissions rate as a performance standard, then they must assess each of these choices on a fully individualized basis. And unlike in attributional life cycle assessments, where third-party information sources can help verify taxpayer claims, there is no authoritative third-party source on a project’s consequential emissions.

In light of the uncertainty and intense regulatory workload of the performance standard approach, it is easy to see why both regulators and regulated parties might prefer to see relatively more rule-based adjudication and relatively less of the nebulous lifecycle emissions standard. To that end, safe harbors and unsafe harbors can provide a zone of rule-like certainty.[174] Safe harbors are common in tax law. For example, the test for whether tax law should treat a partnership as publicly traded is a facts-and-circumstances test that asks whether partners are able to buy, sell, or exchange their partnership interests in a manner that is economically comparable to trading on an established securities market.[175] But within that standard are two zones of relative certainty: a safe harbor based on easily verifiable measures of lack of trading (in which case the partnership is deemed not publicly traded),[176] and an unsafe harbor if partnership interests trade on an established securities market (in which case it is deemed publicly traded).[177] Likewise, inside the tax rule that excludes a portion of gain from the sale of a home when sold due to a change in one’s place of employment, there is a safe harbor that guarantees a sale will be deemed to be by reason of a change in place of employment if the “individual’s new place of employment is at least 50 miles farther from the residence sold or exchanged than was the former place of employment.”[178] Within 50 miles, a facts-and-circumstances test would apply.

As Susan Morse observes, the main effect of a safe harbor is to encourage behavior to converge on the boundary line just inside the safe harbor.[179] Those whose behavior would be noncompliant move into the safe harbor to become compliant and avoid regulatory penalties, while those who were already compliant move to the outer edge of the safe harbor (assuming that being “too” compliant is costlier than just barely complying).[180] This dynamic raises the costs of regulators defining an overly permissive safe harbor: Not only will firms that shouldn’t have received regulatory protection benefit from it, but firms that otherwise would have performed better than the safe harbor will worsen their behavior up to its boundary.[181] In the next Part, I examine how the performance standards created to ease the administration of carbon accounting risk encouraging firms into carbon shelters.

III. The Problem of Carbon Shelters

What could go wrong with universal carrots? The most likely answer is that they will become too universal and enrich high-carbon firms. At best, that would mean wasting some of the money appropriated for the green transition on a counterproductive use. At worst, it would mean subsidizing emissions that outweigh the emissions reductions purchased with other carrots. The worst-case scenario would elevate the critique of universal carrots from futility to perversity.[182] In this Part, I describe the carbon shelters that lawmakers should aim to avoid: transactions that win green-subsidy money despite generating higher emissions than the law intends to allow.

Carbon shelters arise due to the tradeoff between accuracy and administrability in carbon accounting. As explained in the previous Part, carbon accounting is susceptible to gaming because it is a form of performance-based regulation. In the first instance, carbon shelters form because it is difficult for regulators to estimate the attributional and consequential emissions of each taxpayer’s bespoke submission. Regulators respond by crafting safe harbor rules to ease the administrative challenge. However, these rules often create additional, legally authorized carbon shelters.

I argue that regulators should err on the side of caution in their pursuit of administrability. This means disqualifying applicants who potentially have high emissions, even though some members of this category will turn out to have acceptably low emissions. Greater risks of subsidizing net-positive emissions justify greater caution. Because it is generally cheaper for firms to produce energy with higher emissions than lower, the availability of a high-emissions way to qualify for carrots threatens to swamp the attractiveness of the low-emissions options. And because most green tax credits are uncapped, there is no limit to the number of high-emission projects that could be inadvertently subsidized.

A. From Tax Shelters to Carbon Shelters

The bogeyman of tax law is the reviled tax shelter. A suitable working definition of a tax shelter is a tax-motivated transaction that produces a tax loss in excess of any economic loss, in a manner inconsistent with legislative intent or purpose.[183] A tax shelter is an exploitable loophole—a way for a taxpayer to have their cake (reducing tax liability) and eat it too (without reducing economic income). In the realm of carbon accounting, a comparable problem is likely to emerge: carbon shelters. I define a carbon shelter as a transaction where a taxpayer receives a green subsidy despite producing carbon emissions in excess of that supposedly allowed by the subsidy.

To illustrate, consider one of the largest possible carbon shelters lurking inside Congress’s green subsidies: the risk of the Clean Hydrogen Credit subsidizing hydrogen electrolysis powered by a fossil-dominated grid.[184] One study estimates that without proper additionality requirements, the credit could fund tens to hundreds of millions of tons of new GHG emissions annually at a cost of $30 billion annually to the federal government.[185] From the taxpayer’s perspective, a carbon shelter like this is simply a shortcut to getting paid by the government without having to do the hard work of reducing emissions. That is, a carbon shelter is desirable for the taxpayer on the assumption that it is cheaper to produce higher emissions rather than lower. From the government’s perspective, a carbon shelter is an own goal: paying taxpayers to do something socially undesirable.

At a high level, taxpayers can form carbon shelters by violating any of the carbon accounting principles introduced in Part II. Subsidy seekers can shelter emissions by outsourcing indirect emissions to parties outside the scope of the subsidy (undermining attributional accounting), attributing emissions to byproducts and waste products (same), or by offsetting emissions with low-carbon inputs that would have been produced anyway (undermining consequential accounting). In theory, attributional and consequential lifecycle assessment are meant to catch these failure modes. But lifecycle assessment is imperfect, and some emissions fall through the cracks. Moreover, when regulators erect overinclusive safe harbors, they give taxpayers explicit permission to produce emissions that violate the green subsidy principles. Every safe harbor provides a corresponding carbon shelter.

It is important to recognize some key differences between the concepts of tax shelters and carbon shelters. First, while there is a widespread view that taxpayer motive or purpose should be one of the primary indicia of a tax shelter, my use of “carbon shelter” is not limited to transactions entered exclusively for the purpose of obtaining subsidies.[186] Tax shelters generally involve claiming tax losses, and the steps required to generate tax losses tend to have little appeal outside that context. Conversely, carbon shelters generally involve producing energy in a high-emission manner, which is a common business practice with or without the allure of subsidies. Therefore, unlike with tax shelters, it is not useful to analyze the taxpayer’s purpose or intent in determining whether a transaction is a carbon shelter.[187] Taxpayers might rationally pursue the same high-emission practices with or without the possibility of green-subsidy incentives, even though green subsidies can make those practices more profitable. Most carbon shelters reflect a policy failure, not any rule-bending fault on the part of the taxpayers who benefit.

The second important difference is that, unlike tax shelters, carbon shelters are, in general, not the result of ambiguous and exploitable statutory language. Tax scholars have long analyzed tax shelters as problems caused by formalist statutory interpretation, where taxpayers seize on ambiguous language to achieve outcomes that Congress did not intend.[188] When the problem is understood that way, the solution is to invent canons of pragmatic interpretation to recover Congress’s intent.[189] For example, one of the most important weapons in Congress and the IRS’s fight against tax shelters has been the “economic substance” doctrine, an influential anti-abuse principle that allows the IRS to disallow transactions that do not change the taxpayer’s economic position or lack a substantial nontax purpose.[190] To be sure, Congress has created carbon shelters through statutory ambiguity as well. In the “black liquor” scandal, paper firms claimed the alternative fuels mixture tax credit, which provided a subsidy for mixing in biofuels to petroleum fuels by adding petroleum fuel to black liquor, a wood byproduct they had historically used for fuel.[191] That is, the paper firms seized on the ambiguity of the term “mixture” and got paid to make their existing biomass fuel more carbon-intensive—the exact opposite of the type of mixture Congress meant to encourage.[192]

But that scandal is not the sort we should expect in the new era of a zero-emissions performance standard. The paper firms’ gambit could not plausibly meet a standard like “greenhouse gas emissions rate not greater than zero” or the similar language found in all technology-neutral green subsidies—to the credit of the performance-based approach. Congress has now provided language that justifies the most rigorous possible carbon accounting, with the exception of certain carveouts like the exclusion of indirect land use emissions under the Clean Fuel Credit.[193] Instead, for the most part, the current generation of carbon shelters will be made possible by overly permissive administration (either in granting safe harbors or in failing to apply the underlying performance standard).

Even if the new generation of carbon shelters do not arise from ambiguous statutory language, Congress might still have an important role to play in closing them. In tax law, one of the most powerful approaches to attacking tax shelters has been the enactment of substantive default rules ruling out broad categories of behavior that, while not necessarily abusive in every instance, are likely to attract shelters.[194] The prototypical example would be Section 469, the “passive loss rules,” enacted under the Tax Reform Act of 1986.[195] Those rules stop taxpayers from claiming losses from business activities in which they do not materially participate, a strategy that was a dominant feature of tax shelters in the 1970s and 1980s.[196] In a subsequent era, Marvin Chirelstein and Lawrence Zelenak proposed a “silver bullet” rule disallowing deductions in excess of a measurable reduction in the taxpayer’s net worth or that involve the allocation of noneconomic income to a tax-indifferent party.[197] Whether or not that was the right rule for its day, the structure of the proposal is notable. Chirelstein and Zelenak described the contours of tax shelter transactions and proposed a deliberately broad, overinclusive rule to exclude such transactions.

In the next Section, I will examine the subset of carbon shelters that arise due to overinclusive safe harbor rules. Then, in Part III.C, I will draw on the Chirelstein-Zelenak example to develop a broad anti-abuse rule for carbon shelters.

B. Shelters in Safe Harbors

As discussed at the end of Part II, some carbon shelters are explicitly granted to taxpayers in the form of safe harbor rules. These rules give permission to not conduct a thorough and rigorous carbon accounting. The more a green-subsidy law relies on safe harbors, the greater the risk of a divergence between true carbon emissions and the emissions counted by law. If the safe harbors are overinclusive, they will allow activities that produce higher-than-desired emissions. The implications of being underinclusive are less damaging because activities not covered by the safe harbor will still be evaluated according to the underlying performance standard.

1.   Technology Safe Harbors

The Treasury and IRS often create safe harbors meant to provide a regulatory fast track for energy production technologies considered to be presumptively low emission. An important production safe harbor appears in the Clean Electricity Credits, where a set of electricity generation modalities are labeled as “Non-Combustion & Gasification” (Non-C&G) and treated as categorically having a GHG emissions rate of zero.[198] These include wind, solar, hydropower, geothermal, and nuclear fission. It is easy to understand why the Treasury and IRS would not want to spend resources measuring modest differences in the emissions intensity of these technologies, which are generally understood to be very low emission compared to most combustion and gasification alternatives.

But there are potentially significant differences in the emissions intensity of these technologies. In the case of nuclear fission, large amounts of electricity may be used to enrich uranium. The carbon emissions associated with that input fuel should be counted just as the emissions associated with converting water into hydrogen are counted under the Clean Hydrogen Credit.[199]

Likewise, the operation of a hydroelectric dam can yield “reservoir emissions,” which occur when the flooding of a dam area submerges organic matter. The matter decomposes, releases methane, and emits “degassing emissions” when methane-rich deep water is pulled through the turbine.[200] The Clean Electricity Credits preamble acknowledges but disclaims responsibility for these emissions on the grounds that they “are not created by the fundamental transformation of potential energy in flowing water into electricity.”[201] This argument seems especially weak with respect to degassing emissions, which occur when flowing water powers a turbine. The argument could be more reasonable if the creation and maintenance of a reservoir is treated as separate from the operation of the attached power plant. But as advocates point out, taxpayers may claim the ITC on the value of the entire hydroelectric complex, including the reservoir, and therefore should have to count reservoir-related emissions.[202]

In both the nuclear and hydroelectric cases, it seems that a desire for simplicity undercounts relevant emissions (neither of which are unusually difficult to measure). There may be good reason to treat emissions below some de minimis threshold as equivalent to zero emissions for subsidy eligibility purposes, but one cannot apply a de minimis rule without measuring emissions in the first place. To avoid creating a carbon shelter, the regulations should require nuclear and hydroelectric plants to measure their emissions and submit LCA documentation just as combustion and gasification plants must.

Even less justifiable is Congress’s decision to exempt one additional technology, fuel cells, from the carbon accounting procedures of the Clean Electricity Credits entirely. A fuel cell is a machine that converts the chemical energy of a fuel source into electricity. In the OBBBA, Congress specified that fuel cells’ eligibility should be determined “without regard to” a greenhouse gas emissions-rate analysis.[203] Most fuel cells are powered by hydrogen, which raises the same questions about the emissions intensity of upstream hydrogen production discussed throughout this Article.[204] For all the issues with measurement of emissions induced by hydrogen production discussed in this Article, at least the 45V hydrogen regulations try to measure emissions. The credit for fuel cells effectively provides a backdoor hydrogen subsidy that circumvents the Clean Hydrogen Credit’s entire enforcement apparatus.

2. Temporal Safe Harbors

The Treasury and IRS also create safe harbors to simplify their responsibilities in overseeing subsidized projects over long periods of time. The tax authorities sometimes provide that after a certain duration has passed, applicants are no longer responsible for maintaining the low GHG emissions rates which initially qualified them for subsidies.[205] A temporal safe harbor is particularly relevant in case of the ITC, which is awarded when property is placed in service but has not yet begun operating. Before the start of operations, there is no observed GHG rate for regulators to measure; there is only an “anticipated GHG emissions rate.”[206]

There are three main conceptual issues in holding applicants to a performance standard based on future behavior. First, what indicia observable at time t=0 are sufficient to establish that the anticipated emissions rate will be not greater than zero? Second, for how long must the facility’s emissions rate be anticipated to remain at or below zero? Third, assuming that the tax authorities will continue monitoring the facility’s emissions in the future, for how long will the facility be subject to credit recapture for violating its anticipated GHG rate?

The Clean Electricity Credits Rule answers the first question with an open-ended performance standard and the second and third questions with safe harbors. The performance standard allows the IRS to consider all facts and circumstances to determine an anticipated emissions rate. These include physical characteristics of the facility and the existence of long-term contracts to purchase low-carbon fuels or to capture and dispose of carbon.[207] But the performance standard is subject to a safe harbor.

For facilities that use feedstock fuels or carbon capture, the facts and circumstances must indicate that the facility will operate with net zero emissions for ten years.[208] Anticipated increases in emissions after ten years do not count against the anticipated emissions rate. While the rule seems to cover ten years of emissions, the provision for recapturing credits when recipients exceed the allowable emissions rate only operates for five years after the facility is placed in service.[209] The five-year recapture window allows an applicant to qualify for the ITC on the basis of a ten-year contract to purchase low-carbon fuels but then cancel the contract after five years and suffer no penalty.

Awarding subsidies based on an anticipated emissions rate is necessarily an uncertain business, and it would not be reasonable to ask the government to predict emissions much further than ten years into the future. In that light, the first safe harbor discussed above is appropriate. But a safe harbor based on anticipated behavior over ten years is not worth the name if the subsidy recipient is free to change plans after five years. The solution is a recapture period that matches the duration over which the subsidy is meant to regulate behavior: ten years. No safe harbor (in the form of freedom from recapture) should be granted for behavior between years five and ten.

3.   Model Safe Harbors

Model safe harbors are created by the erroneous background assumptions of government-provided emissions models. For most green tax subsidies, the IRS and the Treasury have directed taxpayers to use specialized LCA models created by the DOE; in other words, the regulations incorporate the rules embedded in the models.[210] These models are customized versions of the GREET model, the original lifecycle model developed by the DOE to measure the emissions impact of transportation fuels. In response to the Inflation Reduction Act, the DOE developed 40BSAF-GREET, a specialized model for aviation fuel, and 45VH2-GREET, a specialized model for clean hydrogen. The process of developing an official DOE model for the electricity sector is still underway (this will be a significant departure from prior GREET models, all of which dealt with fuels).

These models work by combining user-inputted data with background assumptions and formulas provided by the DOE. The user inputs characteristics of its production process such as the type of feedstock it is using to produce hydrogen, the quantity of power inputs used in the production process, and the quantity of output yielded (this takes place in an Excel spreadsheet).[211] Using DOE formulas, the model combines these parameters with background assumptions about the carbon intensity of specific production processes in order to output Scope 1, Scope 2, and Scope 3 emissions values. The results depend enormously on the background assumptions.

In particular, the GREET models face a tension between allowing users to input parameters specific to their own operations and fixing parameters in line with what the DOE believes to be nationally representative or, more conservatively, in line with a worst-case scenario. For example, the hydrogen GREET model includes a parameter for the upstream methane leak rate associated with processing and transporting natural gas, because natural gas can be used to power the electrolyzers that produce hydrogen. The higher the leak rate, the higher the carbon intensity of hydrogen production. The current version of 45VH2-GREET uses a fixed nationwide leak rate of 0.9 percent despite evidence that the leak rate varies dramatically by region and producer.[212]

Using the fixed nationwide leak rate means that a project with an abnormally high rate could qualify for the same subsidy as a project with a very low one.[213] To combat this, industry representatives have generally advocated for using site-specific emissions intensity data out of a belief that the nationwide averages tend to lag behind improvements producers make on the ground.[214] Allowing taxpayers to submit their own measurements on a variety of technical parameters, however, would mean that the IRS and DOE would have to audit and verify these numbers—a difficult administrative challenge. In the preamble to the final Clean Hydrogen Credit Rule, the Treasury and the IRS indicated that a future version of 45VH2-GREET would require taxpayers to use their own methane leak rate.[215]

For the sake of administrative simplicity, the Treasury and IRS may find it necessary to keep using fixed parameters in LCA models. In that case, they should nonetheless question whether the nationwide average of a given measure is the right value to fix. If one is risk-averse, the parameters should instead be set at the worst-case option among plausible values. This would set up the model so that worst-case users receive no safe harbor in qualifying for the subsidy.

4. Safe Harbors from Emissions Induced Elsewhere in the Economy

A final category of safe harbor exists in the rules meant to ensure that subsidy recipients do not induce emissions elsewhere in the economy. As introduced in Part II.A.2, the additionality principle requires ensuring that subsidies generate new clean energy rather than moving around clean energy sources that would have been used anyway. The Treasury and IRS’s attempted solution has been to establish proxies for additionality; however, these proxies function as safe harbors because they do not actually guarantee additionality.

In the case of the Clean Hydrogen Credit, the Treasury created a safe harbor that can be accessed by low, if not necessarily zero, market-mediated emissions.[216] The most likely reason for this choice is that attempting to model the marginal grid emissions associated with running each electrolyzer would be difficult to administer.[217] Instead of modeling, the Treasury’s regulations require hydrogen producers prove that their electricity supply satisfies “three pillars.” The electricity supply should be (1) procured from clean power plants in the same region, (2) procured from clean power plants built approximately contemporaneously or after the hydrogen plant, and (3) procured during the specific hours those clean plants are operating.[218] Together, the pillars are meant to ensure that hydrogen electrolysis uses clean electricity that other users did not previously rely on.

Still, despite the three pillars, hydrogen electrolysis can still induce additional grid emissions.[219] Buying power from a new, renewable power plant does not mean that no one else might have wanted to purchase that same clean power at the same moment. If renewables were scarce at that moment, the hydrogen plant using those resources may have pushed others to use fossil resources.[220]

Moreover, the Treasury and IRS offered additional safe harbors in the final clean hydrogen rule. Electricity will count as “incremental” (i.e., the second pillar above) if procured in a state where state law makes induced-grid emissions unlikely, as reflected in a legally binding state GHG emissions cap, or if procured from a nuclear reactor at risk of retirement.[221] These exceptions to the general incrementality rule are meant to reflect circumstances where induced emissions are relatively unlikely but do not guarantee that result. For example, electricity demand is projected to grow over the next decade. A nuclear reactor that was financially struggling in the period 2017–2021 (the period referenced by the rule) might nonetheless have multiple new sources of demand such that selling nuclear-generated electricity to a subsidized hydrogen project might indeed induce emissions elsewhere on the grid because other would-be consumers of the nuclear plant’s energy would need to use fossil fuels.[222]

When it comes to the Clean Electricity Credits, the concern is that biofuel feedstocks that are combusted to generate electricity may induce emissions elsewhere in the economy. Because biofuels and fossil fuels are substitutes, subsidies that increase demand for and reduce the price of biofuels will cause the price of fossil fuels to decline, which could cause more overall energy consumption (including more fossil energy consumption). If the prices of fossil fuels drop far enough, those fuels could displace consumption of renewable energy.[223] Likewise, if biogas is diverted from existing economic uses to produce electricity, other users would need to produce more biogas to replace it.

The Treasury and IRS’s current approach to emissions induced by biogas creates another safe harbor. In the Proposed Clean Electricity Credit rules, the Treasury and IRS contemplated requiring that biogas used to generate electricity originate from the “first productive use” of its biogas source.[224] Of course, there is a limit to this logic: If demand for biogas is increasing economy-wide, the fact that a source of biogas is new, with no prior productive use, does not at all prove that it would not have been used for some other purpose.[225]

In the final rule, the Treasury and IRS dropped the “first productive use” idea and instead specified acceptable “alternative fates” to assign to the biogas, depending on its source (e.g., methane from landfills vs. wastewater plants vs. animal waste).[226] “Alternative fates” is the Treasury’s term for counterfactual scenarios, specifically the emissions that the biogas source would have caused if it had not been used for electricity. Specifying alternative fates at the industry level—rather than requiring each facility to use the actual alternative fate of its biogas—is yet another safe harbor. Any facility whose actual practice is less carbon intensive than the industry norm (e.g., flaring rather than venting, or selling rather than flaring) gets to claim that they have saved the methane from the more carbon-intensive alternative fate and thereby get more credit for using the biogas for electricity.

The only area where Congress has weighed in on additionality analysis is the Clean Fuel Credit, where, in the OBBBA, Congress forbade the inclusion of emissions caused by land use changes.[227] This rule is a boon for producers of ethanol, palm oil, and other fuels, who will no longer be penalized for the deforestation-related emissions caused by pulling new land into feedstock production.

Rather than rely on proxies for induced emissions, Congress, the Treasury, and the IRS should mandate the use of general equilibrium economic models to compute the market-mediated effects of using clean fuels and clean electricity in credit-eligible activities. These models represent the best technique available for estimating the effect of consuming one resource (e.g., biofuel) on substitutable products.[228] There is precedent for using these models: Existing government models include the EPA’s SAGE model, the Energy Information Administration’s National Energy Modeling System, and the Department of Agriculture’s food system general equilibrium model.[229] If possible, the DOE should incorporate such models into its GREET models to provide an interface for calculating market-mediated emissions effects (currently, GREET only addresses only indirect land use changes and not other indirect emissions). Requiring general equilibrium economic modeling would ensure stringent adherence to the performance standard.

C.   Shelters in Carbon Offsets

Closing overly permissive safe harbors is just the first step in policing carbon shelters. It should be the easier half of the battle, since at least safe harbors are simple to identify, having been explicitly granted by regulators. The harder half of the battle is discovering and foreclosing the carbon shelters that taxpayers invent through creative carbon accounting.

The challenge here—which this Section confronts—is for the IRS to avoid being continually one step behind. For every carbon shelter discovered, a new one will be invented. As we have learned in the tax shelter context, a whack-a-mole strategy of outlawing one specific fact pattern at a time is not viable.[230] The alternative is to come up with a broad anti-abuse rule that can apply to varied fact patterns. The goal should be to strike a middle ground between specifying a laundry list of per se problematic fact patterns (“whack-a-mole”) and a rule so broad that judicial discretion is needed to apply it. In the tax shelter context, Chirelstein and Zelenak’s proposal for a rule against noneconomic deductions was intended to be broader and more prospectively useful than a rule focused on the varied Code sections exploited by thematically similar tax shelters (e.g., installment sales in ACM Partnership and tax-free incorporations in Black & Decker[231]) but narrower and more predictably applied than the “economic substance” doctrine.[232] In this Section, I propose that the most vital anti-abuse rule to protect against carbon shelters is as follows: Firms should not be able to use carbon offsets (as defined below) to reduce their GHG rates.

Just as taxpayers “game” the income tax by generating noneconomic losses (i.e., negative income), firms similarly game carbon accounting by generating misleading negative emissions. Negative emissions, or carbon offsets, serve to reduce a taxpayer’s emission rate, thereby allowing them to use some quantity of positive-emission techniques or inputs and still qualify for a net GHG rate of zero.

There are three main ways that taxpayers might seek to offset emissions. First, they might purchase carbon offset credits (i.e., a legal record representing emissions reductions entirely unrelated to the taxpayer’s subsidy-eligible project). The Clean Electricity Credits Rule disallows the use of offsets unrelated to the production of electricity or of any input fuel. Thus, this first option is currently off the board.[233] Second, taxpayers might aim to take credit for using input fuels with supposedly negative emission rates. The idea behind a negative emission rate is that if some input source had not been used for electricity, it would have generated emissions. Thus, the resulting lack of emissions should be treated as even better than zero emissions, but negative relative to the counterfactual. The use of negative-emission fuels to offset other emissions is endorsed by the Rule, at least for certain sources of biogas like landfills, wastewater, and farm animal waste.[234] Third, taxpayers might conduct negative emission activities, like carbon capture, at their own facility. Carbon capture is endorsed by the Rule as a legitimate way of reducing the GHG rate.[235]

At a high level, there are two potential objections to giving subsidies for offsets. First, that giving subsidies for offsets risks violating the additionality principle. Second, that even if it satisfies additionality, technology neutrality is meant to reward innovation, and it violates the spirit of green subsidies to reward emissions reductions that don’t come from more efficient energy generation technology.

1.   Offsets Can Violate Additionality

First, consider the additionality issue. Vast literatures in environmental science and law have documented the pervasive problem of carbon offsets overpromising and underdelivering true additional emissions reductions.[236] The exact reasons differ for each type of offset.

The use of biogas as a fuel source for electricity generation risks becoming the next important example of overpromising and underdelivering. Biogas is produced by the decomposition of organic matter from landfills, livestock operations, and wastewater treatment.[237] Each of these processes has the potential to leak methane (a greenhouse gas) into the atmosphere. Nonetheless, biogas is often assigned negative emission rates based on the assumption that the methane would have escaped into the atmosphere if it had not been used as fuel.[238] There are at least two potential flaws in this logic: that the assumption that the methane would have escaped is unfounded and the negative emission rate therefore overstated;[239] and that because methane is delivered via natural gas pipelines, it is hard to know how much methane (as opposed to gas) a taxpayer is actually combusting.[240]

Since the point of claiming credit for negative emissions is to offset one’s positive emissions, letting taxpayers claim emissions below zero will allow those taxpayers to produce positive emissions elsewhere in their production process and still receive a tax credit for balancing out to zero emissions on net. For example, analysts have calculated that using just a 25 percent share of methane alongside 75 percent fossil gas could be treated as net zero, due to the extreme negative emissions value attributed to some methane.[241] If the logic that justified the negative emissions turns out to be faulty, taxpayers would get a subsidy for producing energy using (mostly) natural gas.

We can now reevaluate the first two distinct methods of offsetting emissions in light of the additionality issue. Both unrelated offsets and offsets pertaining to input fuels like biomethane are susceptible to failure because of the risk of faulty additionality logic. Even worse, because offsets justify the inclusion of fossil fuels in a purportedly net-zero energy system, even a small error in additionality can lead to subsidizing net-positive emissions. The upshot of this realization is that blending negative emission fuels to a facility’s tested emissions rate should be forbidden. Stated differently, the minimum emissions rate that should be assigned to a putatively negative emissions feedstock like biomethane should be zero (never negative). This would mean that biomethane could not be blended with any positive-emissions fuel and still achieve net zero.[242] Indeed, the most valuable carbon accounting provision of the One Big Beautiful Bill Act was an update to the Clean Fuel Credit such that “the emissions rate for a transportation fuel may not be less than zero.”[243] The same change was not made for the Clean Electricity Credits.[244]

2.   Offsets Aren’t Innovation Policy

But even supposing for the sake of argument that offsets are truly additional, there is a further question of whether it is appropriate to credit them against a subsidy seeker’s emissions rate. This is not a climate science question but a policy one: To what extent does it achieve the goals of green-subsidy laws to generate emissions reductions disconnected from the subsidized activity?

Here, there is an analogous issue within tax law. When tax outcomes depend on whether taxpayers fall above or below a minimum effective tax rate, like under U.S. law’s Global Intangible Low-Taxed Income (GILTI) or the OECD’s proposed Pillar 2, there is a question of whether firms should be allowed to combine income earned and taxes paid across multiple countries to calculate the effective tax rate.[245] Under current federal law, firms are able to aggregate income from multiple countries, effectively using excess foreign tax credits from high-tax countries to soak up potential tax liability from low-tax countries.[246] If the GILTI tax were applied on a country-by-country basis, i.e., without aggregating, income earned in a low-tax country would be subject to tax, which is arguably the purpose of a minimum-tax policy like GILTI.[247] The policy purpose of GILTI is for firms to face an incentive to leave low-taxed jurisdictions. Instead, the current rule means that firms can satisfy GILTI by aggregating themselves into a corporate structure with activities in high-tax jurisdictions.[248]

Similarly, if the purpose of a green subsidy is to improve the emissions efficiency of individual power plants, electrolyzers, and fuel producers, linking together those industrial processes with unrelated emissions reductions does not advance the policy goal. This is clearest if one views the green subsidies as innovation policy, with a purpose of catalyzing improvement in power and fuel production.[249] That is a different goal than merely paying for net-zero energy, and indeed there are other policies—including the Carbon Sequestration Credit—directly targeted at achieving negative emissions. Notably, the credit pays less per ton of sequestered carbon than the Clean Hydrogen Credit for an equivalent emissions reduction. This suggests that Congress views offsets as less valuable than energy production efficiencies.[250]

We can now reevaluate carbon capture as a variant of the negative emissions problem. On the one hand, if a firm captures and sequesters carbon generated at their own facility, there is less of a concern about the process failing to satisfy net zero and thus implicating the additionality concern.[251] From an innovation policy perspective, however, giving out green subsidies based on carbon capture blunts the primary purpose of those subsidies.

Lastly, and fundamentally, this is an argument for preserving distinct policy instruments, not an argument against carbon capture on the merits. Preserving distinct policy instruments is important because it is uncertain which technological course (better carbon capture vs. more efficient renewable generation) will prove more valuable in the long run. Offsetting collapses such distinctions. It also turns targeted policies into more diffuse ones. In that light, a broad anti-abuse rule could state that carbon captured in the process of electricity generation should not be allowed to offset that facility’s GHG emissions rate for the purpose of the Clean Electricity Credits, nor should carbon captured while producing input fuels that feed electricity generation count against the emissions rate.[252]

D.   Asymmetric Risk Justifies Anti-Abuse Rules

The previous Section illustrated the logic of developing an anti-abuse rule to prevent carbon shelters. A defining feature of that logic was caution: Because it is impossible to anticipate each carbon shelter fact pattern in advance, the rule must extend across a broad range of fact patterns, not all of which will turn out to be shelters.

The downside of such an anti-abuse rule is that it will be overinclusive, punishing activities that should have been allowed under careful application of the underlying performance standard.[253] In the tax context, that would mean imposing tax on transactions that Congress meant to grant as losses, even if noneconomic.[254] Similarly, Congress requires taxpayers to disclose their participation in “reportable transactions”: transactions that it considers likely to be abusive but are not necessarily abusive in every case.[255]

In the carbon accounting context, using overbroad prohibitions would mean denying green subsidies to applicants whose production process might, on close inspection, turn out to be zero emission.[256] That result contradicts the aspiration to technology neutrality. If, in order to win a subsidy, the taxpayer must not only achieve zero emissions but also avoid guilt by association with a questionable practice like carbon offsetting, they no longer enjoy the flexibility of technology neutrality. The question is whether it is appropriate to sacrifice the absolutist version of technology neutrality in order to prevent carbon shelters.

The precise tradeoff is between the harm of denying subsidies to deserving recipients against the benefit of preventing subsidies from going to undeserving recipients. Framed this way, the question is whether false positive or “type I” errors (a good project is not subsidized) are likely to be more consequential than false negative or “type II” errors (a bad project is subsidized).[257] If one were agnostic about the relative prevalence and magnitude of type I and type II errors—meaning that the IRS would err an equal amount in either direction—then one would proceed to apply the performance standard with no need for special caution.[258] But the nature of carbon accounting suggests that the distribution of errors would probably not be symmetrical.

The brief history of carbon accounting suggests an ongoing pattern where modelers progressively learn of new causal pathways between energy systems and their consequential emissions.[259] As recounted in Part II.A, a previous era of carbon accounting did not recognize emissions resulting from indirect land use changes, but contemporary practice does (even when Congress, for unknown reasons, refuses to count those emissions). Presumably there will be more revelations in the years ahead.

This pattern is especially likely to repeat when a new energy source enters the equation, as is intended to happen under technology neutrality. Based on experience, the first attempt to model a new energy source rarely uncovers all consequential emissions.[260] Of course, it is theoretically possible that advances in carbon accounting will instead reveal that consequential emissions of energy systems are actually lower than modelers currently think. But this is not very plausible as applied to new energy sources, which will be chosen on the basis of seeming to have low or zero emissions and will be treated as such until proven otherwise. The IRS will therefore tend to err on the side of underestimating emissions for each new energy technology before iteratively correcting such errors.

It is nonetheless possible that even if administrative errors are more common in the direction of underestimating emissions, it might still be desirable to be neutral between type I and type II errors. If the magnitude of the emissions reductions gained by approving more good subsidies (i.e., minimizing type I errors) is greater than the emissions improperly subsidized through type II errors, the government would still come out ahead. Which way the scales are likely to balance depends on the exact unsafe harbor rule under consideration.

But, in general, the magnitude of type II errors is likely to be greater for two reasons. First, high-emission firms will tend to have a cost advantage over low-emission firms, because reducing emissions tends to be costly.[261] Thus, if subsidies are made available for both, the high-emission firms will maintain their advantage in product markets and capital markets. For that reason, it should be the goal of green-subsidy policy to create advantages for the low-emission firm, not to give equal carrots to all.[262] Second, the case for overcoming bad subsidies with good ones is especially weak if we suppose that there is a limit on how much can be spent on green subsidies. If there is only (say) $100 billion to spend on green subsidies, one should not waste any of that budget on projects with positive (or higher than allowable) emissions.[263] Instead, one should try to spend the entire budget on projects that are all zero emission, assuming such projects are available. In other words, a budget constraint increases the opportunity cost of committing type II errors.

Choosing deliberately overbroad anti-abuse rules like the rule against offsets proposed above necessarily sacrifices the absolutist version of technology neutrality. For example, if a taxpayer is not allowed to claim negative emissions for using captured methane, a power plant fueled by a blend of biogas and natural gas will not be eligible for tax credits because the blended emissions rate will not reach zero.[264] This would be understandably frustrating to the developers of a blended biogas plant confident that their counterfactual emissions estimate is correct and that their plant truly achieves net zero. But a responsible version of technology neutrality does not require the view that error costs are symmetrical between false positives and false negatives. Instead, technology neutrality should be compatible with a high standard of evidence of the sort embodied in a cautious anti-abuse rule. An openness to innovation, rather than blind permissiveness to risky projects, is the core of technology neutrality that energy tax law should preserve.

Conclusion

The principal appeal of universal energy policies is that they are supposed to be “technology-neutral,” or indifferent between any two technologies that reduce carbon emissions by the same degree. The only way to achieve technology neutrality is to estimate the carbon intensity of every project potentially eligible for carrots. This is the role of carbon accounting, the discipline of assigning carbon emissions to responsible parties. While existing scholarship treats carbon accounting as a problem for environmental law, it is instead tax law where carbon accounting faces its highest-stakes challenge yet. This Article argues that carbon accounting is surprisingly well suited to tax law because the normative and analytical principles behind carbon accounting bear close resemblance to principles within tax law. Concepts from tax theory including a comprehensive tax base, additionality, liability shifting, and rate blending can usefully inform the development of parallel doctrines within emissions accounting.

And yet, just as the income tax is susceptible to “tax shelters” that technically comply with the rules but substantively avoid tax, so too will firms develop “carbon shelters” that qualify for green subsidies while covertly making use of high-emission energy. This Article identifies the features of emerging carbon accounting law that facilitate carbon shelters and proposes reforms to close them down. Because of the difficulty of anticipating every carbon shelter in advance, an antishelter strategy needs deliberately overbroad prophylactic rules, including some modeled on similar rules from tax law, like the passive loss rules. If policymakers are to avoid inadvertently subsidizing unlimited emissions, they must be prepared to sacrifice some of the technology neutrality that motivates universal carrots in the first place. Technology neutrality is a means of reducing emissions; it should not be treated as an end in itself when it is no longer serving that goal.

‍ ‍

‍Copyright © 2026 Jeff Gordon, Assistant Professor, Vanderbilt Law School. I am grateful to Alice Abreu, Anne Alstott, Johnny Buckles, David Carel, Brian Galle, Zach Liscow, Joshua Macey, Daniel Markovits, Alan Mitchell, Morgan Ricks, Jim Rossi, Peter Salib, Jeff Schoenblum, Amy Sinden, Michael Vandenbergh, Larry Zelenak, Jon Zytnick, and participants at the Georgetown Tax Law and Public Finance Workshop and Northwestern Tax Law Colloquium for helpful comments and conversations. I thank the talented editors of the California Law Review for their hard work which has greatly improved the Article.

[1].     See, e.g., Brian Galle, The Tragedy of the Carrots: Economics and Politics in the Choice of Price Instruments, 64 Stan. L. Rev. 797, 812 (2012); see also Gerrit De Geest & Giuseppe Dari-Mattiacci, The Rise of Carrots and the Decline of Sticks, 80 U. Chi. L. Rev. 341, 354–55 (2013) (on the distinction between carrots and sticks).

[2].     See Gilbert E. Metcalf, Tax Policies for Low-Carbon Technologies, 62 Nat’l Tax J. 519, 523 (2009) (“An efficient energy policy should not favor one energy source over another after taking into account any positive or negative externalities associated with its production or consumption.”).

[3].     See New Source Performance Standards for Greenhouse Gas Emissions from New, Modified, and Reconstructed Fossil Fuel-Fired Electric Generating Units, 89 Fed. Reg. 39798 (May 9, 2024) (to be codified at 40 C.F.R. pt. 60).

[4].     In practice, a carbon tax might only be imposed on upstream fossil fuel producers and refiners and still capture the 80 percent of U.S. emissions corresponding to fossil fuel combustion. Gilbert E. Metcalf & David Weisbach, The Design of a Carbon Tax, 33 Harv. Env’t L. Rev. 499, 501 (2009).

[5].     See Colton Poore, Without Guidance, Inflation Reduction Act Tax Credit May Do More Harm Than Good, Princeton Eng’g (Dec. 20, 2022), https://engineering.princeton.edu/news/2022/12/20/without-guidance-inflation-reduction-act-tax-credit-may-do-more-harm-good [https://perma.cc/X8CM-NPDM] (“[W]ithout careful implementation, the credit could backfire by inadvertently increasing nationwide carbon pollution.”).

[6].     See Arnaud Brohé, The Handbook of Carbon Accounting 25–26 (2016).

[7].     A number of articles discuss GHG emissions accounting in relation to specific categories of emissions. See generally Alexandra B. Klass & Andrew Heiring, Life Cycle Analysis and Transportation Energy, 82 Brook. L. Rev. 485 (2017) (discussing ethanol and transportation sector emissions accounting); Daniel A. Farber, Indirect Land Use Change, Uncertainty, and Biofuels Policy, 2011 U. Ill. L. Rev. 381 (discussing indirect land use effects of biofuels); Taotao Yue & Marjan Peeters, Better Greenhouse Gas Emissions Accounting for Biofuels: A Key to Biofuels Sustainability, 6 Climate L. 279 (2016) (discussing biofuels emissions accounting generally); Jonathan Rosenbloom, Outsourced Emissions: Why Local Governments Should Track and Measure Consumption-Based Greenhouse Gases, 92 U. Colo. L. Rev. 451 (2021) (arguing for municipal GHG accounting); Clint Wallace & Shelley Welton, Taxing Luxury Emissions, 109 Corn. L. Rev. 1153 (2024) (arguing for GHG accounting of consumption-based emissions); Lynn M. LoPucki, Corporate Greenhouse Gas Disclosures, 56 U.C. Davis L. Rev. 405 (2022) (cataloguing emissions accounting protocols to which corporations are subject); Madison Condon, What’s Scope 3 Good For?, 56 U.C. Davis L. Rev. 1921 (2023) (discussing the significance of “Scope 3” or supply chain emissions to the SEC’s climate disclosure rule).

[8].     See Shelley Welton, Neutralizing the Atmosphere, 132 Yale L.J. 171, 195–98, 199, 202–07 (2022) (discussing accounting risks that might undermine claims of net-zero). See generally Albert C. Lin, Fixing Net Zero Leakage, 58 Wake Forest L. Rev. 119 (2023) (proposing strategies to combat emissions leakage undermining corporate net-zero pledges, including greater accuracy in carbon accounting); Daniel C. Esty & Nathan de Arriba-Sellier, Zeroing in on Net-Zero: From Soft Law to Hard Law in Corporate Climate Change Pledges, 94 U. Colo. L. Rev. 635 (2023) (proposing principles for evaluating and disciplining corporate net-zero pledges).

[9].     Leehi Yona, Emissions Omissions: Greenhouse Gas Accounting Gaps,49Harv. Env’t L. Rev. 597, 638–39 (2025).

[10].     See, e.g., id. at 647–54 (describing implications for environmental law). But see Condon, supra note 7, at 1928–39 (showing how emissions accounting has become a critical element of securities law).

[11].     David A. Weisbach & Jacob Nussim, The Integration of Tax and Spending Programs, 113 Yale L.J. 955, 996 (2004).

[12].     But see Anne L. Alstott, The Earned Income Tax Credit and the Limitations of Tax-Based Welfare Reform,108 Harv. L. Rev. 533, 535–36 (1995).

[13].     Tax law is a social field in the same way that law more broadly is a social field. See Pierre Bourdieu, The Force of Law: Toward a Sociology of the Juridical Field, 38 Hastings L.J. 814, 828–31 (1987) (describing the socio-professional world of lawyers, or the “juridical space,” as a self-contained site of contestation within distinct internal norms and modes of evaluation).

[14].     See Cary Coglianese, The Limits of Performance-Based Regulation, 50 U. Mich. J.L. Reform 525, 558–61 (2017).

[15].     See infra Part II.C.

[16].     See infra Part III.B.

[17].     See infra Part III.B.

[18].     See infra Part III.D.

[19].     I.R.C. § 469. See generally Joseph Bankman, The Case Against Passive Investments: A Critical Appraisal of the Passive Loss Restrictions, 42 Stan. L. Rev. 15 (1989).

[20].     See infra Part III.C.

[21].     See, e.g., Economists’ Statement on Carbon Dividends, Climate Leadership Council, https://clcouncil.org/economists-statement/ [https://perma.cc/9PE5-JD6S] (reflecting thousands of economists endorsing a carbon tax); see also Gilbert E. Metcalf & David Weisbach, The Design of a Carbon Tax, 33 Harv. Env’t L. Rev. 499, 500 (2009) (acknowledging the scholarly consensus behind carbon taxation and proceeding to technical design issues).

[22].     See generally H. Spencer Banzhaf, A History of Pricing Pollution (Or, Why Pigouvian Taxes Are Not Necessarily Pigouvian) (Nat’l Bureau of Econ. Rsch., Working Paper No. 27683, 2020).

[23].     See generally A.C. Pigou, The Economics of Welfare (1920);Francis M. Bator, The Simple Analytics of Welfare Maximization, 47 Am. Econ. Rev. 22 (1957) (proposing roughly this definition of an externality).

[24].     For a classical exposition of Pigouvian taxation and subsidy, see generally William J. Baumol, On Taxation and the Control of Externalities, 62 Am. Econ. Rev. 307 (1972); Louis Kaplow & Steven Shavell, On the Superiority of Corrective Taxes to Quantity Regulation, 4 Am. L. Econ. Rev. 1, 2 (2002); Kyle D. Logue & Joel Slemrod, Of Coase, Calabresi, and Optimal Tax Liability, 63 Tax L. Rev. 797, 829 (2010) (explaining Pigouvian subsidies); Jonathan Masur & Eric Posner, Toward a Pigouvian State, 164 U. Pa. L. Rev. 93, 96, 104–08 (2015) (“Pigouvian taxes are constantly advocated by economists who seek to influence public policy.”). Some authors spell the word “Pigovian.” The modern theory of pollution charges did not emerge until the 1960s, when economist Allen Kneese studied water pollution and proposed “a system of charges and bounties” in order to set the marginal costs of abatement equal to the marginal damages of foregoing abatement. Allen V. Kneese & Kenneth C. Nobe, The Role of Economic Evaluation in Planning for Water Resource Development, 2 Nat. Res. J. 445, 459 n.16 (1962).

[25].     See William D. Nordhaus, Economic Growth and Climate: The Carbon Dioxide Problem, 67 Am. Econ. Rev. 341, 341 (1977).

[26].     Id. at344.

[27].     Economists’ Statement on Carbon Dividends, supra note 21.

[28].     Jennifer A. Dlouhy, From Greenspan to Yellen, Economic Brain Trust Backs Carbon Tax, Bloomberg (Jan. 16, 2019), https://www.bloomberg.com/news/articles/2019-01-17/from-greenspan-to-yellen-economic-brain-trust-backs-carbon-tax [https://perma.cc/29DA-8MMG]. Nordhaus, who had since won the Nobel Prize for Economics for his pioneering work in climate economics, did not sign the statement. See Economists’ Statement on Carbon Dividends, supra note 21.

[29].     Satyajit Bose, Allison L. Bridges & Kelsie DeFrancia, Columbia Univ., Earth Inst., Carbon Pricing as a Policy Instrument to Decarbonize Economies 14 (2019).

[30].     Id. at 14–15.

[31].     See generally Danny Cullenward & David G. Victor, Making Climate Policy Work (2020) (detailing the successes and shortcomings of cap-and-trade markets).

[32].     See Leah C. Stokes & Matto Mildenberger, The Trouble with Carbon Pricing, Bos. Rev. (Sep. 24, 2020), https://www.bostonreview.net/articles/leah-c-stokes-matto-mildenberger-tk/ [https://perma.cc/97TQ-BMVW] (discussing the political weakness of carbon taxation).

[33].     See Eric Biber, Nina Kelsey & Jonas Meckling, The Political Economy of Decarbonization: A Research Agenda, 82 Brook. L. Rev. 605, 616 (2017) (“Our interpretation of these results is that first, carbon pricing instruments tend to be politically difficult to enact as first steps in jurisdictions dependent on fossil fuel power.”).

[34].     Jonas Meckling, A New Path for U.S. Climate Politics: Choosing Policies That Mobilize Business for Decarbonization, 685 Annals Am. Acad. Pol. & Soc. Sci. 82, 86 (2019).

[35].     See John Kemp, Who Will Pay for the Energy Transition, Reuters (Apr. 27, 2021) (emphasis added), https://www.reuters.com/business/energy/who-will-pay-energy-transition-kemp-2021-04-27/ [https://perma.cc/H876-QGJL] (on the high upfront costs of the energy transition).

[36].     On the problem of fossil fuel assets being locked in due to their long operating life, see Emily Hammond & Jim Rossi, Stranded Costs and Grid Decarbonization, 82 Brook. L. Rev. 645, 652–59 (2017).

[37].     See, e.g., Alan Jenn, Inês L. Azevedo & Pedro Ferreira, The Impact of Federal Incentives on the Adoption of Hybrid Electric Vehicles in the United States, 40 Energy Econ. 936, 941 (2013) (finding that “only sufficiently large incentive amounts” yielded an effect on electric vehicle sales). See generally Eric Lonergan & Corinne Sawers, Supercharge Me: Net Zero Faster 20–22 (2022) (advocating the use of “extreme positive incentives for change”).

[38].     See J.W. Mason, Climate Policy from a Keynesian Point of View, Heinrich Böll Stiftung (Apr. 21, 2022), https://eu.boell.org/en/climate-policy-keynesian [https://perma.cc/KJA2-5SRK] (discussing the choice between carbon pricing and investment-based approaches to decarbonization); Noah M. Sachs, Can We Regulate Our Way to Energy Efficiency? Product Standards as Climate Policy, 65 Vand. L. Rev. 1631, 1659 (2012) (arguing that demand “pull” for rapid technological transformation “is unlikely to be generated by economy-wide cap-and-trade systems or by carbon taxes, unless the carbon price is substantial”); Noah Kaufman, Should the Inflation Reduction Act End the Dream of a National Carbon Price?, Columbia | SIPA: Ctr. on Glob. Energy Pol’y (Jan. 26, 2023), https://www.energypolicy.columbia.edu/should-the-inflation-reduction-act-end-the-dream-of-a-national-carbon-price/ [https://perma.cc/FG72-XWNH]. Other scholars do continue to propose variants of carbon taxes, such as a tax on emissions associated with luxury consumption (e.g., private jets and second homes). Seegenerally Wallace & Welton, supra note 7.

[39].     Lily L. Batchelder, Fred T. Goldberg, Jr. & Peter R. Orszag, Efficiency and Tax Incentives: The Case for Refundable Tax Credits, 59 Stan. L. Rev. 23, 45 (2006) (showing that Pigouvian modification of prices results in the optimal supply of a good). A rare critique of the Pigouvian view comes from Victor Fleischer, who argues that Pigouvian taxes and subsidies fail to target costs and benefits toward those who deserve or need them most. Victor Fleischer, Curb Your Enthusiasm for Pigovian Taxes, 68 Vand. L. Rev. 1673, 1688, 1697 (2015).

[40].     Revenue Act of 1962, Pub. L. No. 87-834, § 2, 76 Stat. 960, 962–73 (1962).

[41].     Thomas W. Giegerich, The Monetization of Business Tax Credits, 12 Fla. Tax Rev. 709, 720–22 (2012).

[42].     Energy Tax Act of 1978, Pub. L. No. 95-618, § 301, 92 Stat. 3174, 3194–201 (1978). In 1986, Congress repealed the generic investment credit but retained the energy version. See S. Rep. No. 99-313, at 275–76 (1986) (noting that even though the oil crisis had abated and fossil fuel prices had steeply declined, giving up on renewable energy at this moment would waste innovation progress from the past decade).

[43].     I.R.C. § 48(l)(2) (1978) (amended 2025). “Alternative energy property” referred to any equipment operated with fuel sources other than oil and natural gas. Id. § 48(l)(3) (1978); see also Energy Tax Act of 1978, Pub. L. No. 95-618, § 301(b), 92 Stat. 3174, 3194–96.“Specially defined energy property” was the most narrowly targeted category, listing eleven specific types of products that could be used to reduce the amount of energy consumed in an industrial or commercial facility, including recuperators, heat pipes, and turbulators. I.R.C. § 48(l)(5) (1978); see also Energy Tax Act of 1978 § 301(b).

[44].     Energy Policy Act of 1992, Pub. L. No. 102-486, § 1914(a), 106 Stat. 2776, 3020–23.

[45].     Id. § 45(a) (current version at I.R.C. § 45(a)).

[46].     Id. § 45(c)(1) (current version at I.R.C. § 45(c)(1)).

[47].     See American Jobs Creation Act of 2004, Pub. L. No. 108-357, § 710(a), 118 Stat. 1418, 1552.

[48].     See generally Austin Brown, Philipp Beiter, Donna Heimiller, Carolyn Davidson, Paul Denholm, Jennifer Melius, Anthony Lopez, Dylan Hettinger, David Mulcahy & Gian Porro, Nat’l Renewable Energy Lab’y, Estimating Renewable Energy Economic Potential in the United States: Methodology and Initial Results 59, 60–61 (2016) (discussing the parameters affecting the economic potential of solar and wind energy, respectively).

[49].     Tax Credits for Electricity Production from Renewable Sources: Hearing Before the Subcomm. on Select Revenue Measures of the H. Comm. on Ways & Means, 109th Cong. 47–48 (2005) (statement of Christopher O’Brien, Vice President, Strategy & Gov’t Rels., Solar Sys. Div., Sharp Elecs., Chairman of the Bd. of Dirs., Solar Energy Indus. Ass’n).

 [50].     The chairman of the solar industry lobbying group testified to Congress in 2005 that “we would willingly be removed from the section 45 credit [PTC] in favor of an expansion of the existing investment credit.” Id. at 48.

[51].     David M. Schizer, Energy Subsidies: Worthy Goals, Competing Priorities, and Flawed Institutional Design, 70 Tax L. Rev. 243, 298 (2017).

[52].     Id. at 280; Jonathan H. Adler, Eyes on a Climate Prize: Rewarding Energy Innovation to Achieve Climate Stabilization, 35 Harv. Env’t. L. Rev. 1, 29–30 (2011) (arguing that the government is ill-equipped to identify and efficiently subsidize promising new technologies); Adam B. Jaffe, Richard G. Newell & Robert N. Stavins, A Tale of Two Market Failures: Technology and Environmental Policy, 54 Ecological Econ. 164, 171 (2005) (advocating that technology policy should be “‘technology neutral,’ encouraging all efforts that achieve specified objectives without focusing on a particular approach”).

[53].     Zachary Liscow & Quentin Karpilow, Innovation Snowballing and Climate Law, 95 Wash. U. L. Rev. 387, 391 (2017).

[54].     David Weisbach, Designing Subsidies for Low-Carbon Energy, 20 J. Env’t. & Sustainability L. 1, 6 (2013).

[55].     Technology Neutrality in Energy Tax: Issues and Options: Hearing Before the S. Comm. on Fin., 111th Cong. 60 (2009) (Statement of Gilbert E. Metcalf, Professor of Econ., Tufts Univ.).

[56].     Weisbach, supra note 54, at 14.

[57].     One way to deal with this problem is to set the subsidy value for green energy technologies equivalent to the implied carbon tax necessary for the dirtiest possible alternative, usually coal. The effect of this approach would be to “over-subsidize” green technologies that compete with natural gas.

[58].     Liscow & Karpilow, supra note 53, at 392; see also Daron Acemoglu, Philippe Aghion, Leonardo Bursztyn & David Hemous, The Environment and Directed Technical Change, 102 Am. Econ. Rev. 131, 132–33 (2012) (arguing that optimal environmental regulation should pair a carbon tax with direct encouragement for the development of specific clean technologies).

[59].     See, e.g., Gregory F. Nemet, How Solar Energy Became Cheap 175–79 (2d ed. 2026) (documenting innovation snowballing in the solar industry).

[60].     Regional Clean Hydrogen Hubs,U.S. Dep’t of Energy, https://www.energy.gov/oced/regional-clean-hydrogen-hubs-0 [https://perma.cc/9TJB-UTFF]; Regional Direct Air Capture Hubs, U.S. Dep’t of Energy, https://www.energy.gov/oced/DACHubs [https://perma.cc/9QZS-JR2V].

[61].     Technology Neutrality in Energy Tax: Issues and Options:Hearing Before the S. Comm. on Fin., supra note 55, at 1 (statement of Sen. Max Baucus, Chairman, S. Comm. on Fin.).

 [62].     S. Comm. on Fin, 113th Cong., Staff Discussion Draft: Energy Tax Reform (2013); see also Staff of J. Comm. on Tax’n, 113th Cong., Technical Explanation of the Senate Committee on Finance’s Staff Discussion Draft to Reform Certain Energy Tax Provisions (2013) (providing explanations of the proposed technology-neutral credits).

[63].     See infra Part II.C.

[64].     Cong. Budget Off., Estimated Budgetary Effects of H.R. 5376, the Inflation Reduction Act of 2022 (2022); Credit Suisse, Treeprint: US Inflation Reduction Act – A Tipping Point in Climate Action 4 (2022).

[65].     See generally Jeff Gordon, Statutory Contracts, 42 Yale J. on Regul. 660 (2025) (explaining the open-ended obligation of these spending programs).

[66].     I.R.C. §§ 40B, 45Q, 45V, 45Y, 45Z, 48E.

[67].     Kyoto Protocol to the United Nations Framework Convention on Climate Change, Dec. 10, 1997, 2303 U.N.T.S. 162; 1 Nat’l Greenhouse Gas Inventories Programme, Intergovernmental Panel on Climate Change, 2006 IPCC Guidelines For National Greenhouse Gas Inventories: General Guidance and Reporting 1.4 (Simon Eggleston, Leandro Buendia, Kyoko Miwa, Todd Ngara & Kiyoto Tanabe eds., 2006).

[68].     United Nations Framework Convention on Climate Change, Draft Handbook For The Review Of National Greenhouse Gas Inventories 54 (2019) (providing for a national inventory report); see also Yona, supra note 9, at 619–20 (on the distinction between measuring, reporting, and accounting for emissions).

[69].     See generally World Bus. Council for Sustainable Dev. & World Res. Inst., The Greenhouse Gas Protocol: A Corporate Accounting And Reporting Standard (rev. ed. 2015).

[70].     Id. at 25; Yona, supra note 9, at 629.

[71].     Energy Independence and Security Act of 2007, Pub. L. No. 110-140, sec. 201, 121 Stat. 1492, 1519–21; Mandatory Reporting of Greenhouse Gases, 74 Fed. Reg. 56260 (Oct. 30, 2009) (to be codified at 40 C.F.R. pts. 86, 87, 89, 90, 94, 98, 1033, 1039, 1042, 1045, 1048, 1051, 1054, 1065).

[72].     U.S. EPA, Fact Sheet: Greenhouse Gas Reporting Program Implementation (2013), https://www.epa.gov/sites/default/files/2014-09/documents/ghgfactsheet.pdf [https://perma.cc/JP35-QREP].

[73].     Renewable Fuel Petition Review Process, U.S. EPA (Apr. 10, 2025), https://www.epa.gov/renewable-fuel-standard-program/renewable-fuel-petition-review-process [https://perma.cc/F4WD-MPPR].

[74].     James S.N. McCubbins & A. Bryan Endres, EPA Biofuel Pathways and Petitions: Failure to Launch?, 3 farmdoc daily 100, 100 (2013).

[75].     See, e.g., Renewable Fuel Standard Program: Grain Sorghum Oil Pathway, 83 Fed. Reg. 37735, 37736 (Aug. 2, 2018) (to be codified at 40 C.F.R. pt. 80) (adding indirect land use effects to the evaluation of a fuel pathway). This methodological issue is discussed further below. See infra Part II.A.2.

[76].     See The Enhancement and Standardization of Climate-Related Disclosures for Investors, 89 Fed. Reg. 21668, 21916 (Mar. 28, 2024) (to be codified at 17 C.F.R. pts. 210, 229, 230, 232, 239, and 249).

[77].     The accounting can become more complicated if one considers the possibility of carbon leaking out of storage.

[78].     Bipartisan Budget Act of 2018 § 41119, I.R.C. § 45Q(a)(2).

[79].     See Renée Cho, Capturing Carbon’s Potential: These Companies Are Turning CO2 into Profits, Columbia Climate Sch.: State of the Planet (May 29, 2019), https://news.climate.columbia.edu/2019/05/29/co2-utilization-profits/ [https://perma.cc/ESR2-WQEM] (illustrating the range of industrial applications for carbon oxides).

[80].     I.R.C. § 45Q(f)(5)(B)(i).

[81].     Id.

 [82].     Inflation Reduction Act of 2022, Pub. L. No. 117-169, 136 Stat. 1818.

[83].     I.R.C. § 40B(b), 40B(e).

[84].     Id. § 45Z(b).

[85].     Id. § 45V(b)(2).

[86].     Id. §§ 45Y(b)(1)(A), 48E(b)(3)(A). Furthermore, the rules under 45Y create an interaction with the carbon sequestration credit, in that the emissions rate for an electricity facility must be lowered to the extent carbon dioxide is captured and stored or utilized pursuant to 45Q. See 26 C.F.R § 1.45Y-5(e) (2025).

[87].     One Big Beautiful Bill Act, Pub. L. No. 119-21, § 70512, 139 Stat. 72, 252 (2025). There is an exception for projects that begin construction within one year of the passage of OBBBA, or up through July 3, 2026.

[88].     See Treas. Reg. § 1.45Y-5(c)(2) (2025).

[89].     See infra text accompanying notes 116–120 (on indirect land use changes), text accompanying note 203 (on fuel cells), and note 251 (on carbon utilization). But seeinfra text accompanying note 243 (on negative emissions treatment for fuels).

[90].     See 42 U.S.C. § 7545(o)(1)(H).

[91].     I.R.C. § 61(a).

[92].     Id. § 4121(a)(1).

[93].     See generally Boris I. Bittker, A “Comprehensive Tax Base” as a Goal of Income Tax Reform, 80 Harv. L. Rev. 925 (1967); Henry Aaron, What Is a Comprehensive Tax Base Anyway?, 22 Nat’l Tax J. 543 (1969).

[94].     See Weisbach & Nussim, supra note 11, at 967 (explaining that the comprehensive base concept is grounded in both efficiency and fairness).

 [95].     I.R.C. §§ 40B(e)(2), 45Q(f)(5)(B)(ii), 45V(c)(1)(A), 45Y(b)(2)(B).

[96].     See Regulation of Fuels and Fuel Additives: Changes to Renewable Fuel Standard Program, 75 Fed. Reg. 14670, 14765 (Mar. 26, 2010) (to be codified at 40 C.F.R. pt. 80).

[97].     42 U.S.C. § 7545(o)(1)(H).

[98].     The Clean Air Act language refers to a “fuel lifecycle” because it pertains to the Renewable Fuel Standard program. When Congress incorporates that language for the purpose of tax credits that do not exclusively pertain to fuels, one should understand “fuel lifecycle” as substitutable for whatever product lifecycle is under evaluation.

[99].     For definitions of attributional and consequential carbon accounting, see Göran Finnveden, Michael Z. Hauschild, Tomas Ekvall, Jeroen Guinée, Reinout Heijungs, Stefanie Hellweg, Annette Koehler, David Pennington & Sangwon Suh, Recent Developments in Life Cycle Assessment, 91 J. Env’t Mgmt. 1, 3 (2009); Tomas Ekvall, Attributional and Consequential Life Cycle Assessment, in Sustainability Assessment at the 21st Century 41, 44 (María José Bastante-Ceca, Jose Luis Fuentes-Bargues, Levente Hufnagel, Florin-Constantin Mihai & Corneliu Iatu eds., 2020).

[100].     42 U.S.C. § 7545(o)(1)(H).

[101].     See Credit for Carbon Oxide Sequestration, 86 Fed. Reg. 4728, 4745 (Jan. 15, 2021) (to be codified at 26 C.F.R. pt. 1) (endorsing the use of a “cradle-to-grave boundary” for the lifecycle analysis under Section 45Q).

[102].     Yona, supra note 9, at 608–09.

[103].     Seeid. at 608; Lambert Schneider, Maosheng Duan, Robert Stavins, Kelley Kizzier, Derik

Broekhoff, Frank Jotzo, Harald Winkler, Michael Lazarus, Andrew Howard & Christina Hood, Double Counting and the Paris Agreement Rulebook, 366 Science 180, 181 (2019).

[104].     See Lambert Schneider & Anja Kollmuss, Perverse Effects of Carbon Markets On HFC-23 And SF6 Abatement Projects in Russia, 5 Nature Climate Change 1061, 1061 (2015) (describing the perverse incentives for waste generation in carbon offset markets).

[105].     I.R.C. § 1(g) (known as the “kiddie tax”).

[106].     Id. § 482.

[107].     Id. § 704(c)(1)(B).

[108].     This approach could draw lessons from a United Nations program meant to incentivize the destruction of HFC-23, a coolant manufacturing byproduct with deleterious climate impacts. In response to the Clean Development Mechanism’s incentives for destroying HFC-23, producers began to produce more of the gas in order to destroy it. The UN’s ultimate response was to refuse to award carbon credits to any new factories built after the announcement of the incentive. Elisabeth Rosenthal & Andrew W. Lehren, Profits on Carbon Credits Drive Output of a Harmful Gas, N.Y. Times (Aug. 8, 2012), https://www.nytimes.com/2012/08/09/world/asia/incentive-to-slow-climate-change-drives-output-of-harmful-gases.html [https://perma.cc/7GKU-MM9M] (“The United Nations, recognizing the temptation for companies to jump into the lucrative business, has refused since 2007 to award carbon credits to any new factories destroying the waste gas.”).

[109].     See U.S. Dep’t of Energy, Guidelines to Determine Well-to-Gate Greenhouse Gas (GHG) Emissions of Hydrogen Production Pathways Using 45VH2-GREET 11–12 (rev. 2024) [hereinafter GHG Guidelines] (restricting the amount of steam coproduct that steam methane reformers can claim based on the quantity of steam that an optimally designed reformer is expected to be capable of producing, in order to avoid incentivizing overproduction of hydrogen co-products).

[110].     Fran Moore, How Much Hot Air?, Briefing Book (Feb. 19, 2024), https://www.briefingbook.info/p/how-much-hot-air [https://perma.cc/Q3TS-PLJX].

[111].     Id.Cf. Welton, supra note 8, at 231 (making a similar point about the unpredictable equilibrium effect of encouraging mass adoption of carbon offset strategies, even if well-understood when undertaken at small scale).

[112].     Pol’y Dep’t for External Rels., Eur. Parliament, Four Briefings on Trade-Related Aspects of Carbon Border Adjustment Mechanisms, at 11 (Apr. 2020) (“[W]hile [carbon border adjustment] almost totally cancels sectoral leakages, its impact on overall leakage is limited, reducing it from 10.8 % to 8.2 % in the best case. Half of the overall leakage (5.9 p.p.) comes from the indirect channel: the emissions increase in the sectors that generate electricity, where emission intensity increases because of reduced fuel and other primary energy prices.”).

[113].     42 U.S.C. § 7545(o)(1)(H). See supra text accompanying notes 78–86 (detailing the tax credit statutes that reference the Clean Air Act definition of lifecycle emissions).

[114].42 U.S.C. § 7545(o)(1)(H).

[115].     Id.

[116].     Seegenerally Regulation of Fuels and Fuel Additives: Changes to Renewable Fuel Standard Program, 75 Fed. Reg. 14670 (Mar. 26, 2010) (to be codified at 40 C.F.R. pt. 80) (also known as the “RFS2 rule”).

[117].     See Stephanie Searle, If We Use Livestock Feed for Biofuels, What Will the Cows Eat?, Int’l Council on Clean Transp. (Jan. 10, 2019), https://theicct.org/if-we-use-livestock-feed-for-biofuels-what-will-the-cows-eat/ [https://perma.cc/5HR8-7EV9] (discussing how if corn oil is used for biofuel, farmers then need to replace it with something else in livestock feed).

[118].     Tyler J. Lark, Nathan P. Hendricks, Aaron Smith, Nicholas Pates, Seth A. Spawn-Lee, Matthew Bougie, Eric G. Booth, Christopher J. Kucharik & Holly K. Gibbs, Environmental Outcomes of the US Renewable Fuel Standard, 119 Procs. Nat’l Acad. Scis. U.S. Am. 1, 2 (2022).

[119].     Regulation of Fuels and Fuel Additives: Changes to Renewable Fuel Standard Program, 75 Fed. Reg. at 14766 (“The definition of lifecycle GHG emissions makes it clear that EPA is to determine the aggregate emissions related to the ‘full’ fuel lifecycle, including ‘all stages of fuel and feedstock production and distribution.’ Thus, EPA could not, as a legal matter, ignore those parts of a fuel lifecycle that occur overseas.”).

[120].     One Big Beautiful Bill Act, Pub. L. No. 119-21, § 70521, 139 Stat. 72, 276–79 (2025).

[121].     See Deepak Rajagopal & David Zilberman, On Market-Mediated Emissions and Regulations on Life Cycle Emissions, 90 Ecological Econ. 77, 77 (2013) (introducing the concept of market-mediated emissions).

[122].     For example, subsidy applicants in the biogas sector should not be able to claim credit for avoiding a counterfactual of venting methane into the atmosphere to the extent the Clean Air Act would have required them to collect or flare such gases. See Inst. for Pol’y Integrity, Comment Letter on Proposed Rule for Section 45Y Clean Electricity Production Credit and Section 48E Clean Electricity Investment Credit (Aug. 2, 2024), https://policyintegrity.org/documents/Policy_Integrity_Section_45Y_48E_Comments.pdf [https://perma.cc/MR9U-2KZ9]. On the concept of regulatory additionality in general, see Brian Joseph McFarland, Carbon Reduction Projects and the Concept of Additionality, 11 Sustainable Dev. L. & Pol’y 15 (2011).

[123].     See, e.g., Laurence H. Meyer, Joel L. Prakken & Chris P. Varvares, Designing an Effective Investment Tax Credit, 7 J. Econ. Persps. 189, 195 (1993); see Jane G. Gravelle, What Can Private Investment Incentives Accomplish? The Case of the Investment Tax Credit, 46 Nat’l Tax J. 275, 282–84 (1993).

[124].     See James Salzman & David Weisbach, The Additionality Double Standard, 48 Harv. Env’t L. Rev. 117, 129–33 (2024) (documenting the ubiquity of the additionality problem within tax subsidies).

[125].     A. Will Frazier, Cara Marcy & Wesley Cole, Wind and Solar PV Deployment after Tax Credits Expire: A View from the Standard Scenarios and the Annual Energy Outlook, 32 Elec. J. 1, 7 (2019).

[126].     This problem is referred to as inframarginality or “buying the base.” See Daniel E. Herz-Roiphe, Virtue’s Reward: A (Qualified) Defense of Infra-Marginal Tax Subsidies, 36 Va. Tax Rev. 1, 2, 5 n.11 (2017).

[127].     Economic Recovery Tax Act of 1981, Pub. L. No. 97-34, § 221, 95 Stat. 172, 241–47.

[128].     I.R.C. § 41(c)(3) (setting the base amount against which increases in research are measured in relation to historical spending on research).

[129].     Laura Tyson & Greg Linden, Ctr. for Am. Progress, The Corporate R&D Tax Credit and U.S. Innovation and Competitiveness 42–43 (2012); Cong. Budget Off., Federal Support for Research and Development 24 (2007).

[130].     Treas. Reg. § 1.45Y-5(e) (2025).

[131].     Credit for Production of Clean Hydrogen, 90 Fed. Reg. 2224, 2254 (Jan. 10, 2025) (to be codified at 26 C.F.R. pt. 1)

[132].     See infra Part III.B.4.

[133].     Salzman & Weisbach, supra note 124, at 119–20.

[134].     Id. at 133–38.

[135].     Examples of “perverse incentives,” or incentives that encourage behavior undermining their intended purpose, abound. In anticipation of the implementation of the Endangered Species Act, landowners preemptively altered land to make it uninhabitable for protected species so they would not be covered by the law. John A. List, Michael Margolis & Daniel E. Osgood, Is the Endangered Species Act Endangering Species? 27(Nat’l Bureau of Econ. Rsch., Working Paper No. 12777, 2006). The CAFE standards increased fuel efficiency, which made driving less costly per mile, but incentivized more driving, which increased emissions relative to the counterfactual. Todd Litman, Efficient Vehicles Versus Efficient Transportation Comparing Transportation Energy Conservation Strategies, 12 Transp. Pol’y 121, 128 (2005).

[136].     Ellen P. Aprill & Daniel J. Hemel, The Tax Legislative Process: A Byrd’s Eye View, 81 Law & Contemp. Probs 99, 103 (2018); see Rebecca M. Kysar, Tax Law and the Eroding Budget Process, 81 Law & Contemp. Probs. 61, 68 (2018) (describing the expansion of the reconciliation process).

[137].     On the appeal of tax-driven regulation in light of judicial hostility to other modes of regulation, seeGillian E. Metzger, To Tax, To Spend, To Regulate, 126 Harv. L. Rev. 83, 109 (2012) (predicting that tax law would emerge as a favored mode of economic regulation after National Federation of Independent Business v. Sebelius, 567 U.S. 519, 587 (2012)); Linda Sugin, The Great and Mighty Tax Law: How the Roberts Court Has Reduced Constitutional Scrutiny of Taxes and Tax Expenditures, 78 Brook. L. Rev. 777, 779 (2013) (noting a potential for tax law to “swallow all government policy” following Sebelius).

[138].     See Mackenzie Hawkins, Commerce Department Is Reaping Hundreds of Chips Act Inquiries, Bloomberg News (Aug. 9, 2023), https://news.bloomberglaw.com/ip-law/commerce-department-is-reaping-hundreds-of-chips-act-inquiries [https://perma.cc/6HBG-USZL].

[139].     See Jody Freeman & Jim Rossi, Agency Coordination in Shared Regulatory Space, 125 Harv. L. Rev. 1131, 1142 (2012) (introducing the idea of interagency coordination); Jason Marisam, Interagency Administration,45 Ariz. St. L.J. 183, 190(2013) (arguing that a rational agency enlists an outside agency to contribute expertise it lacks); Bijal Shah, Uncovering Coordinated Interagency Adjudication, 128 Harv. L. Rev. 805, 822 (2015) (on the need for varied expertise in adjudication as reason to bring in multiple agencies); Blaine G. Saito, Tax Coordination, 38 Ga. St. U. L. Rev. 735, 752–57 (2021) (advocating for interagency coordination in the tax law context).

[140].     See I.R.C. § 50(a)(7)(D)(ii).

[141].     See id. § 6103(l)(21).

[142].     See, e.g., id. § 45(b)(7)(A)(ii). In cases where the Department of Labor has not published a prevailing wage determination for a particular occupation, Treasury guidance directs taxpayers to use procedures established by the Secretary of Labor to request a wage determination. This means that the IRS will play no direct role in making such determinations; instead, it will rely on determinations made by Labor. IRS, Notice 2022-61, Prevailing Wage and Apprenticeship Initial Guidance Under § 45(b)(6)(B)(ii) and Other Substantially Similar Provisions (2022).

[143].     To be sure, there remain some practical differences between how the IRS and other peer agencies administer their respective spending programs. For example, Eric Toder notes that the costs of administering direct expenditure programs are generally more transparent than the costs of administering tax expenditures, because the IRS budget does not include the incremental costs of enforcing individual tax provisions, while agency budgets typically break out costs by program. Eric Toder, Tax Cuts or Spending - Does it Make a Difference?, 53 Nat’l Tax J. 361, 365 (2000).

 [144].     IRS, Notice 2024-37, Sustainable Aviation Fuel Credit; Lifecycle Greenhouse Gas Emissions Reduction Percentage and Certification of Requirements Related to the Clean Air Act; Climate Smart Agriculture; Safe Harbors (2024).

[145].     SeeClimate-Smart Agriculture and Forestry, U.S. Dep’t of Agric., https://www.farmers.gov/conservation/climate-smart [https://perma.cc/V959-PBK4].

[146].     Mandatory Reporting of Greenhouse Gases: Injection and Geologic Sequestration of Carbon Dioxide, 75 Fed. Reg. 75060, 75061 (Dec. 1, 2010) (to be codified at 40 C.F.R. pts. 72, 78, 98).

[147].     Treas. Reg. § 1.45Q-3(b)(2)(ii) (2021).

[148].     Letter from J. Russell George, Inspector Gen., Tax Admin., to the Sen. Robert Menendez 2 (Apr. 15, 2020), https://legacy-assets.eenews.net/open_files/assets/2020/04/30/document_gw_07.pdf [https://perma.cc/MUT2-39ZG].

[149].     I.R.C. § 40B(e)(2).

[150].     See IRS, supra note 144, § 3.03.

[151].     I.R.C. § 45V(c)(1)(B).

[152].     See generally GHG Guidelines, supra note 109.

[153].     Treas. Reg. § 1.45V-4(a)–(c) (2025).

[154].     Seeid. § 1.45Y-5(h)(6) (2026).

[155].     Id. §§ 1.45V-4(c)(3), 1.45Y-5(h)(3).

[156].     Id. § 1.45Q-4(c)(6) (2026).

[157].     See id. (making clear that the IRS determines whether to approve an LCA after DOE review). The fact that the IRS retains ultimate adjudicatory authority may be important for protecting against APA challenges. The D.C. Circuit has held that statutory silence does not constitute an implicit ban on interagency consultation, so long as the agency delegated by statute is the one making the final decision. See Nat’l Mining Ass’n v. McCarthy, 758 F.3d 243, 249 (D.C. Cir. 2014). See generally Bijal Shah, Interagency Transfers of Adjudication Authority, 34 Yale J. on Regul. 279 (2017) (arguing that interagency power transfers are constitutional and beneficial).

[158].     Looking to procedures under the Carbon Utilization Credit as a representative guide, the DOE will choose between two modes of review: a “Conformance Review” if the LCA report has been verified by an independent third party, or a “Critical Review” if it has not. A Conformance Review checks to see that the LCA conforms to applicable standards and to LCA guidance documents published by the National Energy Technology Laboratory, while a Critical Review involves “a detailed technical assessment of the LCA model and supporting data.” IRS, Notice 2024-60, Required Procedures to Claim a Section 45Q Credit for Utilization of Carbon Oxide 11–12 (2024).

[159].     See, e.g., Celanese Corp., Comment Letter on IRS Notice 2022-57 on the Credit for Carbon Oxide Sequestration (Dec. 3, 2022), https://www.bakerbotts.com/~/media/Files/Thought-Leadership/Publications/2022/December/CelaneseCorporation-58.pdf [https://perma.cc/5UR2-UPQZ].

[160].     Coglianese, supra note 14, at 532.

[161].     For an example of a design standard, consider the renewable energy investment tax credit, which names eleven acceptable forms of energy property, including solar energy equipment, geothermal equipment, fuel cell power plants, microturbine power plants, and so on. For each form of energy property, the statute provides further detailed design requirements. I.R.C. § 48(a)(3)(A).

[162].     Cary Coglianese & Jennifer Nash, The Law of the Test: Performance-Based Regulation and Diesel Emissions Control, 34 Yale J. on Regul. 33, 37 (2017).

[163].     Id.

[164].     Coglianese, supra note 14, at 545.

[165].     See Exec. Order No. 12,866, 3 C.F.R. 638 (1994) (urging regulatory agencies to “specify performance objectives, rather than specifying the behavior or manner of compliance that regulated entities must adopt”); Off. of Mgmt. & Budget, Circular A-4, at 8 (2003) (performance standards are “generally superior to engineering or design standards because performance standards give the regulated parties the flexibility to achieve regulatory objectives in the most cost-effective way.”); Exec. Order No. 13,563, 3 C.F.R. 13563 (2012) (directing administration to prioritize flexible approaches including “performance objectives”); see also Cass R. Sunstein, Simpler: the Future of Government 11–12 (2013) (“[F]lexible ‘performance standards’ . . . reduce[] costs [and] promote[] freedom.”).

[166].     Technology Neutrality in Energy Tax: Issues and Options:Hearing Before the S. Comm. on Fin., supra note 55, at 33 (statement of David L. Greene, Visiting Scholar, Univ. of California at Davis, Corp. Fellow, Oak Ridge Nat’l Lab’y).

[167].     See, e.g., Wendy E. Wagner, The Triumph of Technology-Based Standards, 2000 U. Ill. L. Rev. 83, 94–103 (arguing that design standards are easier to implement because they only require the government to identify a set of technological approaches that already work); Coglianese, supra note14, at 553–62.

[168].     Coglianese & Nash, supra note 162, at 45 (discussing the methods of assessment needed to enforce a performance standard).

[169].     See Emily Cauble, Safe Harbors in Tax Law, 47 Conn. L. Rev. 1385, 1398–400 (2015). However, certainty is not necessarily desirable. Tax scholars have observed that providing regulatory certainty helps taxpayers exploit poorly written rules. See Noël B. Cunningham & James R. Repetti, Textualism and Tax Shelters, 24 Va. Tax Rev. 1, 33 (2004); David A. Weisbach, Formalism in Tax Law, 66 U. Chi. L. Rev. 860, 867–72 (1999).

[170].     Coglianese & Nash, supra note 162, at 86.

[171].     Id. at 81.

[172].     Id. at 83.

[173].     See, e.g., GHG Guidelines, supra note 109.

[174].     See Cauble, supra note 169, at 1390, 1405; Susan C. Morse, Safe Harbors, Sure Shipwrecks, 49 U.C. Davis L. Rev. 1385, 1389 (2016).

[175].     Treas. Reg. § 1.7704-1(c) (1995).

[176].     I.R.C. § 7704(b); Treas. Reg. § 1.7704-1(h), (j).

[177].     I.R.C. § 7704(b).

[178].     I.R.C. § 121(c)(2); Treas. Reg. § 1.121-3(c)(2)(ii) (as amended in 2004).

[179].     Morse, supra note 174, at 1389.

[180].     Id. at 1397–402.

[181].     Id. at 1419–24 (discussing the risks of over- and under-inclusion in safe harbors).

[182].     See generally Albert O. Hirschman, The Rhetoric of Reaction: Perversity, Futility, Jeopardy (1991) (classifying three standard objections to reform proposals).

[183].     Joseph Bankman, The Tax Shelter Problem, 57 Nat’l Tax J. 925, 925–26(2004).

[184].     As of the enactment of the OBBBA, the Clean Hydrogen Credit is set to terminate for facilities beginning construction after December 31, 2027. See One Big Beautiful Bill Act, Pub. L. No. 119-21, § 70511, 139 Stat. 72, 252 (2025).

[185].     Dan Esposito, Eric Gimon & Mike O’Boyle, Energy Innovation Pol’y & Tech. LLC, Smart Design of 45V Hydrogen Production Tax Credit Will Reduce Emissions and Grow the Industry 12 (2023), https://energyinnovation.org/wp-content/uploads/2023/04/Smart-Design-Of-45V-Hydrogen-Production-Tax-Credit-Will-Reduce-Emissions-And-Grow-The-Industry.pdf [https://perma.cc/C68U-D3HS].

[186].     See generally Peter C. Canellos, A Tax Practitioner’s Perspective on Substance, Form and Business Purpose in Structuring Business Transactions and in Tax Shelters, 54 SMU L. Rev. 47 (2001).

[187].     See generally Joseph Bankman, The Business Purpose Doctrine and the Sociology of Tax, 54 SMU L. Rev. 149 (2001) (discussing the business purpose doctrine, a predecessor to the economic substance doctrine).

[188].     See Brian Galle, Interpretative Theory and Tax Shelter Regulation, 26 Va. Tax Rev. 357, 380–86 (2006).

[189].     See, e.g., Jonathan H. Choi, Beyond Purposivism in Tax Law, 107 Iowa L. Rev. 1439, 1442 (2022); Jonathan H. Choi, The Substantive Canons of Tax Law, 72 Stan. L. Rev. 195, 209–16 (2020) [hereinafter Choi, Substantive Canons of Tax Law].

[190].     I.R.C. § 7701(o)(1); see also Choi, Substantive Canons of Tax Law, supra note 189, at 225–27 (describing the codification of the economic substance doctrine).

[191].     See Jad Mouawad & Clifford Krauss, Lawmakers May Limit Paper Mills’ Windfall, N.Y. Times (Apr. 17, 2009), https://www.nytimes.com/2009/04/18/business/energy-environment/18sludge.html [https://perma.cc/A9T5-3D7R].

[192].     Id.; see I.R.S. Gen. Couns. Mem. 200941011 (Oct. 9, 2009).

[193].     See supra Part II.B (on the statutory language that supports a comprehensive emissions base).

[194].     See Marvin A. Chirelstein & Lawrence A. Zelenak, Tax Shelters and the Search for a Silver Bullet, 105 Colum. L. Rev. 1939, 1951–52 (2005) (discussing the appeal of “silver bullet” solutions that apply to a wide variety of shelters).

[195].     Tax Reform Act of 1986, Pub. L. No. 99-514, § 501, 100 Stat. 2233.

[196].     See James S. Eustice, Abusive Corporate Tax Shelters: Old “Brine” in New Bottles, 55 Tax L. Rev. 135, 147 n.71 (2002) (stating that Section 469 “deservedly received most of the credit for ending the prior shelter activity”); Chirelstein & Zelenak, supra note 194, at 1951 (“The pre-1986 shelters were all very much of a kind—virtually all involved the creation of artificial (noneconomic) losses for passive investors through the combination of tax preferences (most commonly accelerated depreciation) and interest expense deductions.”).

[197].Chirelstein & Zelenak, supra note 194, at 1952–53.

[198].     Treas. Reg. § 1.45Y-5(c)(2) (2025).

[199].     NYU Tax L. Ctr., Comment Letter on Guidance for Sections 48E and 45Y Technology-Neutral Credits 4–5 (Aug. 1, 2024), https://www.regulations.gov/comment/IRS-2024-0026-1571 [https://perma.cc/VA68-P4TT].

[200].     John A. Harrison, Yves T. Prairie, Sara Mercier-Blais & Cynthia Soued, Year-2020 Global Distribution and Pathways of Reservoir Methane and Carbon Dioxide Emissions According to the Greenhouse Gas from Reservoirs (G-res) Model, 35 Global Biogeochemical Cycles 1, 2–3 (2021). See generally Laura Scherer & Stephan Pfister, Hydropower’s Biogenic Carbon Footprint, 11 PLOS One 1 (2016).

[201].     Section 45Y Clean Electricity Production Credit and Section 48E Clean Electricity Investment Credit, 89 Fed. Reg. 47792, 47802 (proposed June 3, 2024) (to be codified at 26 C.F.R. pt. 1).

[202].     See NYU Tax L. Ctr., supra note 199, at 23–24; see also Consumers Power Co. v. Comm’r, 89 T.C. 710, 726 (1987) (holding that all parts of a hydroelectric plant “must be viewed as one integrated unit” for depreciation and ITC purposes because they “operate simultaneously and as a unit in order to produce electrical power”).

[203].     I.R.C. § 48E(j).

[204].     Types of Fuel Cells, U.S. Dep’t of Energy, https://www.energy.gov/eere/fuelcells/types-fuel-cells [https://perma.cc/P9NR-7K9X] (“Most fuel cells are powered by hydrogen.”); see infra Part III.4.

[205].     For example, the most recent Carbon Sequestration Credit regulations reduced the recapture period from five years to three years for carbon oxide leaking out of geologic storage. Credit for Carbon Oxide Sequestration, 86 Fed. Reg. 4728, 4750 (Jan. 15, 2021) (to be codified at 26 C.F.R. pt. 1) (“The final regulations revise the recapture period from five years to three years.”).

[206].     I.R.C. § 48E(b)(3)(A)(iii).

[207].     Treas. Reg. § 1.48E-5(h)(1), (2) (2025).

[208].     Id. § 1.48E-5(h)(1).

[209].     Id. § 1.48E-4(f)(1).

[210].     See generally Emily S. Bremer, Incorporation by Reference in an Open-Government Age, 36 Harv. J.L. & Pub. Pol’y 131 (2013). Changes to the models could substantially change the operation of the regulations, and so commenters urge the use of notice and comment rulemaking for any changes to the models. See Clean Air Task Force, Comment Letter on Proposed Section 45V Credit for Production of Clean Hydrogen; Section 48(a)(15) Election to Treat Clean Hydrogen Production Facilities as Energy Property 42–44 (Feb. 26, 2024), https://www.regulations.gov/comment/IRS-2023-0066-29565 [https://perma.cc/MH7N-PDV3].

[211].     See GHG Guidelines, supra note 109, at 19–20.

[212].     See generally Luke Hellgren, Pye Russell & Sierra Fraioli, Benchmarking Methane and Other GHG Emissions of Oil & Natural Gas Production in the United States (2023) (finding dramatic variations between companies and basins). That national estimate might itself be too low. See, e.g., Ramón A. Alvarez et al., Assessment of Methane Emissions from the U.S. Oil and Gas Supply Chain, 361 Science 186, 186 (2018); Union of Concerned Scientists, Comment Letter on Proposed Section 45V Credit for Production of Clean Hydrogen; Section 48(a)(15) Election to Treat Clean Hydrogen Production Facilities as Energy Property 12 (Feb. 26, 2024) https://www.regulations.gov/comment/IRS-2023-0066-29522 [https://perma.cc/E2W7-92U3].

[213].     Clean Air Task Force, supra note 210, at 35.

[214].     See, e.g., Coal. for Renewable Nat. Gas, Comment Letter on Proposed Section 45Y Clean Electricity Production Credit and Section 48E Clean Electricity Investment Credit 18–20 (Aug. 2, 2024), https://www.regulations.gov/comment/IRS-2024-0026-1696 [https://perma.cc/V6XL-EZA5].

[215].     Credit for Production of Clean Hydrogen and Energy Credit, 90 Fed. Reg 2224, 2275 (Jan. 10, 2025) (to be codified at 26 C.F.R. pt. 1). The vision seemed to be that the IRS would be able to verify reported leak rates with the EPA, because firms would soon have to report their leak rate to the EPA under a new methane-focused provision of the Greenhouse Gas Reporting Program. However, the EPA has now proposed to end most of the Greenhouse Gas Reporting Program. See generally Reconsideration of the Greenhouse Gas Reporting Program, 90 Fed. Reg. 44591 (Sep. 16, 2025) (to be codified at 40 C.F.R. pt. 98).

[216].     See supra text accompanying note 121 (defining market-mediated emissions).

[217].     See U.S. Dep’t of Energy, Assessing Lifecycle Greenhouse Gas Emissions Associated with Electricity Use for the Section 45V Clean Hydrogen Production Tax Credit 12 (2023) (discussing the use of “reasonable prox[ies]” to estimate the induced emissions of adding new loads to the grid).

[218].     Treas. Reg. § 1.45V-4(d)(3) (2025).

[219].     Letter from Janet G. McCabe, Deputy Adm’r of the Env’t Prot. Agency, to Lily Batchelder, Assistant Sec’y for Tax Pol’y (Dec. 20, 2023), https://home.treasury.gov/system/files/136/45V-NPRM-EPA-letter.pdf. [https://perma.cc/DP2F-ZAVC] (describing the proposed pillars as “a methodological proxy in lieu of quantifying certain indirect greenhouse-gas emissions associated with electrolytic hydrogen production”).

[220].     Wilson Ricks, Qingyu Xu & Jesse D. Jenkins, Minimizing Emissions from Grid-Based Hydrogen Production in the United States, 18 Env’t. Res. Letters 1, 6 (2022) (finding that “when a portion of the highlighted high-quality resource is procured for hydrogen production and cannot be used for grid supply, the system chooses to retire less coal capacity rather than developing significantly more expensive wind resources”).

[221].     Treas. Reg. § 1.45V-4(d)(3)(i). The logic of the nuclear reactor provision is that if the hydrogen project did not buy electricity from a financially struggling reactor, the reactor might retire and be replaced on the grid with a fossil resource.

[222].     Treas. Reg. § 1.45V-4(d)(2)(x)(B).

[223].     Inst. for Pol’y Integrity, supra note 122, at 11–12.

[224].     Section 45Y Clean Electricity Production Credit and Section 48E Clean Electricity Investment Credit, 89 Fed. Reg. 47792, 47804 (proposed June 3, 2024) (to be codified at 26 C.F.R. pt. 1).

[225].     Separately, enforcing the “first productive use” rule would be confusing in practice. One commenter warns of a gaming strategy where dairy farmers might move lagoons of cow manure into new anaerobic digesters in order to claim “first productive use” of the resulting methane production. See Fidelis New Energy, LLC, Comment Letter on Proposed Section 45V Credit for Production of Clean Hydrogen; Section 48(a)(15) Election to Treat Clean Hydrogen Production Facilities as Energy Property 14 (Feb. 26, 2024), https://www.regulations.gov/comment/IRS-2023-0066-29748 [ttps://perma.cc/Q48G-Q8YY].

[226].     Treas. Reg. § 1.45Y-5(e)(3) (2025).

[227].     See One Big Beautiful Bill Act, Pub. L. No. 119-21, § 70521, 139 Stat. 72, 276 (2025).

[228].     See Inst. for Pol’y Integrity, supra note 122, at 12.

[229].     See Energy Info. Admin., U.S. Dep’t of Energy, Renewable Fuels Module of The National Energy Modeling System: Model Documentation 2022, at 19, 76 (2022), https://www.eia.gov/outlooks/aeo/nems/documentation/renewable/pdf/RFM_2022.pdf [https://perma.cc/SKV6-HWPL]. See generally Alex Marten, Andrew Schreiber & Ann Wolverton, U.S. EPA, SAGE Model Documentation (2024), https://www.epa.gov/environmentaleconomics/sage-model-documentation-version-201 [https://perma.cc/JSG5-4TXQ]; Kenneth Hanson, Elise Golan, Stephen Vogel & Jennifer Olmsted, U.S. Dep’t Of Agric., Tracing The Impacts Of Food Assistance Programs On Agriculture And Consumers: A Computable General Equilibrium Model (2002).

[230].     See, e.g., Paul Caron, Moore v. The United States: Will The Supreme Court Join the Whack-a-Tax Shelter Game?, TaxProf Blog (Aug. 6, 2024), https://taxprof.typepad.com/taxprof_blog/2024/08/moore-v-the-united-states-will-the-supreme-court-join-the-whack-a-tax-shelter-game.html [https://perma.cc/945A-U8PQ] (describing the effort to crack down on tax shelters as an “endless game of whack-a-mole”); Brian Galle & Stephen Shay, Admin Law and the Crisis of Tax Administration, 101 N.C. L. Rev. 1645, 1690 (2023) (describing IRS enforcement of tax shelters as “whack-a-mole”).

[231].     See generally ACM P’ship v. Comm’r., 157 F.3d 231 (3d Cir. 1998); Black & Decker Corp. v. United States, 340 F. Supp. 2d 621 (D. Md. 2004).

[232].     The Chirelstein-Zelenak proposal offers more predictability than the economic substance doctrine for two reasons: It describes the circumstances of its application in mathematically precise terms rather than relying on the subjective standard of “change to the taxpayer’s economic position,” and it actively disallows the transactions it describes, rather than merely giving Congress’s imprimatur for the IRS to challenge such transactions at its discretion. See generally Chirelstein & Zelenak, supra note 194. Seealso I.R.C. § 7701(o)(5)(C) (“The determination of whether the economic substance doctrine is relevant to a transaction shall be made in the same manner as if this subsection had never been enacted.”).

[233].     Treas. Reg. § 1.45Y-5(d)(2)(iv) (2025).

[234].     Id. § 1.45Y-5(e)(3)(v) (mandating a negative carbon intensity score of -51 gCO2e/megajoule for combustion of methane derived from animal waste).

[235].     Id. § 1.45Y-5(e).

[236].     See generally Shelley Welton, Neutralizing the Atmosphere, 132 Yale L.J. 171 (2022); Albert C. Lin, Making Net Zero Matter, 79 Wash. & Lee L. Rev. 679 (2022); Yona, supra note 9; Daniel C. Esty & Nathan de Arriba-Sellier, Zeroing in on Net-Zero: From Soft Law to Hard Law in Corporate Climate Change Pledges, 94 U. Colo. L. Rev. 635 (2023); Salzman & Weisbach, supra note 124.

[237].     Renewable Natural Gas Production, U.S. Dep’t of Energy: Off. of Energy Efficiency & Renewable Energy, https://afdc.energy.gov/fuels/natural-gas-renewable [https://perma.cc/CTA6-QJ4P].

[238].     See LCFS Pathway Certified Carbon Intensities, Cal. Air Res. Bd., https://ww2.arb.ca.gov/resources/documents/lcfs-pathway-certified-carbon-intensities [https://perma.cc/M6ZS-23U2] (explaining the negative carbon intensities attributed to biomethane).

[239].     On counterfactual assumptions, one study found that if methane would otherwise have been vented, it would have a highly negative lifecycle emissions value. If, however, it would otherwise have been flared, it would have a negative lifecycle value only if the leakage rate is under 1 percent, which is improbable. Worse still, if the methane were intentionally produced for the sake of receiving the green subsidy (rather than already in existence), it would have a positive lifecycle emissions value. Emily Grubert, At Scale, Renewable Natural Gas Systems Could Be Climate Intensive: The Influence of Methane Feedstock and Leakage Rates, 15 Env’t Rsch. Letters 1, 2 (2020).

[240].     The risk of methane blending with natural gas arises because both are delivered via the same pipelines and methods for differentiating between the two gases once commingled are limited. See Inst. for Pol’y Integrity, supra note 122, at 17.

[241].     See Emily Grubert & Danny Cullenward, The New Hydrogen Rules Risk Opening the Door to Methane Offsets, Heatmap (Feb. 9, 2024), https://heatmap.news/climate/hydrogen-tax-credit-final-methane-offsets [https://perma.cc/A6UW-ZZ8A] (warning about the consequences of allowing methane offsets in the calculation of the Clean Hydrogen Credit).

[242].     See Letter from Sen. Sheldon Whitehouse et al., to Janet Yellen, Sec’y, U.S. Dep’t of the Treasury et al. (Sep. 11, 2024), https://www.whitehouse.senate.gov/wp-content/uploads/2024/09/Letter-to-Biden-Administration-re-45V-Hydrogen-Tax-Credit-09-11-2024.pdf [https://perma.cc/8JMV-5MB3] (arguing for the same restriction); see also Rocky Mountain Inst., Comment Letter on Proposed Section 45V Credit for Production of Clean Hydrogen (Feb. 26, 2024), https://www.regulations.gov/comment/IRS-2023-0066-29657 [https://perma.cc/YX3D-QV88].

[243].     One Big Beautiful Bill Act, Pub. L. No. 119-21, § 70521, 139 Stat. 72, 276 (2025).

[244].     See Treas. Reg. § 1.45Y-5(e)(3) (2025) (continuing to require that taxpayers take into account “avoided emissions”).

[245].     See David Kamin, David Gamage, Ari Glogower, Rebecca Kysar, Darien Shanske, Reuven Avi-Yonah, Lily Batchelder, J. Clifton Fleming, Daniel Hemel & Mitchell Kane, The Games They Will Play: Tax Games, Roadblocks, and Glitches Under the 2017 Tax Legislation, 103 Minn. L. Rev. 1439, 1490–93 (2019) (discussing opportunities under the 2017 tax law to blend income across countries and providing numerical examples).

[246].     See I.R.C. § 951A(a).

[247].     See Stephen E. Shay, J. Clifton Fleming Jr. & Robert J. Peroni, Designing a 21st Century Corporate Tax—An Advance U.S. Minimum Tax on Foreign Income and Other Measures to Protect the Base, 17 Fla. Tax Rev. 669, 706 (2015) (recommending that any minimum tax be determined on a per-country basis).

[248].     See Martin A. Sullivan, Economic Analysis: Will GILTI Save U.S. Multinationals From GLOBE?, Tax Notes (Dec. 23, 2019), https://www.taxnotes.com/featured-analysis/economic-analysis-will-gilti-save-us-multinationals-globe/2019/12/20/2bq6t [https://perma.cc/XP9G-BDL7] (criticizing GILTI for undermining any consequences for activities in low-tax jurisdictions).

[249].     Rocky Mountain Inst., supra note 242, at 18.

[250].     See Mathias Zacarias & Joseph Majkut, 45V or 45Q? How Tax Credits Will Influence Low-Carbon Hydrogen’s Development, Ctr. for Strategic & Int’l Stud. (Feb. 15, 2024), https://www.csis.org/analysis/45v-or-45q-how-tax-credits-will-influence-low-carbon-hydrogens-development [https://perma.cc/5RJK-AZFB](“45V should be the tax credit of choice for all projects with lifecycle emissions below 1.5 kg CO2e per kg H2. These projects are granted a tax credit of $1 per kg H2, higher than what could be attained through 45Q.”).

[251].     This argument pertains only to carbon sequestered in geologic storage. If the taxpayer gets credit for carbon captured and sold to third parties (an option allowed under the Clean Electricity Credits), then another additionality problem surfaces because there is a risk that carbon utilized by third parties will ultimately still escape into the atmosphere. In the OBBBA, Congress amplified this risk by increasing the credit value for carbon utilization up to parity with the credit for geologic storage. See One Big Beautiful Bill Act, Pub. L. No. 119-21, § 70522, 139 Stat. 72, 279–80 (2025).

[252].     The Clean Electricity Credits Final Rule does not allow carbon captured in the course of producing input fuels to count against the emissions rate, but it does allow carbon captured at the generation facility to count against the emissions rate. See Treas. Reg. § 1.45Y-5(f) (2025).

[253].     See Joshua D. Blank & Ari Glogower, The Trouble with Targeting Tax Shelters, 74 Admin. L. Rev. 69, 84 (2022) (describing how attempts to target tax shelters with activity-based rules can be overinclusive and impose unnecessary burdens on fully compliant taxpayers).

[254].     Chirelstein and Zelenak identify the following among noneconomic losses clearly contemplated by Congress that their rule would overinclusively disallow: “the noneconomic loss created by the combination of tax-free imputed rental income and deductible qualified residence interest, see I.R.C. § 163(h)(2)(D) (Thomson 2005); the charitable contribution deduction for appreciation never taxed to the donor, see id. § 170(e); the deduction for amounts contributed to individual retirement accounts, id. § 219(a); and the limited expensing of long-lived assets used in a taxpayer’s trade or business, id. § 179.” Chirelstein & Zelenak, supra note 194, at 1956 n.55.

[255].     See Treas. Reg. § 1.6011-4(b) (as amended in 2010); Blank & Glogower, supra note 253, at 77 (describing how the reportable transactions regime requires the IRS to spend resources investigating transactions that turn out to be legal). For example, taxpayers must disclose any transaction that results in a large tax loss ($10 million in a single year, in the case of corporations, and $2 million in a single year, in the case of individuals)—even though such losses might be fully bona fide. Treas. Reg. § 1.6011-4(b)(5).

[256].     Cf. Julie McNamara, These Are the Critical Issues to Track with the New “Tech-Neutral” Clean Electricity Tax Credits, Union of Concerned Scientists: Equation (Aug. 21, 2024), https://blog.ucs.org/julie-mcnamara/these-are-the-critical-issues-to-track-with-the-new-tech-neutral-clean-electricity-tax-credits/ [https://perma.cc/MG8L-ZL65] (“In the face of uncertainty over input assumptions, analytical boundaries, and/or counterfactuals that have the potential to fundamentally shape the analytical outcome, a conservative approach is appropriate to ensure the tax credits do not inadvertently subsidize a net-harmful outcome for the climate.”).

[257].     See generally Lynn A. Stout, Type I Error, Type II Error, and the Private Securities Litigation Reform Act, 38 Ariz. L. Rev. 711, 711 (1996) (defining the two types of errors).

[258].     These would be variance errors rather than bias errors. See generally Scott Fortmann-Roe, Understanding the Bias-Variance Tradeoff (2012), https://scott.fortmann-roe.com/docs/BiasVariance.html [https://perma.cc/JTL7-K7L3] (distinguishing between bias and variance errors).

[259].     See supra Part II.A (discussing the evolution of carbon accounting to correct past oversights). See generally Yona, supra note 9.

[260].     See, e.g., Charles Cannon, Pieter Gagnon, Gavin McCormick & Wilson Ricks, Assessing the Impact of Voluntary Actions on the Grid 4 (2024) (expressing a lack of consensus on the consequential emissions impact of procurement of clean energy on the grid, which is the central channel of emissions reductions in the Clean Hydrogen Credit).

[261].     Especially once we have already achieved high levels of wind and solar penetration, it tends to be more expensive to produce additional energy with low emissions compared to default fossil fuel alternatives. See James H. Stock, Driving Deep Decarbonization, Fin. & Dev., Sep. 2021, at 12, 14–15 (“The prospect of sustainable aviation fuel competing with petroleum jet fuel . . . is daunting. . . . [N]ew wind and solar power generation is less expensive than coal and natural gas in some but not all parts of the country.”); see also Kenneth Gillingham, Carbon Calculus, Fin. & Dev., Dec. 2019, at 6, 8 (“In the short term, there are some inexpensive ways to reduce emissions, but deeper cuts run up against quickly rising costs.”).

[262].     It is widely understood that the purpose of Pigouvian taxes and subsidies is to make clean energy cheaper relative to fossil fuels, not just cheaper relative to other goods and services. See John Bistline, Neil Mehrotra & Catherine Wolfram, Economic Implications of the Climate Provisions of the Inflation Reduction Act 4 (Nat’l Bureau of Econ. Rsch., Working Paper No. 31267, 2023) (describing the purpose of both the IRA subsidies and carbon taxation as to “lower the relative price of clean to fossil fuel power generation”).

[263].     To be clear, there is no budgeted limit on the possible subsidy expenditure under the Inflation Reduction Act. In the long term, however, Congress’s appetite for green subsidies is presumably subject to whatever budget constraint that applies to fiscal policy generally.

[264].     Cf. Grubert & Cullenward, supra note 241 (explaining that if biogas is allowed a negative emissions rate, a plant that blends biogas with natural gas could qualify for tax credits).

‍ ‍

While a programmer wishing to respond to this argument might attempt to remove both race and all racial proxies from input data, their task is complicated by a key dilemma: The definition of a racial proxy is far from obvious. In response to this dilemma, scholars, computer programmers, and advocates have proffered various approaches to defining a racial proxy and solutions to the problem of racial proxy discrimination. Diverse as they are in their methodologies, these solutions share a common quality. Each relies on an underlying assumption about the relationship between race and the racial proxy, and these assumptions can have far-reaching implications for the development and regulation of machine learning algorithms.

This Article examines the myriad definitions of a racial proxy proffered by courts, scholars, and state and private actors to demonstrate how race and racial assumptions become embedded in the machine learning algorithms that increasingly structure human life. By examining the jurisprudence identifying a racial proxy, the Article explains how those who develop and deploy algorithms assume a powerful adjudicatory role that was once exclusively reserved for judges—that is, they can use their racial intuition to decide what gives a variable its racial quality. The Article shows that the answer to the question of what constitutes a racial proxy requires an explicitly normative and political solution and thus cannot be resolved with the purely technical solutions emerging within econometric literature on algorithmic antidiscrimination.

Ultimately, what is at stake in the ability to define a racial proxy is a novel form of algorithmically driven racial construction, which permits the production of new and meaningful classes of individuals that can later be exposed to differing resources, opportunities, subordination, and privilege. The Article identifies this process as a necessary site of legal thought and regulation.

Table of Contents Show

Introduction

On October 18, 1920, an Oregon district court ruled that Bhagat Singh Thind, an immigrant from India who served in the United States military, was entitled to naturalize as a United States citizen.[1] At the time, federal law restricted naturalization to “free [W]hite persons” as well as “aliens of African nativity and to persons of African descent.”[2] Unsatisfied with the court’s ruling, the federal government appealed Thind’s case to the Ninth Circuit Court of Appeals, which in turn asked the Supreme Court for instruction on the question of whether Thind was, in fact, a White person and thus eligible for naturalization.[3] More precisely, the circuit court presented the following question to the Justices: “Is a high-caste Hindu, of full Indian blood, born at Amritsar, Punjab, India a [W]hite person within the meaning of [the Act]?”[4]

The Supreme Court ultimately answered the question in the negative, effectively ruling that it was common knowledge that Asian Indians were distinguishable from Whites.[5] But the unwieldy composition of the circuit court’s question—a question that explicitly contemplated the role that features like caste, ancestry, religion, and birthplace played in determining racial identity—suggests that, for the judges, racial determinations could turn on the presence or absence of certain proxy variables.[6] The invocation of these racial proxies as part of an assessment of racial identity suggests that the judges relied on proxies for race to determine the very racial categories that were subsequently painted as natural, self-evident, or commonly understood. Importantly, each of these determinations was central to dispensing political, social, and economic rights and privileges, including Bhagat Singh Thind’s right to naturalize.

It may come as a surprise to some that in our digital age antidiscrimination discourse is still troubled by the same questions that perplexed this twentieth-century circuit court. While the constitutional law of equality forbids expressly conditioning the distribution of certain government benefits and burdens on race, the question of how, when, or why certain features like blood, caste, religion, birthplace, residency, and language proxy for racial identity is as salient and vexing as ever in the information age.

By surveying judicial, statistical, and critical race definitions of a racial proxy variable, this Article shows how equality law’s failure to acknowledge the co-constitutive relationship between race and its so-called proxies has opened the door to new and compounded forms of racial inequality in the digital age. Central to this analysis is the idea that race is the outcome produced when those holding immense powers of official decree embed the presence or absence of certain variables with political meaning in their attempt to hierarchically organize and order human life. That race itself has been a proxy for this extraordinary and continuous process of naming, instantiating, and ordering human inequality must be understood in the context of our digital age. Today, such powers of formal decree are undergoing monumental shifts and are, in many cases, driven by globalized technology conglomerates with vast political and economic power.

Indeed, in a world where machine learning algorithms distribute important benefits and burdens, the answer to the question of what features are sufficiently linked to race is no longer exclusively driven by courts, litigators, and juries.[7] Increasingly, a novel set of actors that develop and deploy algorithms make these determinations, and the consequences of their determinations are poorly understood and contained by law and legal thought. Understanding the interaction between antidiscrimination law and technology not only reveals how the collision of the two can reinforce existing forms of racial inequality, but also how novel forms of racialized inequality are made possible by the machine learning algorithms that increasingly structure human life.

I use the term “machine learning algorithm” to capture a diverse array of models and techniques. What unites these algorithms is the centrality of human decision-making to their development and deployment, especially decisions that encompass data selection and use.[8] At a high level, these algorithms are trained on large sets of historical data to search for patterns and complex rules, often to predict the probability of some future outcome.[9] Indeed, the term “machine learning” refers to automated processes of uncovering correlations, relationships, and patterns between variables in a dataset.[10] Determining the source and scope of this historical data is one of the most significant determinations a programmer will make, since a wide variety of data points can be sourced from equally numerous institutions, systems, and contexts. As with many decisions of great consequence to the construction of machine learning algorithms, decisions about data selection and use unfold with limited legal intervention.

An important exception to this absence of legal direction, however, is American equality law’s contemporary legal commitment to formal colorblindness. In practice, this commitment has meant that most machine learning algorithms are programmed not to “see” race.[11] The extent to which this claim to so-called “race-blindness” can be made rests on the fact that an individual’s race is usually excluded as an input variable when developing or deploying an algorithm.[12] For example, an algorithm designed to predict the probability that an individual borrower will default on a loan may exclude the person’s race from the many input variables used to predict their risk of default. This approach to algorithmic fairness is often called “fairness through blindness.”[13] It assumes that excluding the race variable from data will produce an algorithmic decision free from racial discrimination.

But the “fairness through blindness” approach is just one of many contested meanings, possibilities, and constructions of algorithmic fairness,[14] and it has drawn many critics.[15] These critics often argue that even when machine learning algorithms are built to exclude race as an input, these algorithms can still discriminate based on race by relying on so-called proxies for race, a facially race-neutral variable that has a significant relationship with race, such as income, geography, or criminal arrest history.[16]

Concerns about racial proxies impacting algorithmic decisions are buttressed by the technical nature of algorithms. Even when algorithms exclude race as an input, racial patterns and correlations remain visible in formally race-blind data and may recreate the effect of race on algorithmic outputs.[17] The tendency of algorithms to reconstruct an omitted input variable from other unrestricted input variables reinforces the concern about whether (or how) to allow an algorithm access to input data that can “proxy” for an individual’s race.[18] What should be done with these variables that carry the virtually certain risk of introducing racial information into formally race-neutral algorithms?

The technical and legal implications of this question are of great importance in algorithmic antidiscrimination discourse.[19] Scholars have examined the inclusion of potentially problematic racial proxies, discussed the trouble these variables pose for delivering a truly race-neutral algorithm, and proposed potential workarounds.[20] While one obvious approach is to treat racial proxies as one would treat the race variable and simply remove them from an algorithm’s input data, the feasibility of this approach rests on a definition of the term “racial proxy” that remains elusive.[21]

There is no static, universal, or precise conception of what makes a variable a proxy for race.[22] In the statistical literature, there is no consensus among computer programmers about the definition of a racial proxy.[23] Legal scholars have similarly established that even those variables we might intuitively understand to proxy for race, like zip code, income, or criminal history, may be less correlated with race than other, less intuitive variables, such as an individual’s online purchases, location data, or membership in a digital group or community.[24] If solving the problem of racial proxies first requires understanding their fundamental meaning and nature, then computer programmers and developers face the profound theoretical dilemma of deciding precisely what gives any given proxy variable its “racial” quality.

In practice, this dilemma gives rise to what some scholars deem a “haphazard” approach among computer programmers to identifying and dealing with these racial proxy variables.[25] To be sure, the normative decisions of programmers and developers may lack a fine-toothed explanation, but the capriciousness of their decisions ought not to be confused with randomness. Their decisions originate from a set of racial assumptions about the commonsense or objective reality of race and its relationship to a racial proxy: a racial intuition. This racial intuition forms the basis of technical decisions concerning input data that impact both the development of machine learning algorithms and, I will argue, the evolution of race in the digital age.

This Article analyzes the racial theories undergirding definitions of a racial proxy variable in algorithmic and antidiscrimination discourse and discusses their far-reaching implications for legal regulation and racial production. I begin by analyzing the definitions of a racial proxy proffered by courts to show how a powerful adjudicatory role once exclusively reserved for judges is now in the hands of those who develop and deploy algorithms. That role is to use their poorly defined racial intuition to decide what gives a variable its racial quality.

Throughout the Article, I draw connections between the racial common sense used by a programmer to identify a racial proxy and the confused and often contradictory understandings of race that have historically marked the racial assumptions of law and legal actors. The intention behind this examination is not simply to connect past and present but to reveal two significant features of our present-day legal order.

First, the production of so-called “haphazard” technical decisions made by programmers is made possible by the law’s delegation of breathtaking authority to those who develop and deploy algorithms to select a definition of a racial proxy among a host of alternative meanings. Our legal regimes open the door to a programmer’s exercise of extreme epistemic authority by framing the definition of a racial proxy as somehow self-evident or intuitive instead of the product of political and normative reasoning.

Second, the power delegated to those who develop and deploy algorithms goes beyond the power to embed contestable legal concepts like “fairness” and “discrimination” into technical systems.[26] It even extends beyond the power to project historical racial inequality into the future via predictive systems.[27] I argue that those who develop and deploy algorithms can also facilitate novel changes to the racial subject at the center of antidiscrimination law. My argument is that those who define racial proxies do not merely project the racialized past into the future: They construct new and salient classes of individuals who can, on the basis of their digitally observable attributes, be exposed to differentiated access to resources, privileges, opportunities, and hierarchies of status. This process represents a racialization by and for the information age, and the fact that this extraordinary power rests in the hands of modern technology companies with immense political and economic power is a regulatory failure with consequences that are both immediate and yet unknown.

The Article proceeds in four Parts. Part I further explicates the dilemma at the core of the racial proxy debate. Racial proxy discrimination stands as a ubiquitous feature of formally race-blind machine learning algorithms, but addressing this proxy discrimination is seemingly impossible without a precise meaning of a racial proxy. This meaning is far from obvious since there is no inherent quality that turns a variable into a racial proxy.[28] This Section also contextualizes the racial proxy debate, by placing it within a long history of critical scholarship concerned about discrimination based on “proxies” for legally protected categories like race and gender. Finally, I explain how the proliferation of machine learning algorithms has intensified and refashioned these earlier debates in important ways.

Part II turns to the jurisprudential answer to the question of what precisely constitutes a racial proxy. It shows how computer programmers identify racial proxies in a jurisprudential environment wrought with theoretical contradictions and unanswered questions.[29] Understanding the historical development of America’s racial jurisprudence assumes a new salience in a digital age, when legal constructions of the racial proxy are supplemented by algorithmic constructions authored by a new set of actors facing similar normative dilemmas.

Indeed, under American equality law’s reigning attachment to formal colorblindness, race-consciousness done for the purpose of racial animus is tantamount to race-consciousness done for the purpose of racial repair.[30] That each race-conscious action furthers a distinct ideological project is of little consequence under the current colorblind framework. As Cheryl Harris writes, “Under colorblindness race is reduced to color, a biological attribute like height or eye color, and is therefore presumptively normatively irrelevant.”[31] Under this commitment to formal colorblindness, the search for a racial proxy is similarly understood as the hunt for a presumptively fixed attribute that gives a variable its racial quality.[32] The assumption is that judges who identify a racial proxy are merely recognizing a racial fact, rather than demonstrating their investment in one particular conception of the racial proxy among the many that exist.

At a high level, the problem with racial proxies implicates two concerns in equality law. First, a concern about disparate racial treatment is implicated when a racial proxy is understood as functionally equivalent to race.[33] In this perspective, a decision-maker can be formally blind to race, but the functional equivalence between race and another variable means that intentional racial discrimination can still be surreptitiously smuggled in.[34] The racially inflected normative determination for judges (and now for programmers) becomes which variables are functionally equivalent to race and why.

The other legal concern is a worry about the high correlations between race and the racial proxies, which result in racially disparate outcomes.[35] Here, the normative determination for judges and programmers becomes the level of racial differentiation that is permissible before differentiation becomes discrimination. In other words, judges draw a normative distinction between disparities produced by discrimination and disparities attributed to purportedly “raw” racial difference.

Both strands of concern present important definitional challenges for identifying and resolving racial proxy discrimination. Once seen through the unifying project of defining the racial proxy, the distinction between the two strands appears less prominent[36] as courts move between and across these two variants and label a wide array of factors “proxies for race,” from geographic location[37] to ancestry.[38] At the same time, courts curiously decline to recognize racial proxy discrimination based on other highly race-correlated features—like pending criminal charge or certain hairstyles heavily associated with race.[39]

The unpredictable judicial construction of a racial proxy reveals that, apart from judges’ normative determinations, there is no abstract, formal property that turns a variable into a racial proxy. When courts describe something as a racial proxy, they exercise their powers of legal decree to make a political assertion about the necessary relationship between race and the racial proxy. Foundational to this determination is judges’ own “racial common sense.”[40] Translated into a world structured by machine learning algorithms, this jurisprudence gives those who develop and deploy algorithms the authority to identify racial proxies from a potentially infinite number of race-correlated variables whose effect on any given algorithmically produced outcome may be counterintuitive.[41]

From this perspective, what some scholars of law and technology have identified as a haphazard approach among computer programmers to addressing racial proxy discrimination in algorithms can instead be understood as an expression of the judicial reasoning internal to the very development of the legal regimes permeating America’s racial jurisprudence.[42] Judicial thinking about the racial proxy variable frames its definition as somehow obvious or self-evident. Practically speaking, within the context of colorblindness, both judges and programmers can seize on a seemingly intuitive definition of the racial proxy to avoid rigorously justifying their decisions with reference to antidiscrimination principles and moral commitments.

Recognizing the challenges inherent to defining racial proxies, emerging econometric solutions attempt to avoid the definitional question altogether. Part III discusses two methods that stand out because of their uptake in legal scholarship. One method promises to “purge” the proxy effects of race from all algorithmic inputs, allowing computer programmers to isolate and then excise the “racial” effects of all data inputs.[43] The other method is skeptical that excluding, orthogonalizing, or otherwise scrutinizing inputs will create a race-neutral algorithm.[44] Instead, the focus is on the racial outcomes produced by the algorithm.[45]

While both approaches attempt to avoid the normative difficulty in defining the racial proxy, they subsequently rely on their own complex and contestable assumptions about the relationship between race and the racial proxy. Given the potential for these assumptions to be embedded in the technological systems that increasingly structure our world, this Part probes each approach for a deeper understanding of its underlying racial ontology.

In response to techniques that seek to isolate and excise the racial effects of all data inputs, I raise an underexplored conceptual challenge to this approach, suggesting that it relies on a common but paradoxical understanding of race and its relationship to the racial proxy. The issue arises from the assumption that the race variable—which is usually secured via institutional ascription or self-identification—can be successfully isolated from the “proxies” that have historically been used to assign people to their proper race in the first place. This Section recasts race itself as a proxy, standing in for an authoritative background system of racial knowledge that has historically relied on so-called “racial proxies” to determine race. Labeling any resulting algorithm “colorblind” can only be true if one ignores the ways in which many proxy variables are historically constitutive of race itself.

For approaches that assert the fallacy of scrutinizing input variables and instead advocate for a turn to algorithmic outputs, I applaud scholarly skepticism of what is often assumed to be a race-blind algorithm.[46] But I nevertheless identify the outstanding racial questions inherent to scrutinizing outputs, which raise many of the same dilemmas as the identification of racial proxies in input data. In particular, I underscore the issues raised when a programmer attempts to draw distinctions between legitimate levels of racial differentiation and algorithmically produced discrimination.[47] Indeed, the question of how much racial differentiation an otherwise inscrutable algorithm is permitted to produce before it is considered discriminatory implicates contestable notions of algorithmic fairness, algorithmic justice, and algorithmic discrimination.[48] But it also allows a programmer to draw distinctions between algorithmically induced racial disparities and those attributable to purportedly inherent racial differences. Yet, if a programmer accepts any level of innate racial difference, they risk reinforcing a mythical belief in the natural or static nature of race and then embedding this racial ideology in the algorithm.

By placing the racial theories permeating these methods at the center of inquiry, Part III reveals how differences in our theoretical analysis of race shape how machine learning algorithms are developed and deployed. The complex, normative assumptions that undergird econometric approaches to racial proxy discrimination open the door to profound racial dilemmas that programmers and regulators must grapple with. Moreover, these approaches are marked by paradoxical and taken-for-granted beliefs about race that have similarly pervaded judicial reasoning about the racial proxy. Drawing a throughline between judicial and statistical patterns of racial reasoning challenges the belief that algorithmic proxy discrimination can be addressed with a technical fix, rather than an explicitly moral one.

Part IV continues a discussion of the material effects of competing racial proxy definitions, but this time focuses on how differences in our analysis of the racial proxy can instigate changes to the racial subject at the center of antidiscrimination law. I examine the recent and closely watched settlement between the U.S. Department of Justice (DOJ) and one of the largest technology conglomerates in the world—Meta Platforms, Inc. (Meta). I demonstrate how the acceptance of spurious and contestable notions of a racial proxy can limit our ability to recognize the full range of consequences associated with machine learning algorithms and, more importantly, regulate the powerful actors that develop and deploy them.[49] In the settlement, the government’s acceptance of a narrow and ambiguous notion of a racial proxy preserves Meta’s ability to construct new classes of people who, on the basis of digitally observable attributes, can be exposed to differentiated privileges, resources, opportunities, and statuses in some of the most legally significant domains of public life: housing, credit, and employment. I argue that such a power should not be in Meta’s control and that those concerned about algorithmic discrimination ought to attend to the novel political categories catalyzed by new information technologies as racialized categories.

The consequences of this novel approach to group formation are yet unknown, as is the evolving relationship between these groupings and those traditionally understood as politically subordinated.[50] But understanding these groupings as evidence of an algorithmically induced racialization allows us to draw on a long-established tradition of scholarly explorations of race. This tradition teaches that if the overarching objective behind the process of group construction remains the pursuit of economic interests, machine learning algorithms threaten to expand the regime of racism by constructing groups and rendering them differentially vulnerable to exclusion, containment, and exploitation. Referred to in various texts as racial formation,[51] racial construction,[52] or racecraft,[53] among others, what unites these ideas is a focus on the political and economic processes that are obscured by the taken-for-granted “major concept” of race or, in this case, the racial proxy.[54] Understanding, contesting, and regulating these processes ought to be a primary concern of legal thought and action.

I. The Dilemma

Because not all blacks in the United States were former slaves, “freedmen” was a decidedly under-inclusive proxy for race.

—Justice Clarence Thomas, Students for Fair Admissions, Inc. v. President & Fellows of Harvard College

What precisely is a racial proxy? That question has long preoccupied legal scholars, whose responses have been both thoughtful and diverse.[55] Indeed, observing the contested meaning of a racial proxy is not a debate freshly inaugurated by new information technologies, and its implications extend beyond the realm of machine learning algorithms. Scholars have long discussed the consequences that so-called proxies for protected categories like race and gender pose for antidiscrimination law.[56]

Constitutional equality law obstructs overt racial classifications in the distribution of government benefits and burdens.[57] Any intentional attempt to classify by race will be subject to the highest level of judicial scrutiny. Key civil rights statutes similarly prohibit racial discrimination by scrutinizing both racial classifications and racially disparate impacts in areas like housing, employment, and credit.[58] Imported to the world of machine learning algorithms, a concern about racial discrimination means that computer programmers will almost always exclude the formal variable of race from an algorithm to avoid running afoul of antidiscrimination law.[59]

For decades, in response to this legal superstructure, legal scholars have emphasized the disconnect between antidiscrimination doctrine and the reality of racial discrimination, which routinely involves differential treatment based on factors closely related to racial identity even as it stops short of making an explicit racial classification.[60] This differential treatment, they argue, even when ostensibly race neutral, leaves a discriminatory residue of the sort that antidiscrimination law ought to address.[61] And some have even connected this legal framework to an impoverished understanding of race and other meaningful categories of constructed difference.[62]

But the adoption of machine learning algorithms to inform decisions across an array of institutional sites intensifies and simultaneously refashions this debate in novel ways. For starters, the proliferation of machine learning algorithms opens the door to a new set of actors and decision-makers who assign a definition to the racial proxy variable. Programmers, developers, and others involved in the development and deployment of machine learning algorithms can now technologically embed their normative choices into algorithms.

Consider, for example, an emerging trend to remove zip code and other geographic indicators from predictive models in the name of racial and economic equality.[63] Developers may remove or restrict geographic indicators from input data without thorough explanation or confrontation with antidiscrimination principles and interests.[64] In doing so, they exercise their powers of “legal entrepreneurship”[65] to define the universe of racial concern and the nature of racial remedies without fully confronting their underlying equality commitments or racial assumptions.

Programmers who remove geographic indicators—while declining to restrict similarly race-correlated variables—exercise great control over highly configurable technological tools, often without the need to rigorously defend their underlying normative criteria for determinations. To be sure, one need not discount geographic segregation, or its historically state-mandated scaffolding,[66] to observe how the à la carte exclusion of racial proxies represents a powerful form of political line drawing. This power to embed highly contestable political decisions into the technology that increasingly structures human life illustrates the importance of interrogating the meaning of the racial proxy in algorithmic antidiscrimination discourse.

Further intensifying discussion about the precise definition of a racial proxy is the fact that without intervention, proxy discrimination is an unavoidable aspect of machine learning technologies.[67] When potentially relevant data is withheld, an algorithm may “seek out” proxies for the omitted data.[68] This tendency flows from the structure of machine learning algorithms. These algorithms are trained on historical data to learn patterns and uncover relationships upon which future decision-making can rely. The patterns and relationships discovered within the data result in an underlying model, which can then be exposed to new data in order to automate certain processes, including predicting the probability of future outcomes.[69]

Even if the formal category of race is withheld from a dataset, the algorithm may increase its reliance on variables that are correlated with race to account for the omitted race variable.[70] Theoretically, this concern is prevalent in human decision-making as well.[71] In the absence of explicit racial information, human decision-makers may rely on perceived racially inflected signifiers to inform their actions. Yet the sheer quantity of variables processed by algorithms sets them apart from human actors as the number of variables (or combination of variables) available for processing increases the number of potential racial proxies.[72] In fact, nearly all algorithmic inputs can be correlated with race in some way.[73]

Of course, not every data point that is correlated with race will trigger concerns of racial discrimination. This is precisely the critical task of identifying impermissible racial proxies. In a world structured by and through racial difference, nearly all recordable data contains some racial significance. Determining which variables correlate enough with race such that their inclusion in a dataset raises concerns about racial discrimination is a political and normative question that frequently evades scrutiny.

Legal scholars skeptical of actuarial predictions consider certain variables problematic proxies for race, including criminal history,[74] neighborhood crime rate, and other economic indicators.[75] Yet explanations of the precise qualities that give these variables a sufficient “racial” character are often dropped from the discursive frame.[76] The risk is that without a clear articulation of what transforms these variables into racial proxies, one might assume that the designations are somehow self-evident, instead of firmly attached to views about race and the meaning of equality.

Within racial proxy discourse, the legal impulse towards colorblindness combined with the proliferation of machine learning algorithms opens the door to new and surprising theoretical dilemmas. When programmers are tasked with systematic removal of not only race, but also racially inflected inputs from the algorithm, they must determine which variables to exclude and why.[77]

To complicate matters further, from a technical and conceptual standpoint, these determinations are far from obvious.[78] As Talia Gillis explains, even input variables that are commonly understood as uncontestable “proxies for race” (such as a person’s zip code) may be less concerning from an algorithmic standpoint than other, less intuitive inputs, or a combination of inputs, in replicating racial information.[79] It is difficult to articulate how exactly a machine learning algorithm generates individual predictions from a set of input variables.[80] This dilemma is usually understood as part of the “black box” nature of many machine learning algorithms, which places any intuitive judgment about the definition of a racial proxy even further out of reach.[81] This raises a predicament: Without consensus around the meaning of a racial proxy or a definitive understanding of the complex interactions between variables, the exclusion of all racial proxies from algorithmic inputs is not feasible.[82]

Nor is excluding all racially correlated variables desirable when considering an algorithm’s utility. Consider that each data input is usually relevant to algorithmic predictions, and excluding data in the name of legal fairness could impact any resulting knowledge produced by the algorithm.[83] Much of the legal and statistical fairness literature suggests an inevitable tradeoff between “fairness” and “accuracy” to describe the impact of excluding inputs.[84] In this framework, legal regimes regulating input variables threaten to degrade the algorithm’s “accuracy” when they prevent the algorithm from accessing all forms of relevant data.[85]

Taken together, these concerns throw into sharp relief the impracticability of legal conceptions of the racial proxy that assume that racial proxies must be perfectly correlated with race to raise concern about racial discrimination. In the quote that begins this Section, for example, U.S. Supreme Court Justice Clarence Thomas asserts that the Freedmen’s Bureau Acts—passed in the wake of the Civil War to expanded protection for formerly enslaved people—were race-neutral legislation.[86] The Acts should be considered race neutral, Justice Thomas argues, because not all Black people in the United States at the time the Acts were passed were formerly enslaved.[87] That a category as racialized as being formerly enslaved in the post–Civil War United States would be considered race neutral demonstrates the ways in which racial proxy determinations are highly vulnerable to political distortions. In this case, Justice Thomas leaned on this bewildering racial proxy determination to defend a colorblind reading of the Fourteenth Amendment in originalist terms.[88]

Presumably, by this normative reasoning, only a variable that is perfectly correlated with race could be identified as an impermissible racial proxy. Given without explanation or defense, this highly contestable vision of the racial proxy effectively contemplates no other role for courts than to determine the fact of perfect racial correlation. Yet, a perfectly race-correlated variable may not exist. Moreover, because there is often no way to tell if highly race-correlated variables are nevertheless reconstructing the effect of the omitted race variable, excluding only perfectly race-correlated variables from input data may not be enough to render an algorithm race blind.

Yet, in presenting his dubious definition of a racial proxy as a statement of fact rather than as a product of political reasoning, Justice Thomas’s opinion can hardly be considered an oddity in law. Even as judges deviate from Justice Thomas’s narrow and formalistic ideas about perfect racial correlation, they rarely defend their racial proxy determinations with respect to antidiscrimination values, commitments, and principles.

II. Legal Constructions of the Racial Proxy

Choosing not to acknowledge the cause and effect of racial power and adopting a vow of silence is one response, but is no more effective than a child closing his eyes in order to make himself invisible.

—Cheryl Harris, Critical Race Studies: An Introduction

In finalizing the legal definition of a racial proxy, just as in many other crucial choices relevant to the development of machine learning algorithms, there is no strict judicial guidance.[89] Courts have not provided a single theoretical baseline to determine whether any given feature is a proxy for race. Paradoxically, the lack of a precise definition has not prevented judges from invoking racial proxies to achieve some outcome within the framework of American equality law. In this Section, I explain how the irregular and opaque ways that programmers identify the racial proxy variable can be understood as a reflection of a racial commonsense approach internal to the very legal regimes governing antidiscrimination law.

Particularly within the framework of colorblind antidiscrimination law, where any decision-making based on race is highly scrutinized as deleterious,[90] the racial proxy is a powerful technology. It represents a relationship between race and another variable, and seizing on this relationship permits judges to harness what is arguably the most forceful prohibition in American antidiscrimination law. Indeed, because many actions based on race are considered injurious—regardless of their ideological origins[91]—evoking a racial proxy permits legal actors to command judicial outcomes in the name of equality, while ostensibly avoiding moral reasoning or meaningful confrontation with principles of equality and antidiscrimination.

To capture a racial proxy’s legal effect, however, judges must relate a given variable to race, and this has proved a vexing task. Judges appear conflicted about the relationship between race and a racial proxy, leading to contradictions and inconsistencies as they struggle to articulate precisely what gives a variable its racial quality. For example, courts might decide to eliminate prior criminal arrests from state predictions of an individual’s risk of missing a future court date on the grounds that such a variable is a proxy for race. But a court need not make the same decision for other variables that are, by the same measure, heavily race correlated.[92] To reach such an internally inconsistent outcome, however, requires some path of reasoned thought. To clear this path, judges could draw on factors like time, place, and history to explain why some race-correlated variables are identified as racial proxies, while others are not. But, as this Section demonstrates, judges often enlist taken-for-granted beliefs about the relationship between race and the racial proxy, gesturing to a presumably self-evident connection to justify their theoretical leaps and smooth over their conceptual discontinuities.

Courts often frame the relationship between race and the racial proxy as clearly apparent and in no need of political explanation. This framing avoids deeper engagement with the racial proxy’s inherently normative and contingent nature. The resulting opinions unfold in inconsistent, underexplained, and at times paradoxical directions.

The cases examined below demonstrate this phenomenon by introducing some of the varying and internally inconsistent definitions of a racial proxy found in judicial reasoning. These cases span the voting, criminal justice, and employment contexts, and the racial proxies themselves vary as courts consider whether geography, ancestry, criminal arrest history, and hair texture and style constitute racial proxies. The following examples are not meant to be an exhaustive presentation of how judges reference racial proxies in their judicial opinions; instead, these cases serve as examples of a pattern of judicial reasoning in which judges engage with the concept of a racial proxy but rely on definitions they understand to be self-evident or obvious. Not only do these taken-for-granted judgements lead to internally inconsistent results, but they also underscore the broad authority given to an algorithm’s developers to select a definition of the racial proxy intuitively from the many alternatives. From a technical standpoint, the trouble is that relying on their racial common sense represents the very genus of intuitive reasoning that a programmer cannot apply in the machine learning context. Moreover, the ability to make these contestable and significant racial determinations often falls to powerful, profit-driven companies, systems, and actors that develop and deploy algorithms.

The cases discussed below throw into sharp relief a tradition of legal thought that has struggled with the identification of a racial proxy and has therefore presaged this digital era moment of extreme epistemic deference to the racial common sense of an algorithm’s developers. Long before these variables gained significance in algorithmic antidiscrimination discourse, judges struggled to identify the precise relationship between race and a racial proxy.

A. A Racial Proxy as Suspect and Pejorative

Harold Rice was a White man born and raised in Hawai‘i.[93] He was a descendent of missionaries and ranchers who migrated to the Hawaiian Islands in the 1800s.[94] According to the Supreme Court, Rice was a citizen of Hawai‘i “in a well-accepted sense of the term.”[95] He did not, however, have the requisite ancestry to vote in the election for trustees of the Office of Hawaiian Affairs (OHA), a state agency created to manage land and resources reserved for the benefit of Indigenous Hawaiians.[96] Under state law, the only people who could vote in OHA elections were “Hawaiians,” defined as descendants of people who inhabited the Hawaiian Islands in 1778 and continued to live there.[97] Rice challenged his exclusion from voting for OHA trustees under the Fourteenth and Fifteenth Amendments.[98] After losing his case in lower courts, Rice prevailed at the U.S. Supreme Court when the Court held that the State of Hawai‘i had used ancestry as an impermissible “proxy for race” to determine voting qualification, in violation of the Fifteenth Amendment.[99]

On its face, the OHA’s voting plan made no racial reference. The reference to ancestry was designed to capture indigeneity, and it was crucial to the State’s attempts to enshrine protections for Indigenous people by demarcating the borders of the Indigenous group.[100] To grant descendants of Indigenous Hawaiians strategic and political rights to exercise control over their ancestral lands and resources, the State used ancestry to identify the descendants of people living in what is now the State of Hawai‘i prior to its colonization.[101]

The Rice Court,however, reasoned that the State was using ancestry as a “proxy for race” in violation of the Fifteenth Amendment, which forbids states from denying or abridging the right to vote on account of race.[102] It likened Hawai‘i’s statute to Jim Crow–era voting restrictions that used White ancestry to identify voters who would be exempt from literacy tests.[103] In both cases, the Court reasoned, ancestry was used to perpetuate a “racial” exclusion in voting without making an explicit reference to race.

In what sense, then, was the State’s reference to ancestry “racial”? According to the Court, ancestry exhibited a racial quality because conditioning voting rights on ancestral lineage was effectively an attempt to condition these rights on historical belonging to a “race” of people.[104] As the Court wrote, early Hawaiians were a “distinct people, commanding their own recognition and respect,” and the statute evinced “the State’s effort to preserve that commonality [of people] to the present day.”[105]

The State raised several issues with this conceptualization of a racial proxy, including that the voting restriction was a classification “limited to those whose ancestors were in [Hawai‘i] at a particular time, regardless of their race.”[106] The State asserted that in 1778, Hawai‘i was not a racially homogenous place, the Hawaiian Islands were inhabited by an ethnically diverse population that may have migrated from the Marquesas Islands, the Pacific Northwest, and Tahiti.[107] If the original inhabitants of the Hawaiian Islands were not a monolithic race of people, then one could not assert that ancestry was a tool of racial preservation.

Further, the OHA law barred voting for any “person whose traceable ancestors were exclusively Polynesian if none of those ancestors resided in Hawaii in 1778.”[108] Conversely, a White person who could only trace, say, “one sixty-fourth” of their ancestry to people who inhabited the Islands in 1778 would be permitted to vote.[109] This was a crucial point of contention for the State: that race and ancestry do not completely overlap. Some White Hawaiians could satisfy the ancestry requirement, and some non-White Hawaiians could not.

Thus, a rightful dispute emerged about the discontinuities between ancestry, a term referencing questions of descendance, and race, which the Court used to reference a people with common physical characteristics and a common culture.[110] Hardly a statement of fact, the Supreme Court’s insistence that ancestry amounted to a racial proxy necessitated a defense—one that explained why ancestry was a proxy for race despite its reference to a multiracial group of Indigenous inhabitants and its production of a multiracial voter block. But the Court’s defense of its proxy designation drew instead on ideas about a presumably self-evident relationship between race and the racial proxy.

While the Court acknowledged that ancestry was not a perfect substitute for race,[111] it simultaneously held that any incongruence between the variables did not defeat the racial proxy designation.[112] Justice Anthony Kennedy, who delivered the opinion of the Court, wrote that “[s]imply because a class defined by ancestry does not include all members of the race does not suffice to make the classification race neutral.”[113] Rather, ancestry was a proxy for race because classification by race and classification by ancestry evinced the same suspect “purpose and operation.”[114] In the Court’s eyes, an inquiry into both race and ancestry “demeans the dignity and worth of a person” and is inconsistent with “respect based on the unique personality each of us possesses.”[115] This understanding of a racial proxy sees any inquiry into race as fundamentally pejorative, and it is this pejorative information that is captured by the variable of ancestry.

Yet this reasoning still begs a question internal to the Court’s reasoning: If the Court sees race as pejorative and sees ancestry and race as unidentical variables, what information does ancestry capture beyond pejorative race? What nonracial information did ancestry reflect, and why was this information insufficient to overcome suspicion stemming from the use of ancestry in determining voting rights? These questions were taken up in a dissent authored by Justice John Paul Stevens and joined by Justice Ruth Bader Ginsburg.

For the dissenting Justices, ancestry could be understood as a proxy for race, yet it “by no means follows that ancestry is always a proxy for race.”[116] What distinguished the ancestry requirement at issue in Rice v. Cayetano from the literacy requirement at issue in Guinn v. United States[117]were “the realities of time, place, and history behind the voting restrictions being tested.”[118] In Guinn, as part of a flagrant effort to exclude Black voters, the State of Oklahoma exempted from its literacy requirement people whose ancestors were entitled to vote prior to the enactment of the Fifteenth Amendment.[119] In contrast, the OHA voting requirement attempted to use ancestry to politically empower members of a once sovereign, Indigenous people.[120] In this perspective, an inquiry into race would still be pejorative but the purpose of the OHA classification “exists wholly apart from race.”[121] According to legal scholar Addie Rolnick, the Rice Court failed to see the independent significance of using ancestry as a mechanism to define “the beneficiary class affected by the positions being voted upon.”[122] An inquiry into ancestry was meant to capture indigeneity, which was relevant in determining who should qualify for voting rights.[123] Over valid critique about the imperfect correlations between race and ancestry, the Supreme Court simply declared, without full explanation, that “[a]ncestry can be a proxy for race. It is that proxy here.”[124]

To be sure, not even the dissent explained precisely how much racial information must be captured to designate a variable a proxy for race, an omission that may signal the risks it assumed by accepting that there is such a thing as a racial proxy. But rather than embrace the false equivalence of the ancestry requirements in Guinn with those instituted by the State of Hawai‘i in Rice, the dissent engaged with their opposing ideological purposes, suggesting that one can only identify the “racial” quality of proxy variables by referencing historical and societal concerns.

In contrast, Justice Kennedy’s opinion cast ancestry as undeniably “racial,” demonstrating judicial investment in a particular conception of the racial proxy as presumptively normatively irrelevant and existing entirely apart from the process of political line drawing in which the Court itself is engaged. In the context of colorblind constitutionalism, therefore, the racial proxy becomes a powerful discursive tool that permits judges to obscure their contestable normative claims, framing them as self-evident statements of fact and, in Rice, facilitating the erosion of Indigenous legal claims to self-determination and political consciousness.[125]

B. A Racial Proxy as Highly Correlated with Race

In September 2017, the Cook County Circuit Court implemented new pretrial release policies intended to reform cash bail.[126] The Cook County Sheriff, Thomas Dart, disagreed with the revised policies.[127] Acting as penal policymaker, Dart began to conduct his own administrative review of the court’s bail orders, at times substituting his assessment for the court’s and refusing to release individuals who were granted bail by a judge.[128]

The plaintiffs in Williams v. Dart were a group of Black defendants, each charged with felonies and granted release on bail by a judge, only to have the release order ignored by the Cook County Sheriff based on his own “independent reviews.”[129] In conducting his rogue assessments, the Sheriff considered factors like the defendants’ arrest history, the charges they faced, and their neighborhood to determine whether he would comply with a judge’s order granting their release on bail.[130] The plaintiffs challenged this practice through several federal and state law claims, including a claim under the Equal Protection Clause of the Fourteenth Amendment.[131] They argued that the Sheriff targeted them for continued detention because of their race and that the Sheriff’s policy “disproportionately targets African Americans by using charge, prior arrests, and neighborhood to determine eligibility for release.”[132] The result was that out of more than eighty people detained despite a court order directing their release, four in five were Black.[133]

Reversing the lower court’s holding that the plaintiffs did not claim a “plausible, nonconclusory allegations of intentional discrimination,” the Seventh Circuit held that the allegations of racial discrimination in the plaintiffs’ complaint were sufficient to meet the pleading standard.[134] Even though the variables considered by the Sheriff’s bail review policy were facially race neutral, the circuit court wrote that the policy could still raise an “inference of impermissible intent” by relying on criteria that “map[ped] so closely onto racial divisions that they allow racial targeting ‘with almost surgical precision.’”[135]

Although the decision was a step toward vindicating the constitutional rights of the defendants, the notion that a variable’s racial quality stems from its so-called surgically precise overlap with race is hardly a workable standard. Many collectable data points have high correlations with race, and determining which high correlations matter and why is the normative task at hand for any decision-maker, be they judge or programmer. What transformed these highly race-correlated variables into racial proxies?

In criteria reminiscent of a racial commonsense standard, the court took judicial notice that Chicago, the home of all nine plaintiffs, consistently ranked among the most racially segregated cities in the nation, and therefore, “neighborhood was a plausible proxy for race.”[136] In reaching this conclusion, the court cited a previous ruling that relied on evidence of a statistical correlation between race and geography in Chicago, as well as a magazine exposé on the economic pressures exerted on low-income residents of urban metropolises like Chicago.[137]

The court concluded that arrest history is another proxy for race, “in light of Chicago’s alleged history of disproportionately arresting African Americans,”[138] which the court noted was an “allegation endorsed by the U.S. Department of Justice in 2017.”[139] Yet, when it came to the variable of pending criminal charges, the court wrote that although this variable may have been another racial proxy, the precise proxy mechanism was unexplained. The court stated that neighborhood and arrest history were enough to conclude that there was at least a plausible, nonconclusory allegation of intentional discrimination.[140] The court’s hesitation to cite to similar studies supporting the correlation between criminal charge and race warrants some scrutiny for its inconsistency with findings present elsewhere in the opinion.

Racial disparities in charging disposition are a well-documented concern in the criminal legal system[141] such that some states have chosen to respond legislatively.[142] If the Williams court understood statistical evidence or common knowledge as criteria for establishing certain variables as racial proxies, then the court’s hesitation to label a criminal charge as a proxy for race appears inconsistent. Why not use the same criteria—judicial notice or scientific evidence—to label other, race-correlated variables racial proxies? Such internal inconsistency reveals how the transformation of a given feature into a racial proxy, a transformation that raises the specter of impermissible intent, is articulated by courts in a contingent manner. The problem is that such contingency is often obscured in judicial opinions where the borders demarcating a racial proxy are depicted without reference to moral commitments or deeper engagement with time, place, and history.

Williams v. Dart demonstrates how judicial identification of racial proxies frequently avoids the very normative reasoning that is inherent to any interpretation of a racial proxy. This avoidance, however, results in internal inconsistencies, such as the exclusion of some highly race-correlated variables and the unexplained preservation of others.[143] With aspects reminiscent of federal cases determining the boundaries of racial categories, Williams v. Dart exemplifies how judges can selectively and actively construct the racial proxies they claim merely to identify.[144] Even within the same opinion, courts struggle with the question of what gives a proxy variable enough “racial” character such that its inclusion in state decision-making would run afoul of legal commitments to equal racial treatment. While the necessary correlation between race and the proxy variables may appear self-evident, a more thorough reading of judicial reasoning reveals the malleability at the core of these judicial determinations.

C. A Racial Proxy as Immutable

In another case, Equal Employment Opportunity Commission v. Catastrophe Management Solutions, the Equal Employment Opportunity Commission (EEOC) filed suit on behalf of a Black female job applicant whose offer of employment was rescinded when she refused to cut her hair in order to comply with the company’s formally race-neutral grooming policy.[145] The employee wore her hair in a locked style, and the EEOC argued that the company’s actions constituted discrimination on the basis of the employee’s race in violation of Title VII of the Civil Rights Act of 1964.[146] The Eleventh Circuit affirmed the lower court’s decision in favor of the company, writing that “[t]he EEOC’s allegations—individually or collectively—do not suggest that [the company] used that policy as proxy for intentional racial discrimination.”[147]

The EEOC argued that by discriminating against those with locks, the policy was relying on an expressly protected characteristic, not because hairstyle is expressly protected, but because locks are “directly associated” with the immutable trait of race.[148] Thus, according to the EEOC, it was the close association between race and hairstyle that made the company’s actions constitutionally suspect.[149]

The circuit court acknowledged that the EEOC’s appeal required the court “to consider, at least in part, what ‘race’ encompasses under Title VII.”[150] Race, according to the court’s dictionary sources, referred most commonly to “physical characteristics shared by a group of people and transmitted by their ancestors over time.”[151] The definition left open questions about the precise physical characteristics as well as the number of group members that should hold them in order to label any population a “race.” Indeed, one would be hard pressed to find anyone who considers people with green eyes, blonde hair, or short stature a race of people, even though these are all physical characteristics shared by a group of people and transmitted by their ancestors over time. Put another way, some other quality must turn certain shared physical characteristics into markers of race.

Trading the vexation of the precise for the comfort of the self-evident, the court concluded that it was “not much of a linguistic stretch” to understand that these physical characteristics referred to immutable qualities, that is, those features considered “a matter of birth, and not culture.”[152] Although locks are “historically, physiologically, and culturally associated with . . . race,”[153] the court reasoned that this association did not make them an immutable characteristic of race, nor did it comport with a line of legal precedent that understood Title VII to protect against discrimination based on immutable characteristics.[154] Title VII could possibly protect litigants against discrimination on the basis of Black hair texture, the court suggested, but not against hairstyles, even those highly correlated with race.[155]

Thus, rather than engaging with time, place, and history when reasoning about which physical features demarcate the boundaries of a racial proxy, the court leaned on taken-for-granted beliefs about race to conclude that immutability gives a variable its racial quality. This highly contestable and paradoxical definition was justified by the judges through their decree that such a definition amounted to “not much of a linguistic stretch.”[156] But in erecting the putative boundary between immutability and mutability, the court neglected the inconvenient contrary fact that certain immutable qualities—like eye color or height—are not understood as racial proxies in America’s reigning racial regime.[157]

One glaring impact of the transposition of EEOC v.Catastrophe’s definition of a racial proxy into the algorithmic context is the obvious incongruence between algorithmically driven forms of discrimination and the preoccupation with immutability permeating America’s current equality law regimes.[158] Many of the features that antidiscrimination scholars, computer programmers, and advocates deem impermissible proxies for race fall into the category of seemingly mutable characteristics like zip code, income level, or prior criminal arrest history. If a programmer were to exclude all members of an online group whose mission was to share information about Black hairstyles, this would also likely raise the danger of intentional discrimination. The ability of algorithmic tools to rely on input data not previously available or even comprehensible to human actors raises equally vast technical and conceptual problems for equality and antidiscrimination law.[159]

EEOC v. Catastrophe is a paradigmatic example of how courts, in identifying what gives a proxy variable its racial character, make appeals to the self-evident and the obvious. But one need only scratch the surface of these self-evident criteria to uncover critical contradictions, inconsistencies, and unanswered questions. Despite the court’s insistence, such criteria are rarely ever obvious, unless one acknowledges the connective thread as the act of judicial decree that provides the supplemental quality transforming some immutable features into racial markers. These acts of judicial decree, which represent judges’ normative investments and commitments, are too frequently rendered invisible or inevitable behind the powerful taken-for-granted nature of race.

Yet to anyone acquainted with courts’ jurisprudence of racial classification, it will come as little surprise that courts frame the contingent boundaries of a racial proxy as obvious or uncontestable.[160] Just as courts, witnesses, lawyers, and litigants have not required a consistent and exacting definition of race to “know race when [they] see it,”[161] courts evidently do not require a precise definition of a racial proxy before allowing the idea to shape their decision-making.[162]

Understanding how judges engage with definitional disputes surrounding the racial proxy sheds new light on the capricious determinations made by those who develop and deploy algorithms.[163] Rather than seeing their decisions about what constitutes a racial proxy as ad hoc or unsystematic, we might instead notice the digital era emergence of the same political dilemmas that have permeated the very development of America’s racial jurisprudence.[164] Judicial processes of racial classification and assignment that developed over hundreds of years have not engendered a precise definition of race, nor its relationship to other closely related variables. Situating a developer’s technical decisions about the definition of a racial proxy within a broader set of legal dilemmas makes clear how a new set of actors and technologies are assuming this politically powerful and legally significant adjudicatory role.

If programmers, like judges, draw on an obvious or self-evident relationship between race and the racial proxy, they too can cast their normative claims as ostensibly factual ones, replicating a legal ethos that closes one’s eyes to factors like time, place, and history in formulating the relationship between race and the racial proxies.

III. The Racial Proxy and Emerging Statistical Solutions

[T]here is . . . always something about race left unsaid. Always someone—a constitutive outside—whose very existence the identity of race depends on, and which is absolutely destined to return from its expelled and abjected position outside the signifying field to trouble the dreams of those who are comfortable inside.

—Stuart Hall, Race the Floating Signifier: What More Is There to Say About “Race”?, in Selected Writings on Race and Difference

The cases discussed above, divergent as they are in their reasoning about the racial proxy, point to an underlying commonality: A racial proxy is best understood as referencing the relationship between race and another variable. Within the ideological framework of colorblindness, judges can seize on this relationship to powerfully shape legal outcomes. However, courts struggle to articulate the exact shape and nature of this relationship, such that the jurisprudence of racial proxies results in an uncertain picture of the precise quality giving a variable its racial character.

Recognizing the problem for predictive modeling as one rooted in the indeterminate definition of a racial proxy, scholars approaching the issue from an econometric lens have proffered their own solutions to addressing proxy discrimination in algorithms. Two approaches stand out in this literature for the ways they are taken up by legal scholars. The first proffers a more sophisticated and nuanced statistical approach that orthogonalizes inputs rather than haphazardly removing controversial inputs from a dataset.[165] The second deemphasizes the significance of algorithmic inputs altogether and focuses instead on the outcomes produced by the algorithm.[166]

Unlike judicial solutions, these approaches suggest that a computer programmer may avoid the explicitly normative, definitional issues associated with the racial proxy completely. While both approaches are attractive in this way, this Section demonstrates that a shift away from explicitly defining a racial proxy does not avoid the most vexing normative task of racial proxy formation: determining the relationship between race and the racial proxies.

As it were, both methodologies ascribe to a particular understanding of race and its relationship to proxy variables. Yet this act of ascription is frequently obscured by commonly held but paradoxical ideas about race—ideas which are so widespread they rarely receive scrutiny in discussions of algorithmic discrimination.

This Part analyzes some of the racial assumptions and underexamined conceptual beliefs that undergird these proposals. Drawing on historical accounts of racial formation long cited by critical scholars of race and law, I argue that the racial proxy is properly understood as standing in a constitutive relationship to race, a conceptualization that complicates the very possibility of a “proxy for race.” A small group of scholars studying algorithmic discrimination have aptly approached a discussion of proxy discrimination from this vantage point,[167] and here I explain how this critical understanding of race challenges some of the statistical strategies proposed to address proxy discrimination. Indeed, an account of the racial proxy as constitutive of race constitutes a material challenge to statistical techniques that rely on orthogonalizing inputs to achieve the goals of colorblindness. By uncovering the racial ontologies that are omnipresent but rarely acknowledged in algorithmic antidiscrimination reform, this Part shows how the dream of bypassing normative disputes about the meaning of the racial proxy is continuously troubled by unstated racial assumptions and beliefs.

A. Input-Focused: The Racial Proxy Paradox

The statistical method proposed by economists and by some legal scholars permits a computer programmer to use a relatively simple statistical technique to isolate the “true” relationship between a racial proxy and the outcome the algorithm is trying to predict, where the “true” relationship is the correct coefficient between a racial proxy and the outcome of interest.[168] The key to this statistical solution is to “build on the fact that algorithmic predictions are formed through an estimation step and prediction step.”[169]

In the estimation step, the algorithm considers all inputs, including race and all potential racial proxies in estimating any predictive relationships.[170] At the prediction step, however, when the model is applied to a specific individual, racial information is withheld from the algorithm, with an average value of all races placeholding for the race variable.[171] For example, an algorithm designed to predict the risk of arrest within a certain timeframe will include all input variables in its estimation, including race and potential racial proxies like geographic indicators or arrest history.[172] A programmer’s inclusion of race and all race-correlates allows their algorithm to capture all information deemed relevant to the prediction.[173] After the programmer isolates the correct coefficient, meaning the information contained within the racial proxies that is orthogonal to race, they can then use this model to make the risk prediction, and the threat of discrimination posed by a racial proxy is purportedly removed.[174] Proponents of this approach assert that it can solve many dilemmas associated with the algorithmic racial proxy and deliver a truly race-neutral algorithm—one “purged” of both racial discrimination and proxy discrimination, while preserving the algorithm’s accuracy by including all relevant inputs.[175]

Recognizing again that a racial proxy represents a relationship between race and another variable, however, this Section points to an important conceptual paradox associated with this approach. To be clear, the issue is not one of empirical rigor; rather, the conceptual consideration raised here is that this method relies on a common but paradoxical understanding of race and its relationship to the racial proxy. What makes this orthogonalization possible is the assumption that the variable of race can be understood independently of the racial proxies.

To understand the theoretical paradox upon which this assumption relies, it is first important to understand how the race variable is typically understood and secured in algorithms. Within algorithms, race is typically understood as a personal, subjective quality which can be generated from “individual self-identification, institutional ascription, or both.”[176] Even those critical of formalistic conceptions of algorithmic fairness tend to accept this algorithmic understanding of race, although others have problematized such techniques.[177] My aim here is not to renarrate some of the problems inherent to securing a definition of the race variable, which have been laid out by others.[178] Rather, I highlight the paradoxical reasoning in attempts to isolate the effect of race from the effect of a racial proxy when one secures the race variable through institutional ascription or individual identification.

To begin, stopping an inquiry into race at individual identification or institutional ascription continues to beg the question: How do individuals or institutions know to what race one should ascribe? Indeed, few would accept the idea that individuals get to identify themselves as whatever race they choose, nor would one accept the idea that institutions can take such racial liberties.[179] Accepting individual self-identification or institutional ascription as the ground truth of race is possible not because one should be free to select their race; on the contrary, the myth of racial freedom ignites notions of racial theft, appropriation, and trespass.[180] Individual identification or institutional ascription is acceptable as a way to determine the race variable only because of an underlying assumption that individuals will not or cannot lie about their race. In other words, the assumption is that they will or must tell the truth about race based on an authoritative background system of racial knowledge.

It is this authoritative background system of racial knowledge that is represented when one retrieves the race variable through self-identification or institutional ascription.[181] In this perspective, race is not merely a mathematical value garnered from personal or institutional viewpoints. Rather, the race variable is itself a proxy, standing in for the taken-for-granted system of social meaning that powerfully shapes individual and institutional actions and beliefs about racial identification.

This is not an argument against securing the race variable via the techniques of individual identification or institutional ascription—any method of identification will necessarily involve some degree of racial essentialism, which neither can nor should be avoided entirely.[182] The paradox in this technique, however, is that the very beliefs about racial classification that the race variable represents have historically been dependent on the so-called “racial proxies.” That is, rather than exist in some relational posture to race, the racial proxies have historically been the political substrate of race since race itself has been determined by reference to the existence or absence of certain proxy variables.[183]

Isolating racial information from the so-called racial proxies is paradoxical when many of the features that are considered racial proxies by courts, computer programmers, scholars, and advocates—features such as geographic area,[184] hair texture,[185] or ancestry[186]—are the very features that have historically been used to assign people to their proper race in the first place.[187]

Take for example the contested racial assignment cases of the nineteenth century, where courts determined race by looking to morphological features including hair texture or nose shape.[188] These features were not simply “proxies for race” based on the notion that they exhibit a high correlation with some understanding of race already at work in the world. Instead, these variables have historically been constitutive of race itself, meaning a person’s race has turned on the presence or absence of these variables. Geographic indicators have similarly been used to assign individuals to racial groups when a person’s “appropriate” racial category was otherwise ambiguous. Writing about the logic that informed racial ascription in the late nineteenth century census, Angela James explains,

enumerators were instructed to use the socially dependent criteria of residential community to classify “half-breed” Indians. So, if such persons lived with Whites, and had the “habits of life” of Whites, they should be classified as Whites. If however, they lived among Indians they were to be classified as Indian.[189]

The constitutive nature of race and the racial proxies call into question the core framework of statistical methods that suggest the possibility of isolating race from its so-called proxies like geography or hair. Historically, these proxies have been embedded in the political marrow of racial classifications, not by some natural force or biological essence, but by repeated acts of legal enforcement.

Rather than adjudicating the legal position of already existing and coherent racial categories, law plays a central role in shaping the boundaries of the categories themselves. This phenomenon is aptly established by legal scholars who demonstrate how judges, lawyers, juries, and litigants—relying on what they deemed common knowledge or scientific evidence—determined racial identity for a host of litigants based on spurious legal rationale that made reference to the presence or absence of certain proxy features such as ancestry or religion.[190]

Recalling again the way the circuit court framed the question of how to racially classify Asian Indians in United States v. Thind,[191] the judges who were authorized to make racial determinations were confused about how to weigh proxies like “class and caste, blood and birthplace, and even religion” when identifying race.[192] The Supreme Court’s turn towards racial common sense as the deciding criteria in Thind worked to stabilize an inherently unstable racial ideology with a veneer of cohesion,[193] in ways similar to the courts’ holdings in Rice v. Cayetano, Williams v. Dart, and EEOC v. Catastrophe. In each of these cases judges cast their political understandings of race as natural or self-evident, and these determinations were critical to upholding or ameliorating the stratification of political, economic, and social life.

Whether a person was subject to freedom,[194] whether they could enjoy the privileges of citizenship,[195] whether they could testify in court,[196] or whether they could attend a school of their choice[197] have at some point in America’s legal history turned on the legal determination of a person’s race. Through powerful acts of legal decree and everyday performances of ritual enforcement, the illusion of fixed, inherent, or natural races emerged to rationalize and maintain a society organized by and through racial hierarchy.[198]

Seen through this perspective, the very concept of a racial proxy variable is paradoxical since the notion redundantly places race and another variable into a false relationship of “proxy.” The historically constitutive relationship between race and racial proxy variables reveals the contradictions in attempts to leach racial information from the racial proxies. While this statistical method appears to bypass definitional disputes concerning the racial proxy, the normative selection is nevertheless revealed by the decision to place race and a racial proxy into an ahistorical and decontextualized relation of proxy. Only by ignoring the historically constitutive relationship between race and its so-called proxies can this method claim to deliver a race-neutral algorithm.

B. Output-Focused: Racial Ontologies in Algorithmic Outcomes

Some scholars bringing an econometric perspective to the question of algorithmic racial proxy discrimination have similarly cautioned against the “algorithmic myth of colorblindness.”[199] Simulating algorithmic credit pricing using a rich dataset of mortgage loans, Talia Gillis demonstrates some of the problems with excluding, limiting, or otherwise scrutinizing algorithmic inputs as an approach to addressing algorithmic racial discrimination.[200]

When discussing the prohibition of race and racial proxy variables, Gillis shows how, in the big data context, the complex interactions between inputs mean one cannot rely on a racial common sense to determine which variables serve as racial proxies in a dataset.[201] Using this point to reinforce a broader caution against regulatory acquiescence to an “input fallacy,” Gillis urges a turn toward scrutinizing algorithmic outputs. When one cannot be sure about the racial consequences of algorithmic inputs, attention to racialized outputs may be the only workable solution to address racial proxy discrimination.[202] At the very least, scrutinizing algorithmic outputs renders certain normative decisions more transparent.

For Gillis, legally relevant questions—such as the question of which individuals are “similarly situated” and should therefore be treated similarly by the algorithm—can be freshly examined by looking to outputs.[203] Scrutinizing algorithmic outputs also permits regulators to examine how the algorithm performs compared to human decision-makers when evaluating protected groups,[204] another metric of algorithmic bias that occupies a prominent place in contemporary discourse.[205]

While these output-based assessments of discrimination bypass the difficulty inherent to identifying racial proxies in input data, they are hardly an escape from the racial reasoning that makes input scrutiny a difficult task for computer programmers. Consider for example that any scrutinization of outputs evidently returns the programmer back to issues which legal scholars have been keenly attentive to. That is, the lack of consensus over the meaning of key, politically relevant concepts like algorithmic fairness, algorithmic equality, and algorithmic bias within antidiscrimination discourse.[206] This is a point that Gillis acknowledges.[207]

Any approach to regulating algorithmic outcomes necessarily requires selecting definitions of algorithmic equality, discrimination, and bias from the many competing definitions that exist.[208] While legal scholars and computer scientists have been acutely aware of the normative disputes behind measurements of algorithmic fairness,[209] the underlying racial ontologies that become embedded in algorithmic outputs as a result of these disputes are rarely identified as part of the terrain of normative contest.

Take, for example, the question of whether an algorithm’s disparate racial impact must reach a certain magnitude before violating civil rights law, a question that has preoccupied legal scholars and that directly implicates theories of antidiscrimination law.[210] To be sure, this is a question about the normative meaning of algorithmic discrimination. But it is also fundamentally a question about how much “race” can be said to explain algorithmically determined differences between groups once a programmer has presumably accounted for every other variable relevant to the outcome.

It is here that a programmer must decide what amount of algorithmically produced differentiation can be counted as legitimate racial disparity.[211] Drawing these distinctions between warranted and unwarranted levels of racially disparate impact is an enormous conceptual challenge. If racial disparities in the outcome remain beneath a particular limit, the algorithm is not deemed the source of differentiation. Instead, race is the explanation for the difference. But for a programmer, accepting any level of innate racial difference risks embedding an erroneous ideology of natural, static, or intrinsic racial difference within the algorithm. Hence, the determination of how much racial differentiation is tolerable is one way that human decision-makers technologically embed racial ontologies into algorithms.[212] These embedded ontologies can, in turn, lend the algorithm some degree of explainability.

What I want to suggest is that when algorithmic determinations are otherwise difficult for a programmer to articulate or explain—perhaps due to the self-enhancing nature of the algorithm itself[213]—beliefs and assumptions about how race relates to the algorithmic output can form the basis of a programmer’s explanation of an otherwise inscrutable algorithmic outcome. Their beliefs echo legal constructions of race, where judges’ racial intuitions lend a veneer of coherence to otherwise inscrutable reasoning.[214]

Aspects of this debate are not newly initiated by machine learning algorithms. Legal scholars have long established the fundamentally ambiguous nature of antidiscrimination discourse, describing how judges, scholars, and social movements put forth varying and competing definitions of equality that are later absorbed into law and policy.[215] Each legally absorbed definition contains embedded racial beliefs about the source and meanings of racially disparate outcomes.[216] But just as new information technologies usher in new types of disputes, these technologies also initiate new actors, arenas, and processes that resolve these disputes,[217] and each dispute must be clearly defined as a domain for normative contest and political struggle. Even if a turn to an outcomes-based analysis bypasses the programmer’s need to define a racial proxy in input data, it opens the door to a new, output-oriented genre of racial reasoning for computer programmers.

To be sure, these unanswered questions of antidiscrimination law and racial ontology have not stemmed the tide of algorithmic proliferation. If anything, the law’s capricious understanding of a racial proxy leaves the door open to a breathtaking assertion of state and private power to define a racial proxy within the context of algorithms. That is, against this legal, statistical, and discursive backdrop, the definition of the racial proxy, and other questions much debated in equality law, are currently being answered by a novel list of extra-legal actors engaged in the development and deployment of algorithms. Without attention to the theoretical voids left open in the interaction between law and technology, the technologically embedded meaning of race and other constructed categories of difference, and the economic and political processes that guide this meaning, escape critical scholarly, advocacy, and regulatory attention.[218]

IV. The New Racial Proxy Construction

All that you touch you Change. All that you Change Changes you. The only lasting truth is Change.

—Octavia E. Butler, Parable of the Sower

The current judicial, statistical, and discursive understandings of a racial proxy provide capricious and at times paradoxical conceptual tools to define a racial proxy in input data. But such capriciousness ought not to be confused for randomness, nor should it be allowed to conceal the ways in which political and economic domination work to shape understandings of the algorithmic racial proxy with consequences that are both immediate and yet unknown.

In the interstices of legal reasoning about the definition of the racial proxy is space for various “acts of legal entrepreneurship” taken by powerful political and economic actors.[219] To date, the practical result of courts’ inability to proffer a definition of the racial proxy, as well as a popular confusion about its relationship to race, has been a delegation of epistemic authority over racial proxy determinations to those who develop and deploy machine learning algorithms.

To illustrate this point and its far-reaching consequences, this Section offers an examination of the definition of the racial proxy proffered by state and corporate actors in one of the most closely watched legal disputes about algorithmic racial discrimination, one between the DOJ and Meta, one of the largest multinational technology conglomerates in the world.[220] I examine how the U.S. federal government has ultimately sought to respond to racial proxy discrimination and how it has approached the problem of the indeterminate meaning of a racial proxy.

While remaining sensitive to the ways the processes of defining a racial proxy can vary across algorithmic contexts, a close examination of this case can nevertheless offer a microcosmic glimpse into how misguided conceptions of the racial proxy can both reinforce the maldistribution of benefits and burdens along conventional racial lines, while also facilitating a change to the racial subject at the center of antidiscrimination law.

The federal government’s case against Meta demonstrates how myopic understandings of the relationship between race and a racial proxy result in meaningful limitations on the ability to regulate new information technologies, while simultaneously facilitating the emergence of novel forms of racialization whose consequences have yet to be determined.

A. The Lawsuit Against Meta’s Racial Proxy

In October 2016, an article published in ProPublica revealed the results of an inquiry into Facebook’s, now known as Meta, advertising portal.[221] Journalists entered the company’s portal to purchase housing ads, conducting a quasi-audit for discrimination.[222] Once within the portal, the journalists were able to target their ads by including or excluding from the ad’s eligible audience Facebook users with certain characteristics.[223]

The ability to target ads to particular user characteristics was hardly a new or surprising revelation. Targeted advertisement is core to Meta’s business model.[224] The company maintains an extensive and unprecedented system of mass behavioral surveillance and data collection.[225] Meta tracks information related to individuals’ online activity across the company’s technologies (e.g., Facebook and Instagram), their activities with other websites and cellphone applications, their location data from websites and from phones, and their activity with other businesses.[226] The collected data then fuel Meta’s lucrative business of targeted advertising—nearly $135 billion in revenue in 2023.[227]

While ad targeting based on user characteristics is key to the company’s financial success, the specific characteristics presented to housing advertisers drew media and advocacy scrutiny. Facebook’s ad portal allowed the ProPublica journalists to narrow the eligible audience for their housing ads by excluding users who were classified as having a certain “ethnic affinity” such as African American, Asian American, or Hispanic.[228] Housing advertisers could select from a drop-down menu of these categories and exclude people within these groups from the ad’s audience.[229] The legal question at hand was whether Facebook’s ad targeting and delivery system violated the Fair Housing Act (FHA).[230] Among other prohibitions, the FHA makes it unlawful to:

[M]ake, print, or publish, or cause to be made, printed, or published any notice, statement, or advertisement, with respect to the sale or rental of a dwelling that indicates any preference, limitation, or discrimination based on race, color, religion, sex, [disability], familial status, or national origin, or an intention to make any such preference, limitation, or discrimination.[231]

ProPublica likened the ethnic affinity categories to a racial exclusion option that gave advertisers the Jim Crow–era opportunity to place ads only in newspapers that went to White readers.[232] Civil rights lawyers contended that one would be hard pressed to find such a blatant violation of the FHA prohibition against racial discrimination in housing advertisements, and that allowing advertisers to narrow their advertising pool by ethnic affinity constitutes unlawful racial discrimination.[233] The ProPublica article launched a series of investigations and a lawsuit against Facebook filed by a group of civil rights advocates, National Fair Housing Alliance v. Facebook (NFHA v. Facebook).[234]

In formulating the problem with Facebook’s ad system as one of racial discrimination under the FHA, civil rights advocates relied on their understanding of “ethnic affinity” as an impermissible racial proxy.[235] To grasp why the issue of proxy discrimination is relevant to the dispute, one must understand that the “ethnic affinity” categories that Facebook presented to ad buyers were not direct racial markers in the formal sense.[236] According to Meta, the company did not directly collect racial information.[237] Facebook’s ethnic affinity categories were instead algorithmically constructed conclusions about people’s interests or affinities based on insights culled from the massive trove of data that Facebook collects.[238] As the complaint in NFHA v. Facebook argued, “Facebook extracts data from its users’ online behavior, both on Facebook and off, and uses algorithms designed to sort that data, process it, and repackage it to group potential customers into new and salient categories for advertisers to choose from when targeting their ads.”[239]

In other words, Facebook’s algorithms classify people as belonging to a particular “ethnic affinity” based on data about their online behaviors.[240] The company’s algorithms also create and sort users into hundreds of other novel categories based on their data.[241] Once assembled into these categories—like “parents with toddlers,” or “people with an interest in cooking”—advertisers can then include or exclude users from an ad’s audience based on their membership in the constructed category.[242]

The critique at the core of the NFHA complaint is that some of these algorithmically constructed categories reflect groups protected under the FHA,[243] even though the categories stop short of making explicit racial identifications. From this perspective, these algorithmically constructed groupings can track racial categories with almost surgical precision,[244] making them close proxies or pretexts for constitutionally protected groups.[245]

In November 2016, Meta responded with a statement about Facebook’s “ethnic affinity” ad-targeting tool.[246] Framing the issue as a misuse of Facebook’s technology by bad actors who unlawfully sought to narrow ad audiences, Meta explained that it would disable the use of these ethnic affinity categories.[247] But, in 2018, after advocates reported that the problem persisted, the Assistant Secretary for Fair Housing and Equal Opportunity filed an administrative complaint with the U.S. Department of Housing and Urban Development (HUD).[248] Following discussions with Meta, the DOJ brought an action against Meta in district court in June 2022 for violations of the FHA.[249]

Like the suit filed by civil rights advocates, the DOJ complaint alleged that Facebook facilitated housing discrimination based on “characteristics that are proxies for, or closely related to, FHA-protected categories.”[250] Facebook’s algorithmically constructed categories, like its “multicultural affinity” categories, helped advertisers identify African American, Latino, and White users and exclude them from housing advertisements.[251] According to the DOJ, “Facebook continues to rely upon its vast trove of user data to create the audiences for housing ads. FHA-protected characteristics are related to and encoded within this user data.”[252]

The government’s discursive framing of protected characteristics as “encoded” within facially race-neutral data echoes a long line of antidiscrimination critiques of algorithms discussed within this Article. These critiques suggest race is fundamentally present within purportedly race-neutral data. These so-called “redundant encodings,” which occur when certain data is highly correlated with membership in a legally protected class, have been critiqued by computer scientists and antidiscrimination scholars alike.[253] And as previous sections explain, core to the racial proxy debate is the difficulty in determining which redundant encodings ought to be acceptable and which ones should be rejected.[254]

Importantly, the answer to this core theoretical question is glaringly absent from the government’s complaint. Beyond the assertion that these contentious variables are “closely related” to race,[255] the government appears to sidestep this definitional question altogether, opting instead for a self-evident understanding of a racial proxy. And while the complaint plainly asserts that FHA-protected characteristics such as race are “related to and encoded within” Meta’s massive trove of user data,[256] the specific mechanism and extent to which this racial coding occurs is not explicitly stated.

Without giving an account of the histories of exclusion, segregation, and subordination that form the material basis of these redundant encodings—one that links race with its imagined proxies like zip code and geography[257]—the complaint leaves a significant conceptual void. As I argue below, the definition of the racial proxy that prevails in the case’s eventual settlement constitutes a regulatory conceit to the technology company. This regulatory conceit extends beyond inadequate transparency and disclosure requirements secured within the settlement and touches on the authority of Meta to instantiate new political groupings of great consequence to public life.

B. Meta’s Epistemic Authority over the Racial Proxy

Without proffering a precise definition of a racial proxy, the DOJ entered into a settlement agreement (Settlement) with Meta in June 2022.[258] Meta agreed to eliminate ad targeting options that it determined were “direct descriptors of, or semantically or conceptually related to, a person or group of people based on FHA-Protected Classes.”[259] This means that once Facebook’s algorithms assemble their novel and salient human groupings for ad targeting, the company cannot name these groupings after racial or other protected categories, nor can it name these groupings anything that would appear to describe legally protected groups.[260]

While the Settlement still permits Meta to algorithmically assemble meaningful groupings for targeted advertising, under the Settlement’s terms, Meta must share the names of these new groupings with the government.[261] Using criteria that is not explicated in the agreement, the government will have the opportunity to review and contest these categories based on any “semantic or conceptual” overlaps with race and other protected characteristics.[262]

Despite the government’s ability to observe and contest these potential racial proxies from a descriptive standpoint, Meta’s right to maintain the privacy of its proprietary and confidential information is explicitly protected by the Settlement.[263] The explicit protection of Meta’s right to privacy leaves open the question of whether the government has secured the level of transparency and accountability required to contest how Meta algorithmically assembles its groupings.

Without guaranteeing the ability to contest the myriad human decisions that have shaped Meta’s group-classifying algorithms, the Settlement assumes that the threat of the impermissible racial proxy is purely semantic. The “racial” nature of the “ethnic affinity” proxy is addressed as a highly formalized question of naming a particular category, rather than as the human-driven algorithmic design decisions that such naming represents.

To be sure, a shift in naming practices alone is not entirely beside the point from an antidiscrimination standpoint. Renaming these groupings from “ethnic affinity” categories to categories that are less semantically or conceptually related to race may frustrate attempts by those seeking to intentionally exclude racial groups from online housing advertisements. Without a clear, intuitive, or recognizable racial descriptor, intentional discrimination by housing advertisers may be much harder.

Nevertheless, by defining the relationship between race and a racial proxy as purely semantic, the Settlement improperly gives Meta the authority to use machine learning techniques to assemble distinct groupings of individuals that may then be ordered hierarchically. Yet therein lies the government’s regulatory conceit. It is precisely this process—a process that categorizes and differentiates people for the purpose of exposing them to differentiated opportunities, resources, and privileges—that constitutes the racial proxy in need of legal regulation. And it is this process that should not be within Meta’s control.

To be sure, framing Meta’s authority to algorithmically assemble new and salient classes of people as a process of racialization that ought to be removed from the company’s control turns in part on an understanding of race that is not widespread in law and technology literature.[264] While the idea that race is a sociopolitical construction is accepted in most scholarly circles, the full import of this understanding has yet to inform legal codes or much of legal thought.[265] And it has proven unsuccessful at dislodging commonly held assumptions about the role of race within scientific fields.[266]

While there continue to be those that cynically exploit biological theories of race to advance political-economic agendas, the durability of beliefs about the existence of innate races may also reflect just how salient race is as a sociopolitical construction. One would be hard pressed to point to a corner of American life where the reality of race is not glaringly evident—the sheer material import of race may lead some to believe race must have natural origins. Combine race’s material salience with the reality that scholarship tracing the emergence of race as a historically contingent (often violently coercive) sociopolitical processes[267] is difficult to encounter since such accounts are powerfully suppressed.[268]

Despite this knowledge suppression, a focus on these sociopolitical processes has led scholars of race and technology to conceptualize race itself as a “technology,” one that mediates between individuals and techniques to systemize certain outcomes.[269] Emerging from historically contingent processes, systems of racial classifications have been constantly subject to change and revision as social, economic, and political conditions change.[270] At each step, law has provided the ideological and coercive scaffolding.[271]

Today, legal doctrine strictly scrutinizes actions based on overt racial classifications but has failed to acknowledge or correct the role that law has played in constructing the categories themselves.[272] One encounters strict scrutiny even when groups are classified for the purpose of eliminating racial subordination.[273] Law sees race as the cause of some present or future act of discrimination that must be avoided, but not as evidence of the phenomena of differential treatment and discrimination.[274]

To argue that Meta’s group-assembling algorithm constitutes a process of racialization is to assert that the assignment of a group status, based on attributes that are regarded as unchangeable and inherent, for the purpose of hierarchical human ordering, is a process that touches on the heart of what is meant by the term “racial.”[275] Although Meta’s algorithmically constructed categories may be semantically distinct from race, they still operate in a virtually identical way as race,[276] meaning members of these algorithmic groups are assumed to be meaningfully dissimilar from others by virtue of the presence or absence of certain traits. This presumed difference then becomes the basis for the assignment of differential opportunities, resources, privileges, and statuses within hierarchies of disposability.

While the agreement between Meta and the U.S. government settles the definitional dispute about the racial proxy—defining it as a category with a semantic or conceptual overlap with race—it settles it in such a way as to neglect the myriad value-laden choices that allow Meta to create these distinct classes of people whose differentiation can be accepted as uncontestable or obvious. Permitting Meta to algorithmically configure these groupings protects processes of ad personalization—processes that are practically useful to some Facebook users and that are, without a doubt, immensely lucrative for the company.[277] But it does so at the unknown cost of relinquishing control over the algorithmic construction of categories that can order and organize humans in legally relevant areas of public life, including housing, credit, and employment.

Approaches that allow Meta to self-regulate this process risk acquiescence to larger patterns observed by scholars of technology who demonstrate that when powerful companies in the technology sector are occasionally forced to alter their practices in response to some external opposition, their “executives and engineers produce superficial but tactically effective adaptations that satisfy the immediate demands of government authorities, court rulings, and public opinion.”[278] Yet these adaptations may do nothing to alter the company’s fundamental enjoyment of epistemic deference and domination.

Consider that advertisers who run housing, employment, and credit ads on Meta platforms will now encounter restrictions in the creation of their target audiences. According to Meta, an advertiser will:

not [be] permitted to target based on gender, age, or interests that appear to describe people of a certain race, religion, ethnicity, sexual orientation, disability status, or other protected class. If they opt to target by location, that location targeting must have a minimum 15-mile radius.[279]

Nevertheless, the company’s voluntary disclosures reveal almost nothing about what features “appear to describe people of a certain race,” or how the company selected a 15-mile radius as the minimum location-targeting zone, presumably to eliminate the discriminatory effect of place-based racial proxies. While Meta asserts that “the categories that remain available to these advertisers were the result of in-depth conversations with civil rights stakeholders,”[280] the precise details of this consultation process remain obscured. Therefore, the company maintains virtually exclusive epistemic control over the process of racial proxy selection.

From this perspective, the definition of a racial proxy as “semantically or conceptually” related to race may satisfy the narrow dictates of antidiscrimination law, but it does little to shape the underlying value-laden design decisions that structure Meta’s group-constructing algorithm. In this state of affairs, Meta retains authority over the algorithmically constructed categories that hierarchically order and organize individuals and expose groups of people to divergent opportunities and resources in key areas of legal and policy concern.

By defining a racial proxy as a fixed, semantic choice that can be intuitively recognized, the Settlement confines the antidiscrimination debate to questions of Meta’s phrasing. Yet algorithmic processes of sorting and classifying individuals necessitate a plethora of human decisions beyond naming the resulting groups.[281] Each development and deployment decision eventually determines the shape and nature of Meta’s “ethnic affinity” category, and each decision ought to be available for meaningful scrutiny and challenge by regulators when the outcome of these decisions is to sort individuals into novel, materially significant classifications.

Urgently needed is an approach to algorithmic discrimination that turns attention away from the names ascribed to racial proxy groupings and towards the human and algorithmic decisions required to bring these algorithmically constructed groups into existence and assign them material resources in the world. Indeed, it is their constructed nature—which is done for the purpose of distributing resources asymmetrically across groups—that transforms these algorithmically constructed groups into racial proxies.

The importance of understanding how these different conceptual definitions of a racial proxy can result in vastly different regulatory approaches lies partially in the unpredictable potential of new information technologies as both socially repressive and socially redemptive tools.[282] What Meta’s control over the value-based design decisions that construct race-like categories does for the categories themselves remains an open question. It is, of course, impossible to predict the future. But the technical design decisions overseen by Meta will be shaped by the company’s own interests and agendas.

One can expect that Meta’s algorithmically constructed groupings will be driven by the company’s primary pursuit of profit maximization. In other words, the logic of capital accumulation in the digital age is likely to drive these new and capricious forms of racial common sense because, for companies like Meta, the reputational benefits or legal consequences that stem from demonstrating fair algorithms have not proven to outweigh the commercial risks of self-auditing and disclosure.[283] Therefore, the company acts within its profit-making interests when it constructs these categories in ways that prioritize commercial benefits, rather than prioritizing the prevention of racialized economic opportunity and privilege.

In 2023, advertising drove almost all of Meta’s revenue.[284] While this Settlement institutes a racial reform, it is unrealistic to assume the company will engage in the process of category construction in ways that are not guided by economic concerns like protecting the company’s quest for revenue through the continued extraction and monetization of consumer data.[285]

Through a close examination of Meta’s Settlement, we see how law enables a powerful set of economic actors to drive developments in racial formation. This is the digital era instantiation of a historical process. As the law—a major and meaningful agent of racial production—began to enshrine the ideology of race and White supremacy into the American colonies, it was fueled by the real and immediate quest to solidify economic structures and secure profound economic gain. As Barbara and Karen Fields write,

[p]ractical needs—the need to clarify the property rights of slaveholders and the need to discourage free people from fraternizing with slaves—called forth the law. And once practical needs of this sort are ritualized often enough either as conforming behavior or as punishment for non-conforming behavior, they acquire an ideological rationale that explains to those who take part in the ritual why it is both automatic and natural to do so.[286]

The ideological rationale to which the Fields refer is race itself, a powerful technology developed to reconcile a complex and contradictory social and political reality in which the enslavement of African descendants stood as an “anomalous exception” to an otherwise avowed commitment to human liberty.[287] Part of the power of racial ideology is that it assumes its own aura of normality by virtue of its reference to an embodied “underlying and unchangeable essence” that makes racial borders seem otherwise unpassable.[288]

When vision is clouded by the ideology of race, it becomes difficult to recognize that Meta’s presumably fixed categorical borders—those based on data culled from online activity instead of skin, hair, or bone—are in fact porous, contestable, and actively constructed.[289] In the interstices left open by the Settlement’s definition of the racial proxy, Meta retains the power over novel forms of racial formation.

C. Meta’s Output-Focused Approach in Practice

To be sure, a shift in semantic and conceptual framing is not all Meta intends to do to comply with the FHA. The company also intends to deploy a new system—the Variance Reduction System (VRS)—to reduce the variance between an ad’s eligible audience and the audience who ultimately sees the ad.[290]

Broadly speaking, the VRS uses reinforced learning, a type of machine learning, to achieve a predefined outcome. In this case, the system is instructed to minimize the variance in advertisement impressions across demographic groups.[291] Once a housing advertiser selects a target audience for its advertisement, the eligible demographic ratio for the ad is determined. Then, as the ad is being delivered, the VRS will measure the proportion of impressions of that ad across demographic subgroups, including gender, age, and estimated race.[292] Since Meta contends that it does not collect user information about race, race here is estimated using a method known as Bayesian Improved Surname and Geocoding (BISG).[293] Broadly speaking, BISG estimates racial identity using name and location data pulled from the U.S. Census.[294] If the VRS determines that an ad is being shown in a manner that does not reflect the demographic distribution of the eligible audience, a controller system will shift the distribution of the ad’s impressions toward the eligible ratio.[295]

Although the move toward a more outcomes-based test as the primary mechanism for detecting algorithmic racial discrimination is a constructive one, it does not adequately put to rest disputes about racial meaning and effect. Many important disputes about the meaning of algorithmic fairness and the effect of race on a given outcome will ultimately be technically embedded in the VRS. These disputes will be resolved by Meta’s design choices with some engagement from the government. As Meta explains:

Across the industry, approaches to algorithmic fairness are still evolving, particularly as it relates to digital advertising. But we know we cannot wait for consensus to make progress in addressing important concerns about the potential for discrimination – especially when it comes to housing, employment, and credit ads, where the enduring effects of historically unequal treatment still have the tendency to shape economic opportunities.[296]

For Meta, the recognition that there is no “consensus” in the “evolving” notions of algorithmic fairness helps fuel its rationale that the company can fill the theoretical void.[297] And fill the void it does by relocating important debates about the meaning of algorithmic discrimination from “open, public participatory processes” to private development choices.[298]

For example, the question of what degree of variance is required before the VRS system should trigger the controller represents its own distinct racial determination about what level of racial disparity is both ethically and technically permissible.[299] As discussed in previous sections, this technical choice reflects a theoretical perspective on the degree to which algorithmically produced differential outcomes can be attributed to racial differences. According to the Settlement, Meta and the government will “meet and confer in good faith in an effort to agree on metrics for how much the VRS will reduce any variances.”[300] But the Settlement does not clarify which values, commitments, and principles will guide Meta’s design decisions or the government’s regulatory approach. Nor are these infinitely contestable values left open for public debate, discussion, and dissent.[301]

The Settlement also fails to articulate how the VRS’s third-party reviewer—which was suggested by Meta—will engage the public in its review process. Most importantly, the development, deployment, and oversight of the VRS itself highlight the counterintuitive ethics of outsourcing civil rights regulation to machine learning technologies that are designed, owned, and operated by parties accused of civil rights violations.

Ultimately, a study of Meta’s agreement with the U.S. government—particularly its treatment of the racial proxy—highlights what is at stake in the definition of the indeterminate racial proxy. To be sure, at stake are the meanings of contestable terms like algorithmic discrimination, equality, and fairness, but a closer examination reveals that we may also be witnessing Meta’s ability to algorithmically alter the very nature of the political categories that animate these terms. Indeed, regulatory approaches to machine learning algorithms must consider how contested, constitutionally relevant concepts, like intent or colorblindness, translate uneasily to the algorithmic context and can therefore be manipulated by programmers.[302] But there is also a need to understand how the racial subject—the very subject at the center of antidiscrimination law—may similarly be reconstructed by those who develop and deploy machine learning algorithms.

Conclusion

The concern that formally race-blind algorithmic decision-making can perpetuate racial discrimination by relying on racial proxies has become a cornerstone of debates over algorithmic discrimination. In laying the groundwork for a discussion of the theoretical and material implications of this racial proxy debate, this Article has advanced several claims. I have shown how our present jurisprudential terrain gives those who develop and deploy algorithms a powerful adjudicatory role, which was once exclusively reserved for judges—that is, to use their racial intuition to decide what gives a variable its racial quality. Normatively, I argue that the answer to the question of what constitutes a racial proxy requires an explicitly moral and political solution and cannot be resolved with a purely technical fix. Most importantly, what is at stake in the ability to define a racial proxy is the production of new and meaningful classes of individuals that can later be exposed to differing resources, opportunities, subordination, and privilege. Power over this novel process of racial construction ought not to rest with profit-driven technology companies, who can fashion these groups in the cause of their economically lucrative futures. What is urgently needed in both law and technology is an understanding of race as itself a proxy—one imbued with social, historic, and political meaning to be constructed in service of a just human future.

‍ ‍

Copyright © 2026 Fanna Gamal, Assistant Professor of Law, UCLA School of Law. This Article is indebted to the wisdom and generosity of many people. For your time and extraordinary insights, thank you Angela P. Harris, Issa Kohler-Hausmann, Jerry Kang, Jessica Eaglin, Dorothy Roberts, Noah Zatz, Cheryl Harris, Joseph Fishkin, Mark McKenna, and Pauline Kim. This Article benefitted greatly from thoughtful comments, feedback, and conversations in faculty workshops at UC Davis School of Law, UC Berkeley School of Law, UCLA School of Law, UC Irvine School of Law, and University of Pennsylvania Carey School of Law. Many thanks are due to the Privacy Law Scholars Conference for incubating this work and for awarding it the Reidenberg-Kerr Award for Outstanding Scholarship by a Junior Scholar. For invaluable research support, I thank Nicola Haubold Sanz De Santamari, Jada Evans, Sherry Leysen, Brian Raphael, and Jonathan Rogers. The editors of the California Law Review provided exceptional editing and feedback. This research was made possible in part by support from the UCLA Initiative to Study Hate and the UCLA Luskin Institute on Inequality and Democracy. The Mesa Writer’s Refuge provided the gift of time and space to complete this project. Kyle Halle-Erby, thank you for introducing me to this literature all those years ago. This Article is dedicated to my brothers, Zeine and Jamal.

[1].     In re Thind, 268 F. 683 (D. Or. 1920).

[2].     An Act to Establish an Uniform Rule of Naturalization, ch. 3, § 1, 1 Stat. 103 (1790) (repealed 1795); An Act to Amend the Naturalization Laws and to Punish Crimes Against the Same, and for Other Purposes., ch. 254, § 7, 16 Stat. 254 (1870).

[3].     United States v. Thind, 261 U.S. 204 (1923).

[4].     Id. at 206.

[5].     Id. at 215 (“It is a matter of familiar observation and knowledge that the physical group characteristics of the Hindus render them readily distinguishable from the various groups of persons in this country commonly recognized as white.”).

[6].     See Ian Haney López, White By Law: The Legal Construction of Race 62 (10th anniversary ed. 1996) (“The language betrays entrenched beliefs about the racial significance of class and caste, blood and birthplace, and even religion in establishing racial identity.”).

[7].     See generally id. at78–108.

[8].     See David Lehr & Paul Ohm, Playing with the Data: What Legal Scholars Should Know About Machine Learning, 51 U.C. Davis L. Rev. 653, 670 (2017).

[9].     Id. at 671.

[10].     Id.

[11].     Crystal S. Yang & Will Dobbie, Equal Protection Under Algorithms: A New Statistical and Legal Framework,119 Mich. L. Rev. 291, 297–98 (2020) (“Surveying the field, we find that all commonly used predictive algorithms exclude race as an input. The universal exclusion of race as an algorithmic input is unsurprising given the mainstream legal view that the direct use of race as an input would be unconstitutional.”).

[12].     Id.

[13].     See Cynthia Dwork, Moritz Hardt, Toniann Pitassi, Omer Reingold & Richard Zemel, Fairness Through Awareness, in ITCS ‘12: Proceedings of the 3rd Innovations in Theoretical Computer Science Conference 214, 218 (2012) (referring to “fairness through blindness”); Lily Morse, Mike Horia M. Tedorescu, Yazeed Awwad & Gerald C. Kane, Do the Ends Justify the Means? Variation in the Distributive and Procedural Fairness of Machine Learning Algorithms, 181 J. Bus. Ethics1083, 1087 (2022); Amina A. Abdu, Irene V. Pasquetto & Abigail Z. Jacobs, An Empirical Analysis of Racial Categories in the Algorithmic Fairness Literature, in FAccT 2023: Proceedings of the 6th ACM Conference on Fairness, Accountability, and Transparency 1324, 1332 (2023) (explaining that ethical considerations in technical research and AI are influenced by legal compliance among other aspects).

[14].     Arvind Narayanan, Tutorial: 21 Fairness Definitions and Their Politics, YouTube (Mar. 1, 2018), https://youtu.be/jIXIuYdnyyk (on file with the California Law Review).

[15].     Abdu, Pasquetto & Jacobs, supra note 13, at 1324; Pauline T. Kim, Race-Aware Algorithms: Fairness, Nondiscrimination and Affirmative Action, 110 Calif. L. Rev. 1539, 1545 (2022) (“[E]fforts to make a model less biased could involve taking race into account in many different ways. Exactly when and how a given de-biasing strategy does so is critically important for judging its legality.”); Yang & Dobbie, supra note 11, at 297–99; Solon Barocas & Andrew D. Selbst, Big Data’s Disparate Impact, 104 Calif. L. Rev. 671, 695 (2016); Sonja B. Starr, Evidence-Based Sentencing and the Scientific Rationalization of Discrimination, 66 Stan. L. Rev. 803, 817–21 (2014); Anya E.R. Prince & Daniel Schwarcz, Proxy Discrimination in the Age of Artificial Intelligence and Big Data, 105 Iowa L. Rev. 1257, 1265–67 (2020).

[16].     See, e.g.,Bernard E. Harcourt, Risk as a Proxy for Race: The Dangers of Risk Assessment, 27 Fed. Sent’g Rep. 237, 238 (2015); Barocas & Selbst, supra note 15, at 695; Starr, supra note 15, at 838; Prince & Schwarcz, supra note 15, at 1260–63.

[17].     See Betsy Anne Williams, Catherine F. Brooks & Yotam Shmargad, How Algorithms Discriminate Based on Data They Lack: Challenges, Solutions, and Policy Implications, 8 J. Info. Pol’y 78, 83–85, 89 (2018).

[18].     See Talia B. Gillis & Jann L. Spiess, Big Data and Discrimination, 86 U. Chi. L. Rev. 459, 469 (“In very high-dimensional data, and when complex, highly nonlinear prediction functions are used, this problem that one input variable can be reconstructed jointly from the other input variables becomes ubiquitous.”); Prince & Schwarcz, supra note 15, at 1273–76; Yang & Dobbie, supra note 11, at 311–18; Kim, supra note 15, at 1546.

[19].     Prince & Schwarcz, supra note 15, at 1260 (“[O]ne of the most important threats to anti-discrimination regimes posed by big data and AI is largely unexplored or misunderstood in the extant legal literature. This is the risk that modern AIs will result in ‘proxy discrimination.’”); see also Pauline T. Kim, Data-Driven Discrimination at Work, 58 Wm. & Mary L. Rev. 857, 898–99 (2017); Aziz Z. Huq, Racial Equity in Algorithmic Criminal Justice, 68 Duke L.J. 1043, 1099–1100 (2019) [hereinafter Huq, Racial Equity]; Dorothy E. Roberts, Digitizing the Carceral State, 132 Harv. L. Rev. 1684, 1719 (2019) (reviewing Virginia Eubanks, Automating Inequality: How High-Tech Tools Profile, Police, and Punish the Poor (1st ed. 2018)); Sandra G. Mayson, Bias In, Bias Out,128 Yale L.J. 2218,2263 (2019); Barocas & Selbst, supra note 15, at 691;Jessica M. Eaglin, Constructing Recidivism Risk, 67 Emory L.J. 59, 95–97 (2017); Harcourt, supra note 16, at 237; Starr, supra note 15, at 821–42; Yang & Dobbie, supra note 11, 311–18; Aziz Z. Huq, Constitutional Rights in the Machine Learning State, 105 Corn. L. Rev. 1875, 1925–26 (2020) [hereinafter Huq, Constitutional Rights].

[20].     Yang & Dobbie, supra note 11; Talia B. Gillis, The Input Fallacy, 106 Minn. L. Rev. 1175 (2022).

[21].     See Talia B. Gillis & Jann L. Spiess, Big Data and Discrimination, 86 U. Chi. L. Rev. 459, 460 (2019) (“Unlike human decision-making, the exclusion of data from consideration can be guaranteed in the algorithmic context. However, forbidding inputs alone does not assure equal pricing and can even increase pricing disparities between protected groups.”).

[22].     See Gillis, supra note 20, at 1192; Yang & Dobbie, supra note 11, at 312, 315–19.

[23].     Michael Carl Tschantz, What is Proxy Discrimination?, in FAccT ‘22: Proceedings of 2022 5th ACM Conference on Fairness, Accountability, and Transparency1993 (2022).

[24].     Gillis, supra note 20, at 1235–36.

[25].     Yang & Dobbie, supra note 11, at 333.

[26].     Barocas & Selbst, supra note 15, at 728 (“Abandoning a belief in the efficacy of procedural solutions leaves policy makers in an awkward position because there is no definite or consensus answer to questions about the fairness of specific outcomes. These need to be worked out on the basis of different normative principles.”).

[27].     Mayson, supra note 19, at 2224.

[28].     Gillis, supra note 20, at 1236 (“If what we are truly interested in is the ability to recover a person’s protected characteristics, intuitive judgments are insufficient to determine which features to exclude. Features that intuitively feel like proxies might correlate less than features that do not feel like proxies.”).

[29].     For background on the capricious and contradictory manner in which courts, litigants, and lawyers construct racial categories, boundaries, and meaning, see generallyHaney López, supra note 6; Ariela J. Gross, What Blood Won’t Tell: A History of Race on Trial in America (2008).

[30].     See Ian Haney López, Intentional Blindness, 87 N.Y.U. L. Rev. 1779, 1783 (2012) (“Colorblindness today applies when a government actor explicitly employs a racial classification. In practice, this covers affirmative action policies and little else. Under colorblindness, the remedial motives behind affirmative action are irrelevant.”).

[31].     Cheryl Harris, Critical Race Studies: An Introduction, 49 UCLA L. Rev. 1215, 1229 (2002).

[32].     See infra Part II.

[33].     Noah D. Zatz, Disparate Impact and the Unity of Equality Law, 97 B.U. L. Rev. 1357, 1382 (2017) (discussing scenarios that “elicit confusion about whether they constitute disparate treatment” based on the functional equivalency of a characteristic with a protected class).

[34].     Rice v. Cayetano, 528 U.S. 495 (2000) (holding ancestry is a proxy for race because it captures a pejorative purpose to classify).

[35].     Barocas & Selbst, supra note 15, at 714.

[36].     Equality law treats these variants as analytically distinct, but antidiscrimination scholars have long pointed to their conceptual unity beneath overarching principles or frameworks. See, e.g., Zatz, supra note 33, at 1383 (unifying the variants of equality law under the framework of status causation); see also Cass R. Sunstein, The Anticaste Principle, 92 Mich. L. Rev. 2410 (1994) (uniting the variants of equality law under an anticaste principle); Joseph Fishkin, The Anti-Bottleneck Principle in Employment Discrimination Law,91 Wash. U. L. Rev. 1429 (2014) (uniting the variants of antidiscrimination law under an antibottleneck principle, designed to prevent the emergence and perpetuation of opportunity “bottlenecks”); Deborah Hellman, Measuring Algorithmic Fairness, 106 Va. L. Rev. 811 (2020).

[37].     Williams v. Dart, 967 F.3d 625, 638 (7th Cir. 2020).

[38].     Rice,528 U.S. 495.

[39].     See discussion infra Part II.C.

[40].     Gross, supra note 29, at 16.

[41].     Gillis & Spiess, supra note 21, at 469 (“For example, if an applicant’s neighborhood is highly correlated with an applicant’s race, we may want to restrict the use of one’s neighborhood in pricing a loan. A major challenge of this approach is the required articulation of the conditions under which exclusion of data inputs is necessary.”).

[42].     See Haney López, supra note 6, at 56–77 (discussing the variation of Supreme Court racial reasoning in Ozawa v. United States, 260 U.S. 178 (1922), and United States v.Thind,261 U.S. 204 (1923). Decided mere months apart, the two cases reflect inconsistent ideas about what evidences racial differences: racial science or common knowledge).

[43].     Yang & Dobbie, supra note 11, at 346–48; Devin G. Pope & Justin R. Sydnor, Implementing Anti-Discrimination Policies in Statistical Profiling Models, 3 Am. Econ. J. 206 (2011).

[44].     Gillis, supra note 20, at 1180–81 (“The input fallacy creates an algorithmic myth of colorblindness by fostering the false hope that input exclusion can create non-discriminatory algorithms.”); see also Talia B. Gillis, Orthogonalizing Inputs, in CS&Law ‘24: Proceedings of the Third Symposium on Computer Science and Law 1 (2024) [hereinafter, Gillis, Orthogonalizing Inputs].

[45].     Gillis, supra note 20, at 1186 (“For when it is no longer possible to scrutinize inputs, outcome analysis provides the only way to evaluate whether a pricing method leads to impermissible disparities.”); Gillis, Orthogonalizing Inputs, supra note 44, at 7.

[46].     See Jessica M. Eaglin, On “Color-blind” and the Algorithm, 112 Geo. L.J. 1385, 1389 (2024).

[47].     Gillis, supra note 20, at 1250 (“In a world in which there is no credible way to determine at the outset whether a protected characteristic is being used to price, the closest alternative world be to ask: are the prices different for protected groups, controlling for the legitimate grounds for differentiation? . . . Only the unexplained component of price disparity would then be the basis of discrimination and not the raw disparities alone.”).

[48].     See id. at 1187, 1219.

[49].     Press Release, Off. of Pub. Affs., U.S. Dep’t of Just., Justice Department Secures Groundbreaking Settlement Agreement with Meta Platforms, Formerly Known as Facebook, to Resolve Allegations of Discriminatory Advertising (June 21, 2022), https://www.justice.gov/opa/pr/justice-department-secures-groundbreaking-settlement-agreement-meta-platforms-formerly-known [https://perma.cc/8Z26-NG7H].

[50].     See Sandra Wachter, The Theory of Artificial Immutability: Protecting Algorithmic Groups under Anti-Discrimination Law, 97 Tul. L. Rev. 149, 153–55 (2022) (analyzing AI’s construction of new groupings and the challenges this poses for antidiscrimination law).

[51].     Michael Omi & Howard Winant, Racial Formation in the United States: From the 1960s to the 1990s (2d ed. 1994).

[52].     Ian F. Haney López, The Social Construction of Race: Some Observations on Illusion, Fabrication, and Choice, 29 Harv. C.R.-C.L. L. Rev. 1 (1994).

[53].     Karen E. Fields & Barbara J. Fields,Racecraft: The Soul of Inequality in American Life (2012).

[54].     Stuart Hall, Race the Floating Signifier: What More Is There to Say About “Race”?, in Selected Writings on Race and Difference 359 (Paul Gilroy & Ruth Wilson Gilmore eds., 2021) (“Well, to put it crudely, race is one of those major concepts which organize the great classificatory systems of difference, which operate in human societies.”).

[55].     See, e.g.,Richard T. Ford, Race as Culture? Why Not?, 47 UCLA L. Rev. 1803, 1804 (2000); Mari J. Matsuda, Voices of America: Accent, Antidiscrimination Law, and a Jurisprudence for the Last Reconstruction, 100 Yale L.J. 1329, 1348–57, 1360–67 (1991); Angela Onwuachi-Willig & Mario L. Barnes, By Any Other Name?: On Being “Regarded As” Black, and Why Title VII Should Apply Even if Lakisha and Jamal Are White, 2005 Wis. L. Rev. 1283, 1288–89, 1297–1312 (2005); Lauren Sudeall Lucas, Identity as Proxy, 115 Colum. L. Rev. 1605, 1613–34 (2015).

[56].     See Devon W. Carbado & Mitu Gulati, Working Identity, 85 Corn. L. Rev. 1259, 1298 (2000) (discussing racial discrimination based on racial performance); Mary Anne Case, “The Very Stereotype the Law Condemns”: Constitutional Sex Discrimination Law as a Quest for Perfect Proxies, 85 Corn. L. Rev. 1447, 1449 (1999); Ford, supra note 55; Matsuda, supra note 55; Onwuachi-Willig & Barnes, supra note 55.

[57].     Parents Involved in Cmty. Schs. v. Seattle Sch. Dist., 551 U.S. 701, 720 (2007).

[58].     See Fair Housing Act, 42 U.S.C. §§ 3601–3619; Title VII of the Civil Rights Act of 1964, 42 U.S.C. §§ 2000e–2000e-17; Equal Credit Opportunity Act, 15 U.S.C. § 1691.

[59].     Yang & Dobbie, supra note 11, at 297 (surveying the use of predictive algorithms in the criminal legal system and finding that “all commonly used predictive algorithms exclude race as an input”).

[60].     See, e.g.,Camille Gear Rich, Performing Racial and Ethnic Identity: Discrimination by Proxy and the Future of Title VII,79 N.Y.U. L. Rev. 1134, 1140 (2004); Onwuachi-Willig & Barnes, supra note 55, at 1297–1312; Lu-in Wang, Race as Proxy: Situational Racism and Self-Fulfilling Stereotypes, 53 DePaul L. Rev. 1013, 1015–16 (2004).

[61].     W. Kerrel Murray, Discriminatory Taint, 135 Harv. L. Rev. 1190 (2022).

[62].     Matsuda, supra note 55.

[63].     See Complaint at 24, United States v. Meta Platforms, Inc., No. 22-cv-05187 (S.D.N.Y. June 21, 2022) [hereinafter DOJ Meta Complaint] (indicating that zip code would no longer be considered in connection with Facebook’s Lookalike Audience tool); The Los Angeles County Risk Stratification Pilot: An Overview and One Year Update 73 (Aug. 29, 2022), https://dcfs.lacounty.gov/wp-content/uploads/2022/08/Risk-Stratification-One-Year-Update_8.24.22.pdf [https://perma.cc/K29X-K887] (“To help mitigate bias, several modeling and implementation decisions were made, including . . . (3) excluding zip code and geographic indicators from the model . . . .”); see also Barocas & Selbst, supra note 15, at 722 (discussing the elimination of certain geographic indicators in a workforce optimization algorithm).

[64].     See Emily Putnam-Hornstein, Rhema Vaithianathan, Jacquelyn McCroskey & Daniel Webster, Los Angeles County Risk Stratification: Model Methodology & Implementation Report (Model Version 1.0), Child.’s Data Network (Aug. 2022), https://dcfs.lacounty.gov/wp-content/uploads/2022/08/Risk-Stratification-Methodology-Report_8.29.22.pdf [https://perma.cc/HH6A-RXEF].

[65].     I borrow this term from Julie E. Cohen, Between Truth and Power: The Legal Constructions of Informational Capitalism 25 (2019).

[66].     See generally Rashida Richardson, Racial Segregation and the Data-Driven Society: How our Failure to Reckon with Root Causes Perpetuates Separate and Unequal Realities, 36 Berkeley Tech. L.J. 1051 (2021).

[67].     See Yang & Dobbie, supra note 11, at 298.

[68].    Prince & Schwarcz, supra note 15, at 1276; Williams, Brooks & Shmargad, supra note 17, at 106.

[69]. See, e.g., Huq, Racial Equity, supra note 19, at 1062 (discussing “instruments” that are “generally calibrated using one pool of data and then applied to new data as a means of identifying or predicting crime that was previously unknown and that, typically, has not yet occurred”); Kim, supra note 15, at 1551 (“[A] predictive model is developed by applying mathematical tools to extract patterns from an existing dataset called the training data. Those observed patterns are then used to make inferences about what will happen in future cases.”).

[70].     See Anupam Datta, Matt Frederikson, Gihyuk Ko, Piotr Mardziel & Shayak Sen, Use Privacy in Data-Driven Systems: Theory and Experiments with Machine Learnt Programs, in CCS ‘17: Proceedings of the 2017 ACM SIGSAC Conference On Computer And Communications Security 1193, 1194 (2017); see also Gillis, supra note 20, at 1183 (“[I]nformation about a person’s protected characteristics is embedded in other information about the individual, so that a protected characteristic can be ‘known’ to an algorithm even when it is formally excluded.”).

[71].     Cary Coglianese & Alicia Lai, Algorithm vs. Algorithm, 71 Duke L.J. 1281, 1297–98 (2022).

[72].     See generally Alice Xiang, Reconciling Legal and Technical Approaches to Algorithmic Bias, 88 Tenn. L. Rev. 649 (2021) (explaining the processing capacity of algorithms and mentioning their potential negative effects on equality).

[73].     Barocas & Selbst, supra note 15, at 721 (“[I]f you wanted to remove everything correlated with race, you couldn’t use anything.”) (citing Nadya Labi, Misfortune Teller, Atlantic (Jan./Feb. 2012), https://www.theatlantic.com/magazine/archive/2012/01/misfortune-teller/308846/ [https://perma.cc/74PC-T4WP]); see also Yang & Dobbie, supra note 11, at 291.

[74].     Harcourt, supra note 16, at 240 (identifying prior criminal history as a proxy for race: “Current actuarial instruments vary widely in the number and type of risk factors that they include, but all place heavy weight on criminal history. Unfortunately, reliance on criminal history has proven devastating to African American communities and can only continue to have disproportionate impacts in the future”); Starr, supra note 15, at 838 (arguing that factors like criminal history and geographic indicators raise concerns about the impermissible racial proxy).

[75].     Starr, supra note 15, at 838 (“[Even when some actuarial instruments exclude race], the socioeconomic and family variables that they do include are highly correlated with race, as is criminal history, so they are likely to have a racially disparate impact. Given widespread de facto residential segregation and the concentration of crime in urban neighborhoods of color, the neighborhood crime rate variables found in some instruments are particularly disturbing.” (citations omitted)).

[76].     See, e.g., id.

[77].     Yang and Dobbie call this the “formalistic solution” of excluding inputs to the algorithm. Yang & Dobbie, supra note 11, at 298, 343–46.

        [78].     Gillis, supra note 20, at 1185.

[79].     Id.

[80].     Prince & Schwarcz, supra note 15, at 1304 (“Making matters even worse, the black box nature of AIs and the vastness of big data mean that intuition alone will often be inadequate to identify an AI’s use of a proxy variable, even after the fact.”).

[81].     See id.

[82].     Gillis, supra note 20, at 1184–85; Yang & Dobbie, supra note 11, at 315, 334.

[83].     Robert Brauneis & Ellen P. Goodman, Algorithmic Transparency for the Smart City, 20 Yale J.L. & Tech. 103, 125 (2018) (“However, due to residential segregation, zip codes are often proxies for race. Knowing this, agencies may choose to exclude zip codes as inputs to predictive algorithms even where they improve the algorithm’s predictive power.”).

[84].     See, e.g., Gillis, supra note 20, at 1237 (“[I]nput exclusion comes at the price of prediction accuracy, which may hurt vulnerable populations.”); Yang & Dobbie, supra note 11, at 319 (“[C]hoosing to exclude protected characteristics comes at the cost of predictive accuracy.”); Richard Berk, Hoda Heidari, Shahin Jabbari, Michael Kerns & Aaron Roth, Fairness in Criminal Justice Risk Assessments: The State of the Art, 50 Socio. Methods & Rsch. 3, 3 (2021); Richard Berk, Accuracy and Fairness for Juvenile Justice Risk Assessments, 16 J. Empirical Legal Stud. 175, 175–76 (2019); Tameem Adel, Isabel Valera, Zoubin Ghahramani & Adrian Weller, One-Network Adversarial Fairness, in AAAI-19: Proceedings of the 33rd AAAI Conference on Artificial intelligence 2412, 2412 (2019).

[85].     See, e.g., Gillis, supra note 20, at 1237; Yang & Dobbie, supra note 11, at 319.

[86].     Students for Fair Admissions, Inc. v. President & Fellows of Harvard Coll., 143 S. Ct. 2141, 2185–86 (2023) (Thomas, J., concurring).

[87].     Id.

[88].     Id.; Jack M. Balkin & Reva B. Siegel, The American Civil Rights Tradition: Anticlassification or Antisubordination, 58 U. Mia. L. Rev. 9, 11 (2003).

[89].     Lehr & Ohm, supra note 8, at 682 (explaining that in many of the most important decisions in the development of algorithms, analysts will have to make “judgment calls” and “bring subject matter knowledge to bear”).

[90].     Students for Fair Admissions, 143 S. Ct. at 2165–66.

[91].     Id. (strictly scrutinizing race-based admissions designed to advance racial diversity in higher education).

[92].     See, e.g., Williams v. Dart, 967 F.3d 625, 638 (7th Cir. 2020).

[93].     See Rice v. Cayetano, 528 U.S. 495, 510 (2000).

[94].     Seeid.

[95].     Id. at 499.

[96].     Id.

[97].     Id. at 498–99.

[98].     Id. at 510.

[99].     Id. at 514.

[100].     See Addie C. Rolnick, Indigenous Subjects, 131 Yale L.J. 2652, 2667–68 (2022) [hereinafter Rolnick, Indigenous Subjects]; see also Addie C. Rolnick, The Promise of Mancari: Indian Political Rights as Racial Remedy, 86 N.Y.U. L. Rev. 958, 1008 (2011) [hereinafter Rolnick, The Promise].

[101].     Rolnick, Indigenous Subjects, supra note 100, at 2693.

[102].     Rice, 528 U.S. at 510, 514.

[103].     Id. at 513–14.

[104].     Id. at 514–15.

[105].     Id. at 515.

[106].     Id. at 514.

[107].     Id. at 500–01, 514.

[108].     Id. at 514.

[109].     Id.

[110].     Id. at 515–16.

[111].     See generally id. at 514.

[112].     Id. at 517.

[113].     Id. at 516–17.

[114].     Id. at 517 (explaining that the State’s inquiry into ancestorial lines evinced “the same grave concerns as a classification specifying a particular race by name”).

[115].     Id.

[116].     Id. at 539–40 (Stevens, J., dissenting).

      [117].     238 U.S. 347 (1915).

[118].     Rice, 528 U.S. at 540 (Stevens, J., dissenting).

[119].     Guinn, 238 U.S. at 367–68.

[120].     Rice, 528 U.S.at 540 (Stevens, J., dissenting).

[121].     Id. at 541.

      [122].     Rolnick, Indigenous Subjects, supra note 100, at 2695.

[123].     Id. at 2722.

[124].     Rice, 528 U.S.at 514.

      [125].     J. Kehaulani Kauanui, The Politics of Blood and Sovereignty in Rice v. Cayetano, 25 Pol. & Legal Anthropology Rev. 110 (2002).

      [126].     Williams v. Dart, 967 F.3d 625, 630 (7th Cir. 2020).

[127].     Id.

[128].     Id.

[129].     Id. at 632.

[130].     Id. at 638.

[131].     Id. at 637–38.

[132].     Id. at 638.

[133].     Id.

[134].     Id.

[135].     Id.

[136].     Id.

[137].     Alana Semuels, Chicago’s Awful Divide, Atlantic (Mar. 28, 2018), https://www.theatlantic.com/business/archive/2018/03/chicago-segregation-poverty/556649/ [https://perma.cc/NYU5-2TBE].

[138].     Williams, 967 F.3d at 638.

[139].     Id.

[140].     Id. Although, the court was less sure about the “proxy mechanism” in the case of criminal charges.

[141].     See, e.g.,Cassia Spohn, Race, Crime, and Punishment in the Twentieth and Twenty-First Centuries, 44 Crime & Just. 49, 57 (2015); Yu Du, Racial Bias Still Exists in Criminal Justice System? A Review of Recent Empirical Research, 37 Touro L. Rev. 79, 92–94 (2021); Paul Butler, Race and Adjudication, in 3 Reforming Criminal Justice: Pretrial and Trial Processes 211, 215 (Erik Luna ed., 2017).

[142].     Cal. Penal Code § 741 (requiring the state Department of Justice to develop race-blind charging guidelines).

[143].     See Williams, 967 F.3d at 638 (acknowledging that arrest history and neighborhood are proxies for race without explaining why pending criminal charges may not be).

[144].     See id.; Rice v. Cayetano, 528 U.S. 495, 514–15 (2000).

[145].     852 F.3d 1018, 1020 (11th Cir. 2016).

[146].     Id.

[147].     Id. at 1030.

[148].     Id. at 1024.

[149].     Id. at 1022.

[150].     Id. at 1026.

[151].     Id. at 1027.

[152].     Id.

[153].     Id. at 1030.

[154].     See id. at 1033–35.

[155].     Id. at 1030 (“[D]iscrimination on the basis of black hair texture (an immutable characteristic) is prohibited by Title VII, while adverse action on the basis of black hairstyle (a mutable choice) is not.”).

[156].     Seeid. at 1027.

[157].     “Racial regimes are constructed social systems in which race is proposed as a justification for the relations of power. While necessarily articulated with accruals of power, the covering conceit of a racial regime is a makeshift patchwork masquerading as memory and the immutable.” Cedric J Robinson, Forgeries of Memory and Meaning: Blacks and the Regimes of Race in American Theater and Film Before World War II, xii–xiii (2007).

[158].     852 F.3d at 1033–35.

[159].     See Sandra Wachter, The Theory of Artificial Immutability: Protecting Algorithmic Groups under Anti-Discrimination Law, 97 Tul. L. Rev. 146, 154 (2022) (“AI systems frequently use input data containing features that cannot be meaningfully interpreted by a human observer due to their scale (both small and large), volume, complexity, or source.”).

[160].     See Haney López, supra note 6, at 35–55, app. A at 163 (discussing racial prerequisite to citizenship cases such as Ozawa v. United States, 260 U.S. 178 (1922),and United States v. Thind, 261 U.S. 204 (1923), that move between a reliance on so-called scientific evidence and racial common sense).

[161].     Gross, supra note 29, at 98–99; see Haney López, supra note 6, at 35–55, app. A at 163.

[162].     See Students for Fair Admissions, Inc., v. President & Fellows of Harvard Coll., 143 S. Ct. 2141, 2185 (2023) (Thomas, J., concurring) (“[B]ecause ‘not all blacks in the United States were former slaves,’ ‘freedman’ was a decidedly under-inclusive proxy for race.” (quoting Michael B. Rappaport, Originalism and the Colorblind Institution, 89 Notre Dame L. Rev. 71, 98 (2013))).

[163].     See Yang & Dobbie, supra note 11, at 43 (“[T]hese algorithms take a very haphazard approach to dealing with nonrace correlates and proxy effects, sometimes excluding inputs deemed to be correlated with race out of fairness concerns (even if a loss to accuracy) yet also retaining others that are also likely correlated with race, including in particular current offense and criminal history.”).

[164].     Haney López, supra note 6, at 56–77 (discussing the variation of Supreme Court racial reasoning in Ozawa, 260 U.S. 178,and Thind, 261 U.S. 204, which, although decided mere months apart, reflect inconsistent ideas about what evidences racial differences, racial science, or common knowledge).

[165].     Yang & Dobbie, supra note 11, at 298; Pope & Sydnor, supra note 43, at 207.

[166].     Gillis, supra note 20, at 1186.

[167].     See generally Lily Hu, What is “Race” in Algorithmic Discrimination on the Basis of Race?, 21 J. Moral Phil. 1 (2024); Lily Hu & Issa Kohler-Hausmann, What Is Perceived When Race Is Perceived and Why It Matters for Causal Inference and Discrimination Studies, 59 Law & Soc’y Rev. 239 (2025).

[168].     Pope & Sydnor, supra note 43, at 207; Yang & Dobbie, supra note 11, at 343 (“Our first recommended solution purges all algorithmic inputs of the proxy effects of race in the estimation step of the predictive algorithm, and then uses these ‘colorblind’ inputs to predict outcomes in the prediction step.”).

[169].     Yang & Dobbie, supra note 11, at 343.

[170].     Id. at 346.

[171].     Id. at 346–47.

[172].     See id. at 346.

[173].     See id. (“[T]he colorblinding-inputs algorithm does not exclude race and race correlates in the estimation step. In fact, it uses all inputs to estimate predictive relationships . . . .” (alteration in original)).

[174].     See id. at 346–48.

[175].     Id. at 346.

[176].     Sebastian Benthall & Bruce D. Haynes, Racial Categories in Machine Learning, in FAT ’19: Proceedings of the 2019 Conference on Fairness, Accountability and Transparency 289, 295 (2019).

[177].     For work that problematizes approaches to securing the race variable in algorithms, seeJessica M. Eaglin, Racializing Algorithms, 111 Calif. L. Rev. 753, 787–88 (2023); Abdu, Pasquetto & Jacobs, supra note 13, at 1327–33.

[178].     See, e.g.,White Logic, White Methods: Racism and Methodology (Tukufu Zuberi & Eduardo Bonilla-Silva eds., 2008); Laura E. Gómez, Looking for Race in All the Wrong Places, 46 Law & Soc’y Rev. 221, 229–34 (2012).

[179].     See United States v. Johnson, 122 F. Supp. 3d 272, 331 (M.D.N.C. 2015) (suggesting that a statistical study was unreliable because its author identified people as “Hispanic” based on self-reports and observations of who “appeared to be” Hispanic).

[180].     Cheryl I. Harris, Whiteness as Property, 106 Harv. L. Rev. 1707, 1710–11 (1993).

[181].     See Khiara Bridges, Race in the Machine: Racial Disparities in Health and Medical AI, 110 Va. L. Rev. 243, 286 (2024) (discussing how the variable of “black race” in machine learning algorithms can conceal the “persistence of the myth of biological race”).

[182].     Devon W. Carbado & Cheryl I. Harris, Intersectionality at 30: Mapping the Margins of Anti-Essentialism, Intersectionality, and Dominance Theory, 132 Harv. L. Rev. 2193, 2193 (2019) (“[C]ontest[ing] the view that feminism and critical theory must always avoid essentialism to achieve normative commitments to social transformation.”).

[183].     Hall, supra note 54, at 359–73 (showing race as signs, symbols, and language).

[184].     DOJ Meta Complaint, supra note 63 (referencing the usage of zip code in Facebook’s Lookalike Audience tool).

[185].     EEOC v. Catastrophe Mgmt. Sols., 852 F.3d 1018, 1030 (11th Cir. 2016) (explicitly suggesting that hair texture, as opposed to hair style, can be a proxy for race).

[186].     Rice v. Cayetano,528 U.S. 495, 496 (2000) (using Hawaiian ancestry as a proxy for race).

[187].     See, e.g., Hudgins v. Wrights, 11 Va. (1 Hen. & M.) 134 (1806) (assigning race by reference to morphological features other than complexion, including hair and nose shape); Abdullahi v. Prada USA Corp., 520 F.3d 710, 712 (2008) (racially classifying Iranians by reference to certain proxy variables including national origin, accent, and experiences of discrimination); Hernandez v. Texas, 347 U.S. 475, 479–81 (1954) (expanding equal protection rights to Mexican Americans as a racialized class by reference to racial proxies such as surname, community attitudes, and experiences of segregation); Saint Francis Coll. v. Al-Khazraji, 481 U.S. 604, 613 (1987) (racially classifying Iraqi Americans based on racial proxies such as ancestry or ethnic characteristics).

[188].     See Ariela J. Gross, Litigating Whiteness: Trials of Racial Determination in the Nineteenth-Century South, 108 Yale L.J. 109, 129–30 (1998).

[189].     Angela James, Making Sense of Race and Racial Classification, in White Logic, White Methods: Racism and Methodology 31, 36–38 (Tukufu Zuberi & Eduardo Bonilla-Silva eds., 2008).