Professors Ashley Deeks and Kristen Eichensehr have written a fine article that significantly advances discussion of national security federalism.1 The increase in state and local involvement in national security deserves renewed attention,2 and the article offers several insightful thoughts on this area of law and policy. Much of the article is persuasive, especially when it analyzes prevailing approaches to federalism and when it describes what courts have done recently when facing national security federalism issues. That said, the article’s normative suggestions prompt me to offer three thoughts. I write from my vantage point not as a scholar of national security or federalism in general, but rather as someone who studies immigration and citizenship. In that field, preemption issues can arise in ways that shed light on national security federalism generally.
My first point starts with the article’s account of “entrepreneurial federalism” as a way to understand how courts are resolving preemption in cases that involve national security.3 This descriptive account is a persuasive and significant contribution. The authors then go further to suggest that preemption should turn in large part on whether state law, in relation to federal law, provides “useful supplementation” or leads to “productive friction” with federal law.4 I question the utility of this approach, and so the article leaves me still searching for a coherent doctrinal framework for national security federalism.
My second point starts with praise for the article’s suggestion that preemption analysis in national security cases should not differ fundamentally from preemption analysis generally.5 I agree that any approach that leads more aggressively in national security cases to preemption brings “underappreciated perils.”6 In my view, judges will likely be inclined to find preemption in “national security” cases, so the label’s malleability will undermine the coherence of national security federalism. It’s troubling that “national security” may be used to justify outcomes that courts prefer for other reasons.
My third point builds on the first two. I suggest a way to address part of the problem generated by the indeterminacy of preemption doctrine and the malleability of the “national security” label. Here I respond to the article’s premise that judges face a binary choice between either preemption or rights-based decisionmaking.7 I believe, however, that these approaches are related and should be intertwined in doctrine. Because existing approaches to preemption are indeterminate, and because the “national security” label is so malleable, preemption analysis would be more sure-footed if it considered the risk of rights violations if state laws are allowed to stand. This approach can assist preemption analysis not only in “national security” cases, but in all cases where preemption might apply.
I. Addressing Preemption
This Part of my Response addresses the absence of a coherent approach to preemption in cases that implicate national security. Deeks and Eichensehr’s article starts by discussing how several other approaches to federalism fail to explain states’ recent approaches to national security federalism.8 Instead, the authors characterize recent state actions in the national security space as reflecting “entrepreneurial federalism.”9
Against this backdrop, the article identifies the key challenge as establishing a “reasonable middle ground” on preemption.10 The authors argue that “there are underappreciated perils in broad preemption.”11 They explain:
[B]road preemption . . . may inadvertently foreclose two constructive phenomena. The first is states’ useful supplementation of federal efforts to address national security threats, including through gap-filling statutes or enhanced enforcement of the federal scheme. The second is the productive friction that states can provide when they operate in this space.12
Accordingly, the article asks if a state law’s relationship with federal law is “helpful or, conversely, unhelpful and in need of preemption.”13 Relatedly, the authors suggest that a state law that leads to “productive friction” should withstand a preemption challenge,14 in contrast to the preemption that they suggest for state laws that lead to “unproductive friction.”15
The article also explains that courts facing preemption issues should ask first if “the state assert[s] that its actions fall within a traditional area of state responsibility.”16 If a state cites its “traditional police powers”17 or “traditional areas of responsibility,”18 courts should ask if invocations of state responsibility are “a pretext for what is actually a state effort to critique a foreign power or otherwise interfere with federal foreign policy.”19
This explanation prompts me to ask more fundamentally: How do we know if a state law supplements federal law in a “useful” way? State lawmakers may try to “help” the federal government, but the federal government may view state involvement as undermining the compromises that shape the federal approach. Related concerns arise with asking if a state law creates “productive friction” because it “stimulates federal enforcement.”20 The article implies that this effect suggests an absence of conflict between state and federal law, and thus no preemption.21
But all laws reflect compromises and tradeoffs. Federal laws may or may not be meant to be enforced fully as written.22 Even if states “stimulate,” preemption may be appropriate. To be sure, tempered enforcement might reflect resource constraints,23 so a state law that enhances enforcement of a federal law may be supplementing “usefully.”24 If so, preemption may be less appropriate. But in other settings, incomplete enforcement may have little to do with resource constraints and instead reflect deliberate policy choices.
This happens, for example, when complete enforcement requires intrusions on privacy that lawmakers view as impermissible. Imagine that a state wants to “help” enforce a federal law against illicit drug sales by requiring surveillance cameras on every street corner. This state measure might be “useful” to federal enforcement of federal drug laws,25 and the difference between state and federal enforcement may lead to “productive friction.”26 But perhaps the federal law reflects a compromise to limit surveillance cameras. Because intrusions on privacy are inherent in the aggressive surveillance required for zero-tolerance enforcement, federal law may reflect a compromise to define and deter violations without constant surveillance of all public activity.
Similarly, state laws that fill gaps in federal law may or may not be “useful” from a federal perspective.27 By design or practice, such gaps may give federal officers a zone of discretion that would be limited if states were to fill those gaps. A related problem arises with state laws that bring state activity “into line with international law in ways that can reduce friction” with foreign partners.28 From the federal perspective, this harmony may not be useful — and it may be unproductive. A related issue arises with the article’s disinclination to preempt state laws that “motivat[e]” federal action.29 “Motivation,” like state-stimulated enforcement, may conflict with federal law.
Sensitivity to these uncertainties explains the U.S. Supreme Court’s 2012 decision in Arizona v. United States.30 The case involved Arizona Senate Bill 1070,31 which became state law in 2010 and imposed state criminal penalties on noncitizens who violated federal immigration law.32 The law also required state officers to inquire into the immigration status of noncitizens whom the officers encountered and to detain those suspected of being present in the country unlawfully.33 Several lawsuits challenged the legislation by arguing that it violated equal protection or that federal law preempted it.34 The Court considered only preemption challenges to four core provisions of S.B. 1070 and struck down three of them.35 Its analysis emphasized two aspects of federal immigration enforcement.
First, even though federal law makes it unlawful to work without employment authorization, Congress chose to impose criminal penalties on employers and not on employees.36 A state law imposing a criminal penalty on employees would upset this federal legislative compromise, even if federal law made it generally unlawful for the undocumented person to work.37 Second, the federal government may choose to exercise discretion and moderate its enforcement of immigration law against noncitizens whom it could, in theory, lawfully remove from the United States.38 In this way, the Court declined to exempt state laws that “mirror” federal law from preemption challenges. Federal law isn’t just law on the books; it’s also the results that enforcement discretion shapes in practice.39 The law in action may reflect tradeoffs that stop short of zero tolerance.
This view of what “law” is has roots in previous U.S. Supreme Court decisions in immigration cases. In Plyler v. Doe,40 the Court held that the state of Texas violated equal protection by impeding or forbidding undocumented children from attending K–12 public schools in the state.41 The Court relied in part on the likelihood that undocumented children, even without lawful status, would remain in the United States.42 “In light of the discretionary federal power to grant relief from deportation, a State cannot realistically determine that any particular undocumented child will in fact be deported until after deportation proceedings have been completed.”43 This reasoning also appears in immigration cases from the lower courts.44
To be sure, I rely on U.S. Supreme Court decisions that reflected a 5–4 vote, and these precedents may be fragile on these specific points or more generally. But in assessing approaches to national security federalism, it’s appropriate to reflect on whether the analysis is sound. I cite some court decisions less as precedent — this Response is not a brief — but rather as examples of persuasive reasoning.
Relatedly, I have concerns about the U.S. Supreme Court’s 2020 decision in Kansas v. Garcia.45 The case arose from Kansas’s criminal prosecution of three noncitizens who had no federal permission to work in the United States.46 Around the time that they were hired, each used another person’s Social Security number in filling out state and federal tax withholding forms.47 The three were convicted under Kansas law but then appealed, arguing that federal law on the employment of unauthorized workers preempted the Kansas law.48 The Court found no preemption, reasoning that the Kansas law governed taxation, not immigration, and so federal law didn’t preempt it.49
The Court reached this result by focusing on the formal compatibility of state and federal law but disregarding the practical inconsistency between them. According to the Court, it didn’t matter that Kansas wanted to enforce its state tax law with greater enthusiasm than the federal government was enforcing immigration law in a hiring context that required compliance with both federal immigration law and state tax law.50 The result, troubling in my view, was to allow Kansas, through enforcement of a state tax statute, to intensify federal immigration enforcement beyond what the federal government was choosing to do on its own.51
Some have objected that only federal law that is adopted with a certain degree of formality counts as “federal law” for preemption purposes.52 But it’s unclear why the pattern of results created by the discretionary application of law to individual situations isn’t “law” for preemption (or other) purposes. Whether government action counts as “law” doesn’t depend on how it is adopted or the form that it takes. This is one lesson from Erie Railroad Co. v. Tompkins,53 the landmark 1938 U.S. Supreme Court decision that rejected, on constitutional grounds, the prior approach to choice of law in federal courts hearing cases based on diversity jurisdiction.54 The defect in prior doctrine was that it recognized as state “law” state statutes, but not state court decisions, thus relying impermissibly on the form that “law” took.55
I recognize that law that reflects enforcement discretion and doesn’t take written, published form can raise concerns. One is vagueness, which impedes notice about what activities are allowed or forbidden. Another is the absence of transparency, which is essential to ensure respect for the rule of law. These risks are real, and they call for rethinking preemption in ways that I’ll explain later in this Response. And they complicate the determination of what federal law in practice is. But these concerns shouldn’t keep federal law in action from having preemptive effect.
Let me offer one final observation about the article’s analytical approach to national security federalism. As I’ve noted, the authors suggest that preemption is less appropriate if a state law will lead to useful supplementation or productive friction.56 But a state law can influence federal authorities by motivating or helping them, or by creating productive friction — even if the state law is preempted.
Immigration law offers historical examples. Though they may recur only infrequently in such clear form, they show how states, even without formal lawmaking, can influence federal decisions for better or worse. In 1876, the U.S. Supreme Court struck down a California anti-Chinese measure on preemption grounds,57 but California’s congressional delegation responded with successful efforts in Washington, D.C., to enact federal Chinese exclusion laws.58
A much more recent example is Proposition 187, a ballot initiative approved by California voters in 1994.59 It tried to limit undocumented residents’ access to public benefits,60 but a federal district court found it preempted by federal law.61 Political hydraulics redirected much of the support for Proposition 187 to Congress, and new federal legislation in 1996 implemented the same pro-enforcement sentiment that had driven California politics just two years before.62 In short, even if these state laws offered useful supplementation or productive friction, they might still be good candidates for preemption.
II. The “National Security” Label
My second point is that the “national security” label is so malleable that it creates troubling risks. I start by agreeing with one of the article’s basic lessons: that courts deciding preemption issues in national security cases should not reflexively find preemption based on the label.63 The article’s critique of national security exceptionalism is based in part on the great expansion of “national security” to include economic security and more.64 The authors also observe that national security now includes “foreign relations” — once seen often as a matter of international trade or human rights, not national security.65
This expansion of “national security” suggests a cogent critique of national security exceptionalism in preemption analysis. As the article notes, judges “generally grant substantial deference to the executive branch’s claims about and management of” foreign relations and national security.66 As a result, approaches that tilt toward preemption in national security cases run the risk of unbounded application — due to the malleability of the “national security” label.
Examples are abundant in immigration-related settings. The long history of viewing immigration law as implicating national security goes back to the Alien Enemies Act of 1798.67 National security and invasion have been themes in a long line of U.S. Supreme Court decisions since the 1880s that have insulated the federal government’s immigration decisions from full judicial scrutiny by applying the plenary power doctrine.68 National security concerns drove decisions that rejected constitutional challenges to anti-Chinese immigration statutes in the late 1800s,69 to truncated due process during the McCarthy era in the 1950s,70 and to draconian immigration law enforcement after the 9/11 terrorist attacks in 2001.71 The Court’s reasoning in these decisions tied immigration firmly to national security, even as the targets shifted from Asian immigrants to Communists and then to Muslims.72
As another example, the second Trump Administration tags as “invaders” not just new arrivals on the southern border, but also noncitizens, including long-time residents, far inside the country.73 This ploy explains why “invasion” and “inundation” have appeared prominently in many of the Administration’s first immigration-related executive orders and proclamations.74 To normalize this invasion narrative, the second Trump Administration has been determined to view immigration as a matter of national security above all. It has tried to deport noncitizens under the Alien Enemies Act,75 induce noncitizens to self-deport,76 and deploy the National Guard for immigration enforcement inside the United States.77
The Administration has repurposed “national security” justifications for immigration-related initiatives to enhance executive power in other settings. It sent the National Guard first to Los Angeles for immigration enforcement,78 then to “[a]ddress the [c]rime [e]mergency” in Washington, D.C.79 These examples show how the malleability of “national security” should give serious pause if the concept serves as an analytical tool to broaden preemption in general.
The malleability of “national security” is closely related to my critique earlier in this Response of an approach to preemption that relies on useful supplementation or productive friction. As I’ve noted, the authors recognize that judges generally defer to the executive branch when it claims that it is managing foreign relations and national security.80 It’s a troubling mix to combine the malleability of the “national security” label with a preemption analysis that evaluates preemption of state laws by asking if they lead to “useful supplementation” or “productive friction.” The temptation is strong for the federal government to invoke “national security” whenever it wants to preempt a state law. Courts may then dismiss too easily a state’s argument that its law doesn’t implicate national security and therefore should not be preempted.
III. Preemption and Rights Claims
My third point suggests a different way to add content to preemption analysis — by assessing the risk of rights violations. I don’t pretend to suggest a comprehensive approach to preemption, but I do suggest a way to move in some cases beyond the indeterminacy of the approach to preemption that Deeks and Eichensehr’s article analyzes.
I start with its discussion of “[t]he [c]hoice [b]etween [c]ompeting [p]reemption and [r]ights [c]laims.”81 The authors imagine a binary choice between deciding a case on preemption grounds and deciding it on rights grounds if the case raises both.82 The authors urge courts in such cases “to rule on alternative, rights-based constitutional grounds.”83 They urge this especially when a rights-based holding is “comparatively easy to reach” as compared to a “complicated” preemption analysis.84 The article also notes that “there may be symbolic value to vindicating a rights-based injury with a rights-based ruling, rather than a ruling based on constitutional structure and principles of federalism.”85
But how often are holdings regarding the First Amendment, due process, or equal protection “comparatively easy”? “Easy” cases with rights claims are especially hard to find in settings related to immigration and immigrants.86 Immigration status inherently assumes the unequal treatment of noncitizens. And the plenary power doctrine curtails judicial scrutiny of constitutional claims in immigration cases, impeding full appraisal of such arguments.87 This raises a tough question: When does this discrimination reflect impermissible considerations that might include racism, religious bias, gender discrimination, or due process violations? More generally, hard cases involving rights lurk everywhere.
Courts have more than a simple binary choice between deciding on the basis of preemption or rights claims. Rather, in many cases courts should consider rights claims in deciding preemption. When challengers raise both types of arguments, this intermediate option is an analytically coherent approach that infuses preemption analysis with awareness of rights claims.
In cases involving state laws that address immigration or immigrants, challengers often invoke both equal protection and preemption. In deciding these challenges, a court may find the evidence insufficient to show the discriminatory intent that is required for an equal protection violation, but the same court might also find that the risk of unlawful discrimination is sufficient to deny state authority by finding federal preemption. This can be a coherent way to address both equal protection and preemption issues.88
A prominent example is the U.S. Supreme Court’s 2012 decision in Arizona v. United States. Recall that this case was a challenge to Arizona’s Support Our Law Enforcement and Safe Neighborhoods Act,89 commonly referred to as S.B. 1070, which required state and local efforts to enhance federal immigration enforcement through direct means that included arrest, detention, and state and local police demands for the identification papers of some individuals who come into contact with law enforcement.90
Many challengers, including individuals, churches, labor unions, and community organizations, sued to block parts of the law.91 Though most of these suits argued both preemption and equal protection, the only challenge that reached the Supreme Court was a challenge by the federal government that solely raised preemption.92 But it became clear, especially during oral argument, that the arguments for preemption relied heavily on possible equal protection violations in the form of discriminatory policing against Latinx communities.93 In cases of this type, it’s reasonable to be concerned about state laws that enable subfederal actors to make discretionary decisions, if those decisions may reflect racial or ethnic discrimination that will escape detection and remedy. Limiting state laws can prevent discrimination at state hands, as well as First Amendment violations and other impermissible outcomes from government decisions.
A recent U.S. Supreme Court decision, Williams v. Reed,94 seemed to view preemption and a rights claim in a way that suggests that they can and should influence each other.95 The case concerned an Alabama state law that required exhaustion of administrative remedies before suing in court for a denial of unemployment benefits.96 The Court found federal preemption on the ground that applying the state law effectively immunized state officials from liability for due process claims brought under a federal civil rights statute, 42 U.S.C. § 1983.97
This combination of preemption and rights claims resembles other legal rules that are prophylactic by reducing the number or magnitude of undetected or unremedied constitutional violations.98 Examples include Miranda warnings, which protect suspects by inhibiting self-incrimination and involuntary confessions.99 Constitutional doctrine invalidating vague criminal statutes is similar. Such statutes are troubling because they discourage constitutionally permitted behavior,100 but the extent of this harm is unknown because we don’t know what someone would have done absent the vague laws.
This reasoning also explains the success of a challenge to two local ordinances in Hazleton, Pennsylvania.101 The city required all individuals to prove U.S. citizenship or lawful immigration status if they were living in rental housing or working in the city.102 The challengers alleged both federal preemption and unlawful discrimination in violation of the Equal Protection Clause.103 To support their equal protection claim against one of the ordinances, the challengers introduced evidence that the ordinances would disproportionately affect Latinx residents, not just undocumented residents but also U.S. citizens and lawfully present noncitizens.104 But the validity of such claims is hard to establish. Challengers must show more than disparate effects; they must prove discriminatory intent.105 The court struck down the ordinances on preemption grounds instead.106
Fundamentally, discrimination law requires courts to identify actors and practices that discriminate. These cases involve high stakes. Judges may hesitate to blame culprits, perhaps as racists, especially given the difficulties of ascertaining what state or local officials were thinking. But this outcome is deeply troubling if the risk of undetected and unremedied constitutional violations is real. In such settings, preemption can inhibit unlawful discrimination, not by directly monitoring and penalizing decisions, but instead by limiting who may decide.
The article sometimes seems open to blending preemption with rights. The authors observe that “because some of the recent state actions seem to have been motivated by racism or xenophobia, courts may view those actions with added skepticism and find preemption an especially attractive tool,”107 or that “some of the current state actions merit preemption, including because of their discriminatory underpinnings.”108 The article later notes that preemption may be more justified when state decisions are “impermissibly motivated by discrimination”109 or are backed by “xenophobic[] rhetoric.”110 Yet these observations are in tension with the article’s apparent view that preemption and rights are distinct.
My suggestion to blend preemption with rights has its critics. One concern is that challengers — or possibly judges — will use a weak rights claim to press a weak preemption claim when neither is persuasive. Professor David Rubenstein has articulated this view cogently. Under this view, the blend stretches both equal protection and preemption beyond their doctrinal bounds.111 One aspect of this critique contends that I argue for de facto expansion of equal protection — with no basis in current doctrine — by allowing equal protection considerations to influence decisions on preemption.112 This means that challenges may succeed if they rely on a combination of preemption and rights claims, even if neither claim would be doctrinally well-founded on its own.
I appreciate why this outcome can prompt doubts. In Rubenstein’s view, I treat inadequate direct evidence of an equal protection violation as relevant to finding preemption.113 To be sure, my approach calls for the law to mesh preemption and rights claims in ways that seem to go beyond received understandings of doctrine. And yet, Arizona, Lozano v. City of Hazleton,114 and other cases show how this already happens in preemption cases with sensible results.
When would a preemption analysis that considers rights claims not lead to preemption? It would not in any situation that raised preemption issues but not rights-based concerns that would be heightened and escape remedy if state laws went into effect. A useful touchstone is whether a preemption claim would have a parallel rights claim by a person with standing to sue under Article III.
For example, state laws that authorize discretionary enforcement against the undocumented run the risk of racial, ethnic, or language/accent profiling that strongly suggests preemption. Other state laws address the situation of undocumented persons in the state but are unlikely to have parallel rights claims. Those who object to these laws wouldn’t face harm that is sufficient for standing to sue in federal court. State laws that authorize non-discretionary enforcement should survive preemption challenges. So should state laws that don’t involve enforcement at all.
Lessons from history inform my approach. Ever since the Civil War, an essential constitutional role for the federal government has been to ensure that regions, states, and localities do not rely improperly on race when they decide who belongs and who is respected under the law.115 On equal protection, my approach makes it more likely that state discretionary enforcement would be invalidated based on preemption, even if the evidence is insufficient to prove the discriminatory intent required for a violation. Similarly, my reasoning could make evidence of other constitutional rights violations relevant to preemption. I agree with the authors that “there may be symbolic value to vindicating a rights-based injury with a rights-based ruling.”116 But there is also symbolic value in preventing a rights-based injury through preemption.
IV. Combining Three Points
My three points are closely related to each other. First, the article’s proposed approach relies too much on whether state law leads to what is useful or productive. Second, the problem of indeterminacy in national security preemption cases is exacerbated by the malleability of the “national security” label and the inclination to preempt that follows. Third, a binary choice between preemption and rights grounds too readily dismisses a blended approach: treating rights claims as bearing on preemption. The most compelling reason for this approach is the risk that discriminatory intent against certain nationalities or ethnic groups is behind the law’s enactment, but that this intent will escape detection and remedy if evaluated only as a matter of equal protection.
Letting rights claims influence preemption can go a long way to addressing the problems that my first and second points highlight. Preemption would turn less on the method chosen to distinguish federal from state turf and also less on inquiries into what is useful or productive. Instead, preemption would depend on whether state activity runs the risk of rights violations that might go undetected and unremedied when committed by states.
This middle ground doesn’t require changing or conflicting with equal protection doctrine or any other rights-focused doctrines themselves. Any shift affects preemption doctrine by allowing it to consider the risk that hidden but real injustices come from delegating authority to actors whose decisions will escape scrutiny. This approach would evaluate the preemption of state authority with this risk in mind. The effect is a coherent middle ground — in cases of state laws that raise preemption issues — between the requirement to show discriminatory intent and the risk that discriminatory effects can suggest but not prove that intent.
Crucially, this middle ground protects someone potentially injured by a rights violation without requiring judges or juries to identify and blame specific perpetrators. This compromise is especially appropriate when a focus on identifying perpetrators obscures systemic origins and makes remedies depend on willingness to name the especially blameworthy.117 The U.S. Supreme Court in Arizona gives this middle ground precedential foundation — at least in cases involving immigration or immigrants,118 or perhaps more generally, as decisions like Williams may suggest.119 Of course, preemption doesn’t apply to federal decisionmaking, but that limit on the reach of this approach isn’t an analytical flaw.
This Response doesn’t allow full discussion of this approach to preemption, but I can flag items that should be on an agenda for serious thought. First is discussion of whether this approach would be appropriate for rights claims other than equal protection. Second is evaluation of ways to decide if implementation of state law depends on discretionary decisions — by government employees or by private actors — that lack the transparency needed to detect and remedy constitutional violations. Third is the question whether cases related to immigration and immigrants are exceptional, or if gaps between law on the books and law in action in other settings make discretionary decisions crucial and seriously impede scrutiny for rights violations. Fourth is whether decisions that seem ministerial will still evade the scrutiny that supports confidence in the lawfulness of those decisions.
I am grateful to the authors for their many insights into national security federalism and for prompting this discussion of ways forward and an agenda for future inquiry.