Article

Chevron’s Liberty Exception

I.     Introduction

When the Senate considered Neil Gorsuch’s nomination to the Supreme Court in 2017, much of the opposition to him focused on his criticism of Chevron deference.1 Since the late 1980s, Chevron has been the central doctrine in administrative law. It calls on courts to defer to executive branch agencies on the interpretation of ambiguous statutes. In the process, it gives agencies more space in which to craft public policy, and seems to minimize the role of the judiciary in saying what the law is.

While the general public probably does not know that “Chevron” is anything but a petroleum company, Judge Gorsuch’s criticism of a “titanic administrative state”2 seemed to fit a broader political narrative. Around the time when President Trump nominated Judge Gorsuch, presidential advisor Steve Bannon declared that the Trump Administration would be working toward the “deconstruction of the administrative state.”3 The Gorsuch nomination seemed to be a step in this direction. This narrative has been explicitly embraced and applauded in Breitbart News, which has become a prominent institution in right-wing media.4 Reflecting an analogous, if more critical, understanding of Gorsuch’s views, Sen. Maria Cantwell, a Democrat, protested that Gorsuch’s desire to overturn Chevron “could make it easier for courts to overturn important agency decisions protecting public health and the environment.”5

For those who are familiar with Chevron as a legal doctrine, there are obvious ironies in this political narrative. For one thing, the original Chevron decision upheld a Reagan Administration policy that environmentalists opposed. But perhaps even more interesting given the politics of the Trump Administration, then-Judge Gorsuch issued his broadside against the Chevron doctrine in an immigration case, to defend the interests of a Mexican citizen who was trying to adjust his status to become a legal resident.6 This is just one recent indication that there is potential for alliance between immigrant rights advocates and conservative critics of Chevron.7

It is difficult to find a more striking example of a largely unchecked administrative state imposing itself against the liberty of individuals than immigration enforcement. In a deportation case, the Department of Homeland Security operates as police, jailer, prosecutor, and deporter, while the Department of Justice plays the role of judge through its Immigration Courts. Both departments answer to the same Chief Executive, and can easily work together in pursuit of a more aggressive enforcement policy.8 Recently, the Attorney General­­—who supervises the Immigration Courts—has been one of the loudest voices in favor of stricter enforcement of immigration laws.9 A person detained and subject to deportation through the immigration system only reaches the judicial branch of government late in the adjudication process, as a last resort when all administrative appeals have been exhausted.10 The legal issues that can be reviewed in federal court are strictly limited by statute, but Congress explicitly preserved judicial review on “constitutional claims or questions of law.”11 It is here that Chevron deference comes in. If Chevron deference applies, it means that the federal courts will defer back to the Attorney General on difficult questions of law, further minimizing the degree to which meaningful judicial review is available, and strengthening to a corresponding degree the power of the executive branch over people.

At a surface level, the Supreme Court has sent mixed messages about whether deference would be appropriate in these cases. On the one hand, the Court has repeatedly said, clearly and strongly, that courts should defer to the Attorney General’s interpretation of immigration laws, as the Chevron doctrine prescribes.12

On the other hand, the Court appears to honor this prescription in the breach. This kind of inconsistency with the application of Chevron deference is typical for the Supreme Court, and is not unique to immigration, but it has particular consequences for lower courts, which tend to make more consistent efforts to follow deference doctrines in administrative law cases.13 In the specific context of deportation cases, one can certainly see the Court’s failure to adhere to a central doctrine of administrative law as one of many examples of immigration law’s tendency toward exceptionalism.14 It has led to the descriptive observation that “deportation is different” for the Supreme Court, although the Court itself has never quite said so explicitly.15 Moreover, the description raises a deeper question: Why is deportation different? 

This Article intervenes in this confusing situation with two main points.

First, the Supreme Court’s practice with regard to Chevron in immigration cases follows a discernable pattern. The Court does not meaningfully apply Chevron in cases concerning deportation, and also seems reluctant to do so in cases concerning immigration detention.16 But the Court does apply Chevron deference in other kinds of immigration cases. This pattern suggests important differences exist between certain immigration cases which influence the extent to which Chevron deference applies, if at all. Any statement of a rule that claims that Chevron does (or does not) apply in immigration cases is likely to be overbroad. It is important to differentiate the specific issues raised in different types of immigration cases. This Article identifies several possible distinctions between different kinds of immigration issues on which deference may be more or less justified, including grounds for removal, relief from removal, detention, and eligibility for visas to enter the country.

Second, this pattern of avoiding deference in deportation and detention cases makes normative sense. I will argue that Chevron deference is inappropriate when courts review the legality of a government intrusion on physical liberty. This is not an example of immigration exceptionalism, but rather an application of deeply rooted constitutional principles. Immigration is the rare field of administrative law in which an executive branch agency can arrest, detain and force an individual to go someplace against their will. But there may be other arenas where courts consider government requests for deference on legal interpretations that lead directly to incarceration, and there are indications that courts have similarly avoided deferring without setting out an explanation for why.17 Courts should be willing to state clearly that deference on questions of law is inappropriate in the immigration context for the same reasons why it is inappropriate in questions of criminal law.18 Ambiguity from the Supreme Court on this question is a serious problem, given that most immigration cases will be decided by lower courts. Rather than state, falsely, that Chevron deference applies in all immigration cases, the Supreme Court should articulate a rule explaining its actual practice: There is a physical liberty exception to Chevron.

There are three main parts to this Article. Part II describes the current unsteady status of the Chevron doctrine and argues that there is considerable value in identifying patterns from the Supreme Court’s seemingly inconsistent rulings. Part III examines the role of Chevron in the Court’s immigration decisions and demonstrates the existence of a pattern and practice that is best explained by a reluctance to defer to the executive branch when physical liberty is at stake. Part IV highlights some of the implications of this normative theory, including its application beyond deportation cases, the question of whether Chevron should apply in relief from removal cases like asylum eligibility, and the claim that the rule of lenity should govern interpretation of certain immigration statutes.

II.     Chevron’s Unsteady Status

A.     Inconsistent, Criticized, and Canonical

In a broad sense, my purpose in this Article is to make sense of an area of law in which the Supreme Court has not practiced what it has preached, but has nevertheless followed consistent doctrinal patterns. To understand why the Court has behaved this way, we need to begin with an overview of how the Chevron doctrine developed up to this moment. Chevron has become indispensible to administrative law. And yet, neither its doctrinal justifications nor its practical importance have ever been completely secure. As Kent Barnett and Christopher Walker wrote recently, Chevron deference is “both [an] untouchable [doctrine] and yet always under attack.”19 In one of his last opinions, Justice Kennedy wrote: “[I]t seems necessary and appropriate to reconsider . . . the premises that underlie Chevron and how courts have implemented that decision.”20

The basic rule of Chevron requires a two-step analysis when an administrative agency interprets and applies a statute.21 The first is whether the intent of Congress is clear.22 Second, “if the statute is silent or ambiguous with respect to the specific issue, the question for the court is whether the agency’s answer is based on a permissible construction of the statute.”23

To make this determination, “a court may not substitute its own construction of a statutory provision for a reasonable interpretation made by the administrator of an agency.”24

Although the essential rule is spelled out fairly clearly in the original Supreme Court decision, this doctrine has a somewhat peculiar pedigree. Chevron is a paradigmatic canonical case—it has taken on a meaning quite different, or at least quite a bit more important, than someone who had only read the decision itself might understand.25 The Chevron doctrine as we know it emerged through subsequent interpretation by lower courts, helped along by energetic promotion by Justice Antonin Scalia when he joined the Court two years after the initial Chevron decision.26 Interestingly, there is a plausible case that we should really be talking about the General Motors Doctrine, in honor of the 1984 D.C. Circuit en banc decision that seems have been the first to cite and explain Chevron as a major change in administrative law.27 That lower court decision is important because when the Supreme Court announced Chevron, it was understood as a more narrow decision about environmental law. Gary Lawson and Stephen Kam, the authors of an authoritative history which explains how the Chevron case developed into the Chevron Doctrine, explain that “the process by which Chevron became law—a series of lower court decisions and then default acceptance in the Supreme Court—prevented . . . ambiguities from being vented and resolved in an authoritative forum; instead, they remain to this day largely submerged and unaddressed.”28 In other words, the Supreme Court did not necessarily intend to launch a new doctrine with its initial ruling.

Despite its odd birth, for roughly the first two decades of its existence, the Chevron doctrine seemed to have unanimous support on the Supreme Court. This is somewhat remarkable, if one considers the political implications of courts deferring to administrative agencies.29 If we assume that the political left favors more robust government programs and the political right favors smaller government, then the idea that courts should defer to (rather than be a check against) government agencies seems to favor the left. Despite this, Justice Antonin Scalia became (at least until the end of his career) the foremost promoter and defender of Chevron. For him, Chevron was more about judicial restraint, in that it offered a means of recognizing the superior role of the elected branches of government in making policy choices, as we will see in more detail in Part III. But the superficial consensus should perhaps have always been treated with some skepticism.

The Chevron doctrine is classically expressed as a rigid algorithm (the famous two steps) which makes any deviation by the Court quite noticeable. On the surface, this algorithmic approach makes Chevron appear different from Skidmore deference, one of its primary alternatives in the administrative law canon. Skidmore dictates that informal interpretations by agencies will be given weight by courts in light of “all those factors which give it power to persuade, if lacking power to control.”30 This is an inherently amorphous standard, especially when compared to the apparent rigidity of the textbook version of the Chevron doctrine. Nevertheless, it is common for questions to arise about how strictly Chevron should be applied as a precedent in different types of cases.31 Despite all the fanfare, it is now well known that the Supreme Court itself applies Chevron inconsistently at best.32 Once this inconsistency became apparent, leading scholars sought to reframe Chevron as a looser set of jurisprudential principles rather than a rigid formula.33 The most influential of these efforts is probably Peter Strauss’ concept of “Chevron space” as contrasted with “Skidmore weight.”34

But inconsistency has also led to doubts about how much Chevron matters. Michael Herz summarizes the situation: “Despite all the attention . . . the “Chevron revolution” never quite happens. This decision, though seen as transformatively important, is honored in the breach, in constant danger of being abandoned, and the subject of perpetual confusion and uncertainty.”35

Given this background, it should come as little surprise that there is now open criticism of Chevron and its progeny on the Supreme Court, and that the criticism has emerged mostly from justices on the conservative side of the Court, albeit in different flavors. Justice Thomas has directly questioned the constitutionality of Chevron.36 Justices Scalia, Alito, and Thomas indicated an interest in reconsidering Auer deference, an offshoot of Chevron (also known as Seminole Rock deference),37 under which a court defers to an agency’s interpretation of its own regulation.38 There is a school of thought that suggests Justice Scalia was beginning to re-consider Chevron itself toward the end of his life, which would have been remarkable given that he was Chevron’s strongest advocate during his early years on the Court.39 Regardless of whether that is true, Scalia was replaced by a justice who had very recently criticized Chevron in terms arguably stronger than those used by Thomas.40

Christopher Walker has suggested that it is possible that Chief Justice Roberts has his own, albeit different, issues with Chevron.41 Walker noted that in 2013 Roberts had disagreed with Justice Scalia about whether an administrative agency should get deference when it is interpreting the boundaries of its own authority.42 Sounding much like a critique of Chevron, the Chief Justice worried about the “vast power” of the administrative state over everyday life.43 He was joined in this critique by Justices Kennedy and Alito. Discussing the ambiguous “major questions” exception to Chevron that the Chief Justice announced in King v. Burwell, Walker writes: “Perhaps the narrowing of Chevron deference in King v. Burwell was not really about major questions. Instead, it could have been the start of a much more systemic narrowing of Chevron’s domain and the Chief Justice’s attempt to relitigate the battle he had previously lost to Justice Scalia . . . .”44

Walker is not alone in thinking that the Chief Justice has been quietly moving the Court away from Chevron deference and toward judicial empowerment vis-à-vis the administrative state.45

If you’ve been keeping score, you will note that in a very short period of time the Chevron doctrine went from having no open opponents on the Supreme Court to having five justices who are willing to either limit its reach or destroy it altogether. This tally does not count Justice Breyer, who has at times urged a flexible context-specific approach to Chevron, which may be partially compatible with some of the conservative justices’ desire to limit deference in specific ways.46

To be clear, this does not mean Chevron is dead.47 Even conservative justices are not always critics of Chevron. In Pereira v. Sessions, the 2018 decision in which Justice Kennedy called for a re-consideration of the doctrine, Justice Alito complained that the Court is ignoring Chevron: “In recent years, several Members of this Court have questioned Chevron’s foundations. But unless the Court has overruled Chevron in a secret decision that has somehow escaped my attention, it remains good law.”48 There are still only two justices (Thomas and Gorsuch) fully on record supporting its reversal. Indeed, there are good reasons to think that a doctrine similar to Chevron deference is unavoidable given the tendency by legislators to deflect difficult policy questions to executive agencies.49 But it does mean that Chevron’s vitality and its reach are suddenly in doubt.

However, new doubts about Chevron at the Supreme Court arguably are not as significant as one might think, given that the Supreme Court has been inconsistent in applying Chevron deference at times. But it is important to remember that the Supreme Court is not the only court in the land. The doctrine has had far more impact in the great number of administrative law cases that are resolved in the circuit courts of appeals.50 Unlike at the High Court, in the circuit courts the choice of deference level (Chevron, Skidmore or de novo), is far more predictive of case outcomes.51 These courts pay close attention to the marching orders they receive from the Supreme Court, and thus it matters a great deal if Supreme Court justices change what they say about when, if, and how this doctrine should apply.

B.     Theoretical Foundations, Tensions and Critiques

In 1989, Justice Scalia delivered an influential lecture that both trumpeted and re-explained the Chevron decision.52 That lecture is remembered and often cited as a key moment in the solidification of the doctrine. Scalia’s justified judicial deference explained in terms of congressional intent and separation of powers would become extremely influential—a rationale that the Court eventually adopted in United States v. Mead.53 Yet, in a less cited passage of his lecture, Scalia frankly acknowledged that there were reasons for constitutional doubt about Chevron:

It is not immediately apparent why a court should ever accept the judgment of an executive agency on a question of law. Indeed, on its face the suggestion seems quite incompatible with Marshall’s aphorism that “[i]t is emphatically the province and duty of the judicial department to say what the law is.” Surely the law, that immutable product of Congress, is what it is, and its content—ultimately to be decided by the courts—cannot be altered or affected by what the Executive thinks about it.54

The important thing about this passage is that it recognizes a baseline in which deference is not the norm. Because Chevron has become so widely accepted over the past quarter century, it may seem that deference to agencies is assumed, and anyone who wants to avoid deference is swimming upstream. But in these earlier days, Justice Scalia acknowledged that non-deference should actually be the starting point. Put another way, courts should give the authoritative interpretation of the law, unless there is good reason to do otherwise. There are three main justifications that have been given for why the judicial branch should defer to executive agencies on the interpretation of ambiguous statutes: technical expertise, political accountability and congressional intent.

The expertise explanation carries the oldest pedigree, and is the rationale that most clearly anchors Chevron in a longer administrative law heritage. Appellate review of administrative agencies evolved with regulation of the railroad rates more than a century ago.55 After considerable struggle, the Court eventually realized that the Interstate Commerce Commission was better suited to set railroad rates.56 Despite having sometimes derided technical expertise as a reason for deference, Justice Scalia himself sometimes fell back on it.57

However, as Lawson and Kam have explained, “[e]pistemological deference . . . does not require any specific doctrine for implementation.”58 If a doctrine is needed, Skidmore deference would appear to suffice; a judge should defer to the technical experts because they are persuasive (and only if they are persuasive). It should be noted that many scholars argue that expertise may consist of an agency’s intricate knowledge of a statute’s technical background, which I explore in more detail in Section III.B. For present purposes it is enough to note that rigid deference to technical expertise seems to undermine the judiciary in favor of technocrats, since it is ultimately the job of judges to make decisions about the law, even when someone else has a better chance of getting the technicalities right.59

The Chevron decision itself emphasizes political accountability as at least as important as technical expertise as a justification for deference:

Judges are not experts in the field, and are not part of either political branch of the Government. Courts must, in some cases, reconcile competing political interests, but not on the basis of the judges’ personal policy preferences. In contrast, an agency to which Congress has delegated policymaking responsibilities may, within the limits of that delegation, properly rely upon the incumbent administration’s views of wise policy to inform its judgments. While agencies are not directly accountable to the people, the Chief Executive is, and it is entirely appropriate for this political branch of the Government to make such policy choices—resolving the competing interests which Congress itself either inadvertently did not resolve, or intentionally left to be resolved by the agency charged with the administration of the statute in light of everyday realities.60

This rationale has become a prominent argument in defense of the administrative state, advocated prominently by Elena Kagan (during her days as a law professor), among others.61 Agencies can change their minds, and voters can change the agencies by voting for a new president. The political accountability rationale for congressional delegation is rooted in foundational constitutional principles governing separation of powers. Courts say what the law is, but the political branches make the policy choices. When a statutory ambiguity left by Congress leaves a policy choice open—better for the courts to defer to the other political branch to make that choice. A related argument for judicial deference is that it may lessen the influence of the political preference of judges.62

The political accountability rationale for Chevron may have been given a recent re-boost by the D.C. Circuit. In its June 2017 decision in Global Tel*Link v. F.C.C., the D.C. Court of Appeals held that when an agency no longer asks a court to give its rule deference, “it would make no sense for this court to determine whether the disputed agency positions advanced in the Order warrant Chevron deference when the agency has abandoned those positions.”63 Critically, this would allow an agency to change course without actually going through the full process necessary to promulgate a new rule. The agency can simply tell a reviewing court that it does not want judicial deference. Such a change is especially likely to occur (and in Global Tel*Link, did occur) when a new presidential administration takes office, and thus is empowering to voters.

The rationale that the Supreme Court has relied on most in recent years is one of congressional intent, particularly since its 2001 decision in United States v. Mead Corp.64 This theory is utilized when a statute gives an agency responsibility for a particular area of public policy and simultaneously leaves a particular policy question, with regard to how that agency should operate, ambiguous—the theory supposes that Congress implicitly, but nevertheless, intentionally, delegated to the agency the authority to decide how to act. Mead’s emphasis on congressional intent gave rise to the recognition that in addition to Chevron’s two steps, there is a Step Zero.65 After all, if congressional intent to delegate is the foundation for deference, then judges must start their analysis by asking if Congress actually intended to delegate the relevant authority in the first place. This necessary preliminary step addresses a tension between Chevron’s presumption that all ambiguous statutes imply a delegation of power by Congress, while the Mead decision seemed to insist on more explicit terms of delegation.66

The three rationales for deference do not exist in isolation from each other. For example, Congress might delegate a policy choice to an agency because of its technical expertise. Congress might delegate because it thinks that the public will have more confidence in a policy if it is administered by a trusted agency.67 This is how Scalia explained the delegation rationale for Chevron in his 1989 lecture.68

He argued that when a statute is ambiguous, the interpretation may involve a policy choice.69 He said, “[u]nder our democratic system, policy judgments are not for the courts but for the political branches.”70 In this way, Chevron is really about judicial modesty and restraint.

A focus on the justifications for deference has become important as it has become increasingly clear that the Court does not apply deference consistently. The famous two steps of the Chevron doctrine (plus the newer Step Zero) imply a rigid formula, but that conceptualization of how the doctrine should have been implemented may never have been realistic. Moreover, the steps are not created equal. We know from empirical research that when Chevron applies, its first step (is the statute ambiguous?) is nearly always decisive.71 But a lot goes into whether judges are actually willing to apply Chevron, and whether they consider statutes to be ambiguous. This is why recent scholarship has leaned toward looser approaches to Chevron, such as Strauss’ notion of “Chevron space.”72 As Strauss describes it, despite apparent inconsistency, Chevron embodies a coherent approach to allocating authority to agencies.73 To some extent this is simply a change in terminology that reduces expectations for perfect consistency. But it also emphasizes the need for judges to think carefully in each case about how much “space” a particular agency should be allotted. Put another way, Chevron is a doctrine embodying important values of judicial restraint, much as Scalia advocated back in 1989, but not a rigid rule.

Even for some justices who ostensibly are committed to Chevron, the desire to respect congressional intent co-exists with an increasingly evident “inclination toward judicial empowerment.”74 This certainly can be ascribed to simple self-interest; judges may only be willing to restrain the powers of their branch so much. It also has links to growing confidence among judges in their ability to reach definitive interpretations of statutory texts, either through textualism or purposivism.75 These methodological approaches allow judges to determine meaning from language in statutes that might easily be considered ambiguous under Chevron’s first prong. But there is also a core separation of powers argument that the judiciary should not be too restrained, because it is needed as an independent judicial check on the power of the administrative state.76

 C.     An Increasing Context-Specific Orientation?

Justice Breyer has advocated a “context-specific” approach to deference, in which the Court should not always presume that statutory ambiguity warrants deference to an agency.77 Breyer would assess a number of broad factors, including the nature and importance of the legal question, the nature of the agency’s expertise, the complexity of the policy in question, and the process by which the agency reached its interpretation.78 Walker surmises that Breyer’s context-specific approach might often be roughly compatible with the inclinations of the Chief Justice.79 This raises the possibility that the Court may be moving toward a context-specific Chevron, and lends credence to those who have called for a “tailored approach” in which norms of administrative law may be adjusted to different policy and agency contexts, including immigration.80

The Chevron doctrine as it is classically presented presumes that a single set of analytical steps will be able to cope with the myriad legal problems presented by our sprawling administrative state. An approach used by judges in an environmental case involving air pollution by an energy company (Chevron) gets extended in a case concerning net neutrality and an internet service provider (Brand X)81 and extended again in a labor law case (Auer).82 This one-size-fits-all understanding may obscure the degree to which the legitimacy of the administrative state is built on several competing theories about accountability, each of which competes to a certain extent with the others.83 If our constitutional baseline is actually non-deference as Scalia suggested in his 1989 lecture, it stands to reason that the arguments to depart from that norm—expertise, political accountability, and congressional intent—will be more persuasive in some contexts than in others. Moreover, some areas of agency action may present additional compelling reasons for courts to not act deferentially.

Inherent in a context-specific approach (and in the existence of a Step Zero) is that deference has limits. For an illustration of a situation where there is a constitutional reason to not defer to an agency despite a congressional delegation, consider United States v. Booker, where the Court found that sentencing guidelines issued by the Federal Sentencing Commission could be advisory only.84 The Court rationalized that if the guidelines were mandatory, those mandatory guidelines would violate the Sixth Amendment right to have factual allegations determined by a jury.85 The Government tried to defend the sentencing guidelines because they were issued by a special commission, rather than by Congress itself, but the Court held that this delegation “lacks constitutional significance.”86 In other words, a role assigned to a jury by the Constitution could not be delegated by Congress to anyone else.

Since Booker focused on criminal sentencing, it would be reasonable to ask what relationship it has with Chevron. As we will see later in Section III, criminal law and immigration cases are not particularly easy to separate. But for present purposes, the point is simply to illustrate that the congressional delegation theory has limits. In the context of Booker, the Sentencing Commission was established and acted much the same way a congressionally-authorized agency would act in administrative law. Other scholars have critiqued Booker by invoking broader administrative law principles, arguing that the courts should not re-delegate to themselves a role that Congress delegated to a specialized commission or agency.87 They may very well be right—but only if it is a role that Congress has the authority to delegate in the first place.

The connection to criminal law brings us to Justice Gorsuch and the critique he levied against Chevron when he was a Court of Appeals judge, just a few months before his nomination to the Supreme Court.88 In the Tenth Circuit case of Gutierrez-Brizuela v. Lynch, then-Judge Gorsuch returned to the baseline rule that Scalia articulated in 1989, emphasizing the primary role of the judiciary to articulate what the law is.89 But Gorsuch expressed considerable suspicion about rationales for deference that Scalia had once found persuasive. For instance, Gorsuch expressed concern citizens would be unable to influence agencies “without an army of perfumed lawyers and lobbyists.”90 In other words, Gorsuch suggested that political accountability—which is what makes the political branches open to lobbyists—might be an institutional weakness in some situations. He endorsed the critique, well-developed in legal scholarship, that Chevron may conflict with the Administrative Procedure Act.91 He indicated that he shares Justice Thomas’ skepticism about whether Congress can actually delegate its lawmaking functions to the executive branch.92

At the center of his critique of Chevron, Gorsuch focused on the dangers to liberty posed by too much judicial deference. For this, he referenced the rule that Chevron does not apply to criminal law.93 Rather than treat this as a formalistic and mechanical line between the civil and criminal contexts, Judge Gorsuch looked at the reasons for the rule, primarily the fear that criminal sanctions carry a special danger of abuse.94 He saw no reason why this rationale would be limited strictly to criminal prosecutions, given that civil, administrative actions can also “penalize persons in ways that can destroy their livelihoods and intrude on their liberty.”95 Judge Gorsuch wrote: “Under any conception of our separation of powers, I would have thought powerful and centralized authorities like today’s administrative agencies would have warranted less deference from other branches, not more.”96

I would point out that, just as the arguments for Chevron may be more persuasive in some contexts than in others, the arguments against it may be context-specific as well.97 In short, the more power an agency acquires and perhaps the less capable its subjects are of being heard through the political process, the more important it is for the judiciary to be a robust check and balance. In this light, it is not irrelevant that Gorsuch raised concerns about the power of the administrative state in the context of an immigration case.

D.     Loud and Soft Approaches

At the Supreme Court, Chevron has long been dogged by a gap between the Supreme Court’s proclamations, and the Court’s subsequent actions. This inconsistency can certainly be explained, and to some extent probably should be explained, by simple human nature. It is easier for Supreme Court justices to state a norm of judicial restraint than it is to practice restraint consistently. Chevron’s two steps ask judges to recognize that the law is often not clear and that there is more than one reasonable interpretation. But these are justices who are accustomed to definitively stating what the law is, and explaining their conclusions (and sometimes arguing with each other) with high levels of apparent self-confidence. It short, perhaps the Supreme Court will not consistently defer to anyone.98

It is also possible that the Court’s inconsistency reflects an underlying ambivalence. The justices wrestle with challenging problems and may sometimes be unsure of their footing—they may change their perspective as they see the doctrine’s potential application in different situations. This possibility is consistent with the fact that over time individual justices argue for both sides of the issue and appear to contradict themselves. Perhaps this is most evident with Justice Thomas, who reiterated a highly deferential rendition of the Chevron doctrine in his Brand X decision for the Court: “If a statute is ambiguous, and if the implementing agency’s construction is reasonable, Chevron requires a federal court to accept the agency’s construction of the statute, even if the agency’s reading differs from what the court believes is the best statutory interpretation.”99

The Brand X articulation of Chevron is especially deferential because it states bluntly that judges should permit legal interpretations that they sincerely do not think are the best. And yet, ten years later, Justice Thomas wrote this, derisively quoting back to his own majority opinion:

Interpreting federal statutes—including ambiguous ones administered by an agency—“calls for that exercise of independent judgment.” Chevron deference precludes judges from exercising that judgment, forcing them to abandon what they believe is “the best reading of an ambiguous statute” in favor of an agency’s construction. It thus wrests from Courts the ultimate interpretative authority to “say what the law is,” and hands it over to the Executive.100

Given the organic manner in which Chevron grew from an air pollution case into the primary organizing canon of administrative law, it seems reasonable to conclude that the justices have been working through the dynamics of the doctrine case by case. While sometimes the inconsistencies can be attributed to natural human vacillation, the justices have likely re-analyzed the issue, or discovered that Chevron was not as appropriate in a particular case as the original doctrine might have previously implied. Perhaps the Court’s articulation of the Chevron doctrine may have been premature. Whereas the justices might have once been enamored with Chevron’s apparently broad usefulness, “the Court’s recent treatment of Chevron [has been] as a doctrine to ignore, disparage, or distinguish.”101

Despite this recent trend, for most of the justices, the current situation may be more about discovering Chevron’s limits, rather than dismantling it entirely. But this makes for a somewhat difficult interpretive moment, because we still have sweeping articulations of the doctrine on the books in cases like Brand X. Therefore, methodologically one might consider paying attention to two different indicators of the state of the law—what I will call loud approaches to Chevron and soft ones.102

Loud approaches include cases where the Court (or an individual justice) directly articulates a particular understanding of Chevron’s reach or limitations. The majority opinions in Mead, Brand X or City of Arlington would be paradigmatic loud, pro-Chevron decisions. Justice Thomas’ opinion in Michigan v. E.P.A. or then-Judge Gorsuch’s opinion in Gutierrez-Brizuela would be loud anti-Chevron opinions. They leave no real ambiguity about where those judges stand at the moment they issued them. We should also include here Chief Justice Roberts’ majority decision in King v. Burwell, because it openly articulates a significant limitation on Chevron’s application through the major cases exception. For reasons explained elsewhere, this exception is not well defined, nor is it entirely clear where all of the justices who signed the majority opinion really stand.103

But it is “loud” in the sense that it explicitly states a rule even if it is a poorly defined rule. Anytime a court explicitly adds (or takes away) an element of Chevron’s new “Step Zero,” it is issuing a loud decision about Chevron’s reach. The D.C. Circuit’s recent decision in Global Tel*Link to not apply Chevron in cases where a new Administration does not request such deference would be another example.104

Soft approaches to Chevron (or, more likely of late, against Chevron) can be harder to interpret but are quite common. Other scholars have noted that there are glaring and fairly obvious cases where the Court should have applied Chevron, according to the way it has been articulated, but simply failed to do so.105 “Failure to apply Chevron where it would seem to apply” could be a way for justices to indicate their concern with a “full-throated Chevron doctrine.”106 These instances come in two varieties, as we will see in our discussion of recent criminal removal cases. In one type, call it Type I, the Court simply decides the case without even mentioning Chevron. As we will see, this does not lead to a situation where the agency always loses. But it does mean that the Court acted as if Chevron deference did not exist. The silence on Chevron in these cases is especially noticeable because the Solicitor General typically asks for deference to the agency’s interpretation.

A different variety of soft anti-Chevron cases, call them Type II, are those in which the Court mentions Chevron, but renders it irrelevant, but without articulating a new formal exception or limitation on the doctrine. This is typically done through application of Chevron’s Step One, when the Court concludes that the statutory text is sufficiently clear. If the statute is actually clear, then this result is what we would expect. However, the cases that I include in Type II are those where the statute is honestly ambiguous, if ambiguity is to have any real meaning. The language is oblique or open-ended. Often, there will be a circuit split, or at least a division between the agency and a circuit, which is a likely reason why the Supreme Court might take on such a case. Of course, it is the Supreme Court’s job to resolve such disputes about what a statute means. But such disputes make it hard to accept that the statute is not really ambiguous, if ambiguity is to have any concrete meaning. In 2018, Justice Alito criticized his colleagues for failing to meaningfully apply Chevron in precisely this kind of case in his dissent in Pereira v. Sessions.107

Another example of Type II can be seen from a 2017 Supreme Court decision in Esquivel-Quintana v. Sessions, in which the Court dealt with the statutory meaning of “sexual abuse of a minor.”108 The case presented the question of whether the rule of lenity should impact the application of Chevron’s Step Two, a question that also arose at the Court a year earlier in Torres v. Lynch.109 In Torres, the Court simply ignored the issue and never even mentioned Chevron, rendering a Type I soft anti-Chevron decision. A year later, in Esquivel-Quintana, the Court dispensed with Chevron as follows:

[P]etitioner and the Government debate whether the Board’s interpretation . . . is entitled to deference under Chevron . . . . We have no need to resolve whether the rule of lenity or Chevron receives priority in this case because the statute, read in context, unambiguously forecloses the Board’s interpretation. Therefore, neither the rule of lenity nor Chevron applies.110

This is a quintessential Type II soft decision.

Type II cases are interesting in that they highlight a longstanding and well-known reality of Chevron: The action is usually at Chevron’s Step One. Moreover, increasing interest in the meaning of statutory texts can work to reinforce a trend toward a stronger judicial role, and thus less deference.111 Armed with strong enough tools of statutory interpretation, a judge (or a group of nine judges) can always arrive at a conclusion about what a statute means, after which the statute will no longer be ambiguous.112 But then, what is Chevron for? What happened to the range of acceptable policy choices that should be left to the political branches?

Such soft approaches to Chevron are nothing new.113 During Chevron’s first formative decade, Thomas Merrill wrote that the Court’s inconsistency was as an indicator that there were problems with “the draconian implications of the doctrine for the balance of power among the branches, and to practical problems generated by its all-or-nothing approach to the deference question.”114 Of course, for quite a while the Supreme Court justices were ostensibly all committed to the Chevron doctrine. A later study concluded that continued inconsistency by the Supreme Court justices called for “caution about the Court’s collective ability to follow any doctrinal framework consistently.”115 Nevertheless, lower courts attempted to energetically follow Chevron because the Supreme Court had prescribed it,116 while the Supreme Court itself often “honored in the breach.”117 While somewhat confusing, this asymmetric application of Chevron was not necessarily fatal to the doctrine.118

However, currently some justices have openly called Chevron an abdication of judicial duty, while several others have indicated interest in carving out clearer limits to the doctrine. This puts the Court’s longstanding inconsistency in the actual application of the doctrine in a new light, suggesting that Merrill got it right early on. When the Court does not apply Chevron in cases where it appears it should—a soft anti-Chevron decision—we should understand it as a possible sign that the justices might have found the doctrine less than useful in those cases. Seen this way, these soft anti-Chevron decisions can play a useful purpose in the refinement of the doctrine. If it is correct that Chevron’s most fatal flaw is its rigidity and one-size-fits-all orientation, this process of quiet testing across the wide variety of administrative law cases is potentially quite healthy. But it needs to be followed up in two ways. First, lawyers need to look for patterns in the inconsistency, so that we learn when Chevron is applied and when it is not.119 Second, eventually the Court needs to explain what it is doing, so that lower courts know their marching orders, and so that the Supreme Court does not appear permanently confused or arbitrary.

III.     Chevron in Immigration Cases

A.     The Wide Variety of “Immigration Law”

In theory, there could be strong arguments for judicial deference in immigration cases. The two most obvious are that immigration touches on foreign policy concerns and that it is an immensely complicated area of law and policy. And yet, not all immigration cases are the same. Although often cited as a rationale for federal power, the impact on foreign policy is not clear in all immigration cases.120 Recent scholarship has raised doubt as to whether foreign policy remains as important a rationale for immigration jurisprudence as it once was.121 If it is convincing in any situation, the foreign policy rationale would seem to be most convincing in cases that involve requests for visas for foreigners who are actually abroad. Those situations most directly invoke the relationship of the United States to the rest of the world, in that it involves consulates dealing directly with a non-citizen in a foreign country. The foreign policy implications are signaled by the fact that Congress assigns the State Department to play a significant role in the process. Immigration law has long recognized a significant constitutional difference for non-citizens who are already in the United States and those who are trying to enter.122 When the government takes action against a person who is inside the country, the foreign policy implications are more attenuated. That is not to say there are never foreign policy ramifications of domestic immigration policy, but there are occasionally foreign policy dimensions of many types of domestic policy.

These distinctions are just the beginning of a complicated taxonomy of issues in immigration cases. Some questions are explicitly committed by Congress to the discretion of the Department of Homeland Security or the Attorney General; these matters are typically beyond the jurisdiction of the courts to begin with. Other questions involve applications for relief by people who are otherwise legally removable from the United States. Central to this Article’s argument is that some questions of immigration law determine whether a person who would otherwise be free in the United States may be arrested, incarcerated (for months or years) and removed.123 Others do not. Finally, some matters of immigration law are grounds for criminal prosecution.124 As Matthew J. Lindsay has written, the presumption that all laws regulating non-citizens form a single, coherent body of law may itself be the source of doctrinal confusion.125

The diversity of contexts within the field of immigration law is extremely important if the Chevron doctrine is moving toward a context-specific orientation, and away from a rigid one-size-fits-all paradigm, as discussed in Section II.C.

B.     Process Problems with Chevron Deference in Immigration Cases

Before delving into different concerns raised by different types of immigration cases, it is important to recognize that there are procedural issues that may weaken the justifications for judicial deference in many immigration cases. A typical removal case (or in more plain language, a deportation case), begins in Immigration Court, which is part of the Executive Office of Immigration Review (“EOIR”) within the Department of Justice.126 The Immigration Judge can issue an order of removal that can be appealed to the Board of Immigration Appeals (“BIA”), which is also part of EOIR.127 Once the BIA resolves an appeal, any order of removal becomes final.128 A final order of removal may be appealed by filing a petition for review with the U.S. Court of Appeals, which has jurisdiction over the state where the immigration court was located.129

In practical terms in these cases, the BIA is the agency whose decision is usually under review by a court; thus, if Chevron deference were applied, a federal court of appeals would defer to the BIA on its interpretation of immigration law. This raises questions about the justifications for deference. Technical expertise is certainly a potential rationale for the BIA’s role, but it is not necessarily compelling. Immigration laws are certainly complicated, but they are not technical in a scientific sense like the subjects of other areas of administrative regulation. To be clear, the expertise rationale for deference need not be limited to scientific knowledge. Instead, because agencies often have unique knowledge of the legislative history and policy context for legislation, it may be justifiable to defer to their interpretation of that legislation.130 However, the BIA is designed to be fairly isolated within the immigration bureaucracy. It is not involved in crafting immigration legislation, nor in setting enforcement policy—a matter left to the Department of Homeland Security. It does not have a major role in granting or reviewing most visa applications, because these roles are left to DHS and the State Department. It interprets the statute through case-by-case adjudication, much like the courts. The federal courts of appeals are experts in statutory interpretation, so it is harder to argue that the BIA has a technical expertise advantage in interpreting immigration law.131 There is also a cogent argument that the BIA has no particular expertise on state criminal law, which weakens the argument for deference when a removal case focuses on a state criminal conviction.132 The bottom line from all this is that, while immigration law is complicated, it is not self-evident that it is beyond the capacities of federal judges to master, nor that the BIA brings anything to the table beyond what a federal court could provide.

Doubts about the BIA’s comparative competence relative to the judiciary were heightened in 2002 when the George W. Bush Administration streamlined the BIA, the result being that the agency now devotes far fewer resources to deciding each case.133 Since then, federal judges have issued scathing decisions about the quality of BIA decision-making.134 At the same time, Immigration Courts are famously backlogged and under-resourced.135 Some courts have raised these resource concerns as reasons to doubt the utility of deferring to the BIA.136 Moreover, there is substantial empirical evidence that the Department of Justice’s decision-making may be highly inconsistent from one immigration judge to the next, suggesting that “the most important moment in an asylum case is the instant in which a clerk randomly assigns an application to a particular asylum officer or immigration judge.”137

We should not be particularly surprised that some judges who review many BIA decisions are less willing to defer to the BIA. In 2005, the Seventh Circuit’s Judge Richard Posner wrote: “[T]he adjudication of these cases at the administrative level has fallen below the minimum standards of legal justice.”138 According to a study by David Zaring on the Court of Appeals for the D.C. Circuit, judges may become less likely to affirm decisions by agencies that appear before them more frequently.139 After the 2002 reforms to the BIA, petitions to the federal courts surged.140 Immigration cases came to represent nearly a fifth of the dockets of the Second and Ninth Circuits.141 One result of this is that any judge on those courts is likely to become familiar with immigration law issues. Of course, judges on circuits that see fewer immigration matters might not develop the same familiarity. But they could just as easily rely on the precedent rulings of their sister circuits, rather than defer to the executive branch.

It was noted early in Chevron’s history that the political accountability theory does not work well with agencies that are independent and thus insulated from legislative or presidential influence.142 According to its governing regulation, the BIA should not consider politics in deciding how to interpret immigration law: “Board members shall exercise their independent judgment and discretion in considering and determining the cases coming before the Board.”143 BIA members are career appointees, meaning that federal law prohibits considering political affiliation in hiring.144 Electing a new president thus would do little to change the BIA. It is also relevant that the BIA has no influence on foreign policy decisions, which makes reference to the Executive Branch’s supremacy in international relations a less convincing rationale for deference to the BIA.

These objections to the BIA require an important caveat: The Attorney General can overrule the BIA.145 The Attorney General retains the authority “to exercise full decision-making upon review.”146 This includes the power to make independent findings of facts and law.147 President Obama’s Attorneys General used this authority sparingly, but there have been calls to use this power more aggressively.148 Attorney General Jeff Sessions is doing just that.149 Beyond issuing precedent decisions that change legal interpretations, during the first year of the Trump Administration, Department of Justice leadership took several steps to assert stronger control over the operational management of the Immigration Courts, which led to protests that their independence was in question.150

During the George W. Bush Administration, the Department of Justice emphasized the fact that members of the Board are mere employees “who [are] appointed by, and may be removed or reassigned by, the Attorney General.”151 Indeed, there have long been arguments that the BIA was never as independent from political interference as it should have been.152 This view of the BIA raises an interesting thought experiment about the various explanations for deference. In most cases, the personal involvement of the Attorney General would not add technical expertise relative to the more specialized BIA, but it would add political accountability. Unlike BIA members, voters have the ability to change the Attorney General by electing a new president, and by electing senators who must consent to the appointment. But does this mean that courts should defer more to decisions actually made by the Attorney General, and less to those more common decisions made by the BIA? Or does the mere potential for the Attorney General to intervene create a presumption that the BIA decision belongs to the cabinet-level appointee? For what it is worth, the courts imply this is the case in the naming of the cases. In the federal courts, immigration petitions for review are captioned “[NAME OF PETITIONER] v. [NAME OF CURRENT ATTORNEY GENERAL].”

The key point is that on close examination the standard rationales for Chevron deference do not apply with equally persuasive force to all agencies and to all decisions. In particular, the technical expertise rationale is less persuasive in immigration cases than it might be in a regulatory arena requiring more scientific or technical knowledge beyond the training of federal judges. The political accountability rationale might be coherent in immigration cases, but only if we rely on the theoretical involvement of the Attorney General. But these questions are not always decisive in any case. In terms of the interpretation of immigration law, congressional intent to delegate is clear and explicit. Congress has explicitly entrusted “questions of law” under the Immigration and Nationality Act to the Attorney General.153

 C.     Non-Removal Immigration

Immigration cases have been a challenge for the Chevron doctrine essentially from the beginning. In 1987, the Court decided I.N.S. v. Cardoza-Fonseca, which remains one of the seminal cases in U.S. asylum law.154 The Court overruled an interpretation by the BIA, but in a majority decision authored by Justice Stevens the Court gave two explanations for this holding. First, the Court concluded that the question at hand could be decided based on the language of the statute, and thus “there is simply no need and thus no justification for a discussion of whether the interpretation is entitled to deference.”155 This appears to be a clear application of Chevron Step One. But the decision also said that Chevron was not applicable to “a pure question of statutory construction.”156Chevron, Justice Stevens suggested, was more appropriate when an agency applies laws to particular facts.157 This was an early expression of doubt, from no less than the author of Chevron itself.158 Had the Court embraced this approach, we might have had an early Step Zero limitation on the doctrine. But the Court abandoned it quickly.159

Although Chevron deference was not applied in Cardoza-Fonseca, it was cited and discussed extensively by the Court, suggesting that it was potentially applicable to immigration cases.160 Yet, in a 1992 case concerning eligibility for asylum, the government asked for application of Chevron deference,161 but the Court made no reference to Chevron in its decision.162 The Court clarified matters in I.N.S. v. Aguirre-Aguirre, another case concerning eligibility for asylum.163 This time the Court affirmed that “[i]t is clear that principles of Chevron deference are applicable to this statutory scheme.”164 The Court explained that Congress has explicitly delegated to the Attorney General authority to decide questions of law under the Immigration and Nationality Act.165 Deference did not seem determinative of the result, however. The Court agreed with the BIA’s interpretation of the statute as the best reading of the text.166 When a Court would have reached the same interpretation of a statute as the agency did, deference does not impact the result. But Aguirre-Aguirre nevertheless seemed to firmly establish that Chevron applies in immigration cases.

Since Aguirre-Aguirre, the Court has more consistently deferred to the Attorney General in cases concerning eligibility for asylum. Even in cases where the government lost, the Court sometimes applied the ordinary remand rule to send the case back for administrative interpretation in the first instance.167 The ordinary remand rule is a means by which a court can vacate an agency decision, but still respect deference by asking the agency to give an interpretation of an ambiguous statute, rather than have the court impose one.168 The Court has used this rule in three asylum cases, Negusie v. Holder and two other per curiam decisions.169 I should also note that I am not including the 1988 case of I.N.S. v. Abudu, which concerned the standard of review on a procedural motion to re-open an asylum application.170 Because there was no statutory question involved, Chevron was not applicable, and thus may not be relevant to this analysis. But to the degree that it matters, the Court ruled for the government and applied the deferential abuse-of-discretion standard of review, and so it broadly fits the same general pattern.171

The Supreme Court’s most recent full throated embrace of Chevron in an immigration case came in its 2014 decision in Scialabba v. Cuellar de Osorio, which concerned eligibility for a visa based on family sponsorship.172 This decision includes the Court’s most robust articulation of a general rule requiring deference in immigration cases: “Principles of Chevron deference apply when the BIA interprets the immigration laws. Indeed, ‘judicial deference to the Executive Branch is especially appropriate in the immigration context,’ where decisions about a complex statutory scheme often implicate foreign relations.”173 In addition to this broad statement, deference appears to have been important to the result in Cuellar de Osorio, at least for the plurality. The BIA had adopted an interpretation of the statute that caused some immigrant families to wait years to be reunified in the United States. Justice Kagan wrote that the statute “makes possible alternative reasonable constructions.”174 She emphasized that the BIA did not have to decide the matter as it did. In a clear statement of deference, Justice Kagan wrote: “[W]e hold only that § 1153(h)(3) permits—not that it requires—the Board’s decision.”175 The Chief Justice, joined by Justice Scalia, concurred in the judgment, disputing to some extent whether the statute was ambiguous, but agreeing that Chevron deference applied.176

It would not be true to say that the Court has been consistent in these cases, since the Court failed to address Chevron in an early case where it would seem to be applicable.177 But after an early false start, and especially since Aguirre-Aguirre, the Court seemed to re-affirm the applicability of Chevron deference in immigration cases, especially in cases like Negusie and Cuellar de Osorio. Moreover, in several of these cases the deference impacted the result. But there is a very large caveat. None of these cases concerned grounds of deportation, nor detention of immigrants. Rather, they concerned eligibility for immigration benefits, or (in the case of asylum) relief from removal for immigrants who were deportable for some other reason.178 These cases are summarized in Table 1. But as we will see in Section III.D, the Court behaves very differently in how it applies Chevron in deportation related immigration cases. 

Kagan Pic 1

D.     Criminal Grounds of Removal Cases

Many of the Supreme Court’s recent encounters with immigration law concern the intersection of immigration and criminal law, a field sometimes known as “crimmigration,”200 and often focusing on analysis of state criminal codes. For purposes of this Article, the main legal problem, and the primary dilemma concerning the appropriateness of Chevron, concerns the categorical approach to analyzing criminal convictions. Immigration law shares an interpretative problem with federal sentencing law. In both areas of law, heightened federal action against a person (e.g., deportation or enhanced prison terms) may be triggered by convictions for certain offenses under state law. The difficulty is that state criminal codes do not define crimes the same way federal law does. That is where the categorical approach comes in.

For example, the Armed Career Criminal Act severely enhances the sentence for someone convicted of unlawful possession of a firearm and who has also been previously convicted of “burglary.”201 Similarly, the Immigration and Nationality Act lists “burglary offense” as an aggravated felony that would trigger both removal and mandatory detention.202 The problem is that definitions vary as to what constitutes a burglary.203 The Model Penal Code defined burglary as breaking into any occupied structure, unless the building was open to the public or the perpetrator had permission to enter.204

But some states define burglary more broadly so as to include entry to any building, even one that is open to the public, so long as the entry is made with an intent to commit another crime.205 These differences make it difficult to decide whether a state burglary conviction should count as a burglary conviction for the purposes of federal sentencing and immigration law.

The Court began to address this problem in 1990 with Taylor v. United States, giving birth to the categorical approach.206 The categorical approach requires a court to define the elements of a particular crime under federal law, and then compare this to the elements required for conviction under the state law.207 If the state definition of the crime is broader—meaning it criminalizes more conduct than the federal definition—then there would be no categorical match.208 Initially, federal authorities were able to overcome this problem through the “modified categorical approach,” through which they could submit evidence that the person had engaged in conduct violating the more narrow federal definition.209 But in 2013 and 2014, the Court decided several cases that reinforced the categorical approach and imposed strict limits on when the “modified” approach could be used.210 This generally benefits immigrants, because it means that grounds of removal must be interpreted quite strictly and that fewer state convictions should lead to detention and deportation by the Department of Homeland Security. But first the Court had to make clear whether this approach applies in immigration cases.

In 2013, the Court returned to the burglary question in Descamps v. United States, a case that—just like Taylor—did not concern an immigrant. In Descamps, the Court found that when a state burglary statute does not require that a defendant have entered a building unlawfully, it is categorically not a generic burglary under federal law.211 The state burglary conviction at issue was “missing an element” from the generic definition of the crime.212 In Descamps, this meant that a federal sentencing enhancement did not apply. This was a criminal case, and the word “Chevron” appears nowhere in the decision.

However, less than two months before Descamps, the Court had decided an immigration case, Moncrieffe v. Holder.213Moncrieffe concerned a legal resident of the United States who was arrested for marijuana possession.214 He was convicted under a Georgia statute that punished both simple possession and distribution and sale of marijuana.215 The Department of Homeland Security sought to deport Mr. Moncrieffe as an aggravated felon, because the Immigration and Nationality Act’s definition of aggravated felon includes convictions for “illicit trafficking in a controlled substance.”216 However, the federal law contained an exception for “distributing a small amount of marihuana for no remuneration.”217 Because the Georgia statute was broader—it swept in simple possession and distribution without remuneration, as well as actual sale—it was not a “categorical match.”218 Justice Sotomayor’s opinion for the Court stated:

Under this approach we look not to the facts of the particular prior case, but instead to whether the state statute defining the crime of conviction categorically fits within the generic federal definition of a corresponding aggravated felony. . . . Whether the noncitizen’s actual conduct involved such facts is quite irrelevant.219

The import of this was that there was justification for the BIA to consider any other evidence once it was clear that the statutory elements of the conviction were overbroad.

There is good reason to think that the categorical approach inherently leaves little room for judicial deference.220 In Moncrieffe, just as in Descamps, the word “Chevron” does not appear. This absence was more noteworthy in Moncrieffe because this case involved a petition to review an administrative order of removal, and as we have seen Congress has nominally entrusted questions of law in this arena to the Attorney General. Following these two decisions, the BIA decided that it was bound to follow Descamps in immigration cases.221 The Department of Homeland Security argued that Descamps should not apply outside the criminal context, but the Board concluded that “Descamps itself makes no distinction between the criminal and immigration contexts.”222 This could have turned out quite differently, if the civil-criminal distinction had determined the reach of Descamps, or if the Board had concluded that it was free under Brand X to reach a different answer to an ambiguous question than that prescribed by the Court. Moncrieffe is thus a prototypical example of a soft anti-Chevron decision, and a fairly potent one at that.

This is a pattern. In at least seven decisions (including Moncrieffe) concerning the BIA’s interpretation of criminal grounds of removal the Supreme Court has simply failed to even mention the existence of Chevron.223 To be clear, these cases are not all alike, and several of them on their own might not raise doubts about the viability of Chevron. In some cases the Department of Justice did not ask for deference, usually because there was no published BIA decision at issue.224 In another case, the government asked for deference, but only in a footnote to its brief, and did not appear to demand Chevron deference specifically.225 But those factors cannot explain the pattern. In two cases, Nijhawan v. Holder and Torres v. Lynch, Board decisions had been published and the government asked vigorously and at length for Chevron deference, the Court still ignored Chevron entirely in its decision.226

In addition to these six cases, there are two other criminal grounds of removal cases in which the Court mentioned Chevron, but did not actually give deference. Most recently, in Esquivel-Quintana v. Sessions, the Court again dealt with the statutory definition of an aggravated felony, specifically whether certain state statutory rape offenses qualified as “sexual abuse of a minor.”227 The government sought to deport an immigrant on the basis of a California statutory rape criminal statute which required only a three-year age difference if the purported victim was under 18 years old, thus criminalizing sex between a 21-year-old and a 17-year-old.228 This case contained a published agency decision and a circuit split on the question.229 A divided Sixth Circuit panel upheld the BIA.230 The Supreme Court dispensed with Chevron in one sentence, stating: “[T]he statute, read in context, unambiguously forecloses the Board’s interpretation.”231 Superficially, this is a Chevron Step One decision, finding that the statute was not ambiguous. Yet, the statute contains no definition of “sexual abuse of a minor,” and the phrase hardly offers a self-evident meaning on its face. While it seems safe to assume that rape of an elementary school child would qualify, marginal cases that involve older teenagers and lesser forms of assault had long troubled the lower courts.232 If this statute is not ambiguous, it is difficult to imagine exactly what kind of statute would be considered ambiguous.233

The other case, Mellouli v. Lynch, concerned whether a Kansas misdemeanor conviction for possession of drug paraphernalia counted as a ground of removal for a “controlled substance violation” (but not an aggravated felony).234 The BIA had previously published a decision which was on point.235 But the Board’s approach differed from the categorical approach that the Court had required in Moncrieffe.236Moncrieffe involved an alleged aggravated felony of drug trafficking, which in the immigration statute is more serious than simple possession.237 The Supreme Court said that this would mean that an immigrant might be deportable for a low level paraphernalia conviction, but not for higher level drug trafficking.238 The Court thus dispensed with the government’s request for deference by briefly stating that “[b]ecause it makes scant sense, the BIA’s interpretation, we hold, is owed no deference under the doctrine described in Chevron . . . .”239 On the surface, this appears to be an emphatic Step One decision, except that the reason why the BIA’s interpretation made “scant sense” was due to the Supreme Court’s own decision in Moncrieffe, where the Court insisted on the categorical approach.240

These two groups of cases are summarized below in Table 2. They illustrate the two types of soft anti-Chevron decisions, those where Chevron is entirely ignored in the decision, and those where it is mentioned but seems irrelevant to the result. The pattern here is quite strong, especially when compared with the immigration cases that did not involve criminal grounds of removal, which I discussed in Section III.D. Consistently, in case after case in this category, the Court does not defer to the Attorney General (or the BIA). It would always be possible to quibble about Chevron’s non-application in an individual instance. For instance, perhaps sometimes the statute really is clear. The “without remuneration” exception for marijuana distribution in Moncrieffe might be such an example.241 But if the Court cared about Chevron in these cases, it could simply categorize this is a Step One decision. In other cases the statutory language is decidedly ambiguous, if the concept of ambiguity is to have any coherency. To borrow Chief Justice Roberts’ famous statement that to be a judge is to simply “call balls and strikes,” 242 these are cases generally right down the center of the Chevron strike zone. And yet in those cases, Chevron does not impact the Courts’ ruling, and is typically not even mentioned by the Court.

Kagan Pic 2 

While in these cases the Court is consistent in not deferring to the agency, the government still often wins. This is consistent with an assertion of judicial supremacy on questions of law, but it does not necessarily favor a particular policy outcome (deportation or non-deportation) over another. The lack of a pattern as to the result, as contrasted with the clear pattern of non-deference, is relevant to a currently open question over whether criminal grounds of removal should be interpreted according to the rule of lenity. This is explored in more detail in Section IV.D. But for present purposes, the necessary next step is to describe the pattern so as to then identify potential doctrinal rules that might explain it.

The cases in Table 2 all involve criminal grounds of removal. The Sixth Circuit’s Judge Sutton has argued that Chevron can only apply to the interpretation of civil statutes, but not criminal statutes.266 Immigration is generally seen as falling on the civil side of the line, but the definition of aggravated felonies is a hybrid. In these cases, “the same statute has criminal and civil applications.”267 Aggravated felonies are defined in one part of the Immigration and Nationality Act.268

Another section refers to this definition as setting out a ground of removal in administrative proceedings.269 But other sections incorporate this definition as an element of an immigration-related crime, such as severely enhancing the maximum sentence for illegal re-entry.270 That means that when a court is asked to interpret the definition of an aggravated felony in a deportation case, it is also defining the elements of a crime. This may explain why Chevron deference is inappropriate, as well forming the basis for applying the rule of lenity, which I discuss in Section IV.D.

However, the dual-use statute theory does not explain the Supreme Court’s lack of deference in Mellouli, which did not involve a statute with a dual application in criminal law.271 Also, immigration is not the only field of administrative law that has dual use statutes. Judge Sutton’s colleagues on the Sixth Circuit rejected his view because the Supreme Court had previously found that deference applies to the Department of the Interior’s interpretation of the Endangered Species Act, even though that act has criminal as well as civil applications.272 Nevertheless, the majority of the Sixth Circuit panel recognized that the Supreme Court had appeared to move away from deference, but relied on the non-removal immigration case law discussed above in Section III.C, which suggested that the application of Chevron was still required.273

The dual-use statute theory relies heavily on the traditional distinction between civil and criminal areas of law. But the civil-criminal distinction was pointedly attacked by five Supreme Court justices in the 2018 decision in Sessions v. Dimaya.274 The key question in Dimaya was whether the same void for vagueness standard applied to deportation statutes in immigration cases as in criminal cases. The Court held that the standard is the same, albeit with a spit decision. Justice Kagan’s opinion, which attracted four total votes, argued that the higher standard was necessary because of the “grave nature of deportation.”275 Justice Gorsuch concurred, but resisted carving out an exception only for deportation cases. He argued that many civil cases carry serious consequences, and thus questioned whether the civil-criminal distinction would ever matter.276

If the dual-use statute theory is not fully satisfactory, how else can we explain the Court’s reluctance to apply Chevron deference? A slightly broader description of the pattern would suggest that in immigration cases the Supreme Court does not defer with regard to any grounds of removal based on criminal convictions, regardless of whether they are dual-use statutes. Descriptively, this captures Mellouli, Moncrieffe and other aggravated felony cases discussed in this Article. But while it describes the decisions that the Supreme Court has given us, it may have theoretical problems. Whether we take this slightly broader view, or Judge Sutton’s theory about dual use statutes, tremendous stress is placed on the formalistic civil-criminal distinction. The Court has said that “criminal laws are for the courts, not for the Government, to construe.”277 But even in the criminal context the Court has occasionally shown that there can be some room for the executive branch to autonomously proscribe conduct, at least when stringent safeguards are in place, including preserving a role for the courts.278 Rather than focus on a formalistic bright line that may not exist, it is important to understand not just the general rule, but also the reason why deference is generally inappropriate in criminal law.

Deferring to the executive branch on matters of criminal law would invite arbitrary use of draconian state power, by allowing an administrative agency to “create (and uncreate) new crimes at will.”279 This was at the heart of Judge Sutton’s concern regarding the intersection of Chevron and dual-use statutes:

[Applying deference] would leave this distasteful combination: The prosecutor would have the explicit (executive) power to enforce the criminal laws, an implied (legislative) power to fill policy gaps in ambiguous criminal statutes, and an implied (judicial) power to interpret ambiguous criminal laws. And it would permit this aggregation of power in the one area where its division matters most: the removal of citizens from society.280

The pivotal question seems to be whether it matters, for separation of powers purposes whether the government aims to remove a person to prison as a matter of criminal punishment, or to remove him from the country. The Supreme Court has long recognized that rigid application of the civil–criminal distinction is not always appropriate in immigration cases.281 Justice Gorsuch has argued that the separation of powers concerns that weigh against deference on criminal law apply with equal force to some immigration contexts (and, apparently for Gorsuch, in all administrative contexts).282 Some commentators have speculated that in avoiding deference in certain immigration cases, the Court may be acting on an unspoken inclination that “deportation[ ]is[ ]different.”283 This suggests that the real explanation for non-deference in criminal grounds of removal cases is not the criminal law issues but rather the removal. In order to understand why that might be the case, it is important to put deportation itself in a broader enforcement context.

E.     Physical Liberty and Separation of Powers

Up to this point I have argued that the Supreme Court’s practice in immigration cases follows a consistent pattern: If deportation is at stake, the Court either ignores Chevron deference entirely or fails to apply it in any meaningful way. In this section, I will argue why this pattern makes normative sense. Deportation means the government uses force to expel a person from the United States, to a place she does not want to go.284 Much like imprisonment, this kind of deprivation of physical liberty calls for strong checks and balances between the judiciary and the executive branches, which makes judicial deference to administrative interpretations of the law especially indefensible. As the Court has said in a different context, deprivation of physical liberty “is a penalty different in kind.”285

My normative assertion could be stated this way: If one branch of government infringes a person’s physical liberty (either by detention or deportation) she should have the right to go before a separate branch of government for an assessment of whether this action was justified under law. That is a basic check and balance, a feature of our constitutional separation of powers. Immigration enforcement distorts this separation, however. In immigration, people are arrested, confined behind bars, judged, and deported all by the executive branch. Chevron would mean that even in the limited judicial check that exists on the immense power that the federal government wields over the physical liberty of individuals, the judiciary should defer back to the executive branch on questions of law. This is too much power for one branch of government to have.

I trust that few would seriously contest that physical liberty is a sacrosanct constitutional value. For purposes of Chevron deference, the real problem is to connect physical liberty with the role of the judiciary. At a general level, it has been recognized since the early days of the Republic that courts play a special role in safeguarding liberty. In Ex Parte Bollman Chief Justice Marshall wrote:

Of a tribunal whose members, having attained almost all that the constitution of their country permits them to aspire to, are exempted, as far as the imperfection of our nature allows us to be exempted, from all those sinister influences that blind and swerve the judgments of men—have nothing to hope, and nothing to fear, except from their own consciences, the opinion of the public, and the awful judgment of posterity? It is in the hands of such a tribunal alone, that in times of faction or oppression, the liberty of the citizen can be safe.286 

In a similar vein, Alexander Hamilton quoted Montesquieu for the maxim that “there is no liberty, if the power of judging be not separated from the legislative and executive powers.”287 In a passage that reads today like a swipe at Chevron, Hamilton wrote:

[A]s liberty can have nothing to fear from the judiciary alone, but would have every thing to fear from its union with either of the other departments; that as all the effects of such a union must ensue from a dependence of the former on the latter, notwithstanding a nominal and apparent separation; that as, from the natural feebleness of the judiciary, it is in continual jeopardy of being overpowered, awed, or influenced by its co-ordinate branches. . . .288

The precise nexus between physical liberty and separation of powers can be seen in more recent illustrations of the point. In particular, useful guidance may be the Fourth Amendment, because this amendment concerns physical seizures before a person has been convicted in a criminal trial.289 The landmark case of Gerstein v. Pugh is especially informative because the Court there wrestled with what amounted to a separation of powers question.290 In this case, the State of Florida used a procedure by which a suspect could be detained after a warrantless arrest based solely on a prosecutor having filed charges, with no judicial review of probable cause.291 Florida argued that the prosecutor’s involvement was a sufficient safeguard, but the Court found this unconvincing, insisting on maintaining separation of powers through a “neutral and detached magistrate.”292

IV.     Beyond Deportation Cases

A.     Immigration Detention

Given the competing potential explanations for why deportation cases are treated differently, immigration detention cases offer a critical test of alternative theories about the limits of Chevron. In the Chevron era, the Supreme Court has heard fewer cases concerning detention of immigrants than cases related to the deportation of immigrants. However, these cases have the potential to provide a window of explanation into how we should interpret the reason for the Court’s apparent hesitation about Chevron deference. If the pivotal issue is physical liberty, then detention cases should be handled much like deportation cases, e.g., without deference.

As Alina Das has noted, the federal government has often succeeded in persuading lower courts to apply Chevron deference in habeas cases concerning the mandatory detention of immigrants.293 Not all circuit courts have been equally receptive to Chevron in immigration detention cases, though (much like the Supreme Court) they sometimes avoid Chevron by simply failing to mention it in their decisions rather than explaining why they chose not to apply it.294 As Das argues, the application of Chevron in habeas cases undermines the role the judiciary has traditionally played in reviewing deprivations of liberty:

Any habeas challenge to the scope of an immigration detention statute—whether it focuses primarily on constitutional concerns or involves broader tools of statutory construction—ultimately requires review of the lawfulness of the executive’s deprivation of an immigrant’s physical liberty. . . . In the context of immigration detention challenges, the interpretive choice is almost always between an agency view that would result in continuing detention and a countervailing interpretation that would result in the detainee’s freedom or more robust procedural protections. The application of Chevron deference in immigration detention cases thus operates as a presumption in favor of detention, at least in the absence of countervailing norms that would give weight to the physical liberty interest at stake.295

The Supreme Court’s rulings thus far follow the same pattern illustrated in deportation cases. In the landmark case of Zadvydas v. Davis, the Court wrestled with whether a statute authorized indefinite detention of deportable immigrants.296 The government asked for Chevron deference in favor its interpretation of the statute, which would have allowed indefinite detention.297 The Court acknowledged that the statute in question was ambiguous,298 but resolved the case in favor of the immigrants using the doctrine of constitutional avoidance.299Chevron appears nowhere in the decision.

In Martinez, the majority opinion by Justice Scalia similarly made no mention of Chevron deference.300 It was in this decision that Scalia famously said that a statute cannot be “a chameleon”301 because the government wanted to give the detention statute a different meaning than it had been given in Zadvydas. In dissent, Justice Thomas thought that Chevron should have applied.302

Zadvydas and Martinez both tackled the issue of detention after an order of removal, with the Court finding against indefinite detention in both cases. In Demore v. Kim, the Court affirmed temporary mandatory detention while a removal case was pending.303 The Court again made no mention to Chevron, but it is less clear if the government asked for such deference. In this instance, the government asked for deference based on its plenary power over immigration, not based on Chevron.304 

The pattern held in the 2018 decision in Jennings v. Rodriguez, which involved a challenge to long-term mandatory detention while cases are pending in Immigration Court. The Court’s decision focused extensively on statutory interpretation, and it declined to immediately resolve the central constitutional challenge to mandatory detention.305 In this case, the government made it clear that Chevron deference should apply.306Jennings, a case about statutory interpretation concerning administrative agencies, would appear to be a good candidate for Chevron to play some role. But, true to the historical pattern, neither the words Chevron nor defer appear anywhere in the decision.307

 B.     Other Liberties

If I am correct in arguing that a physical liberty exception to Chevron deference exists, a question will arise about whether there are other administrative law contexts where there is an interest at stake so weighty that a similar Step-Zero limitation to Chevron is called for. For those who believe the entire Chevron enterprise is flawed, this will be of little concern. But for those who believe Chevron has real value in many contexts, making an exception for physical liberty will be a real concern.308 Certainly, the Constitution protects the right to own property as well as physical liberty, and many regulatory policies impact property rights. There are also strong critiques of Chevron that focus on broader conceptions of liberty, rather than the narrower focus on physical liberty presented in this Article.309

In the abstract, it is certainly possible that a deportation or immigration detention case could be the occasion for the Court to kill off Chevron entirely. Moreover, I would not attempt here to contest Judge Gorsuch’s broad assertion that administrative agencies generally have the capacity “to penalize persons in ways that can destroy their livelihoods and intrude on their liberty even when exercising only purely civil powers.”310 This broader liberty argument would call for a wholesale dismantling of deference. I have not attempted in this Article to wrestle with this larger attack on Chevron, except to observe how it makes the doctrine’s future reach less certain. My purpose has been to develop a narrower and thus more modest argument focused on physical liberty only, rather than all forms of liberty.

As explained already in Part II, while Chevron is on shakier ground today than it once was, it is still ambitious to suggest that it is on the verge of being overturned, rather than merely having its wings clipped. We already know from King v. Burwell that Chevron does not apply in certain “major questions” involving matters of great social and economic consequence.311 The Court has not clarified the parameters of this exception, but the existence of the exception shows that the Court thinks there are some matters too weighty for deference to be appropriate. The physical liberty exception is similar.

There is no need to up-end all of Chevron in order to recognize that a government intrusion on physical liberty entails more protections than purely monetary intrusions. This is a line that the Supreme Court has drawn in other contexts, especially in criminal law, which is the field of law most attuned to governing the power of the government to seize and detain people.312 Moreover, the fact that the Court has given consistent force to Chevron deference in immigration cases that do not involve deportation or detention should be seen as an affirmation of the doctrine’s vitality in contexts for which it is appropriate.

C.     Relief from Removal

A removal proceeding in Immigration Court proceeds in two stages. The Immigration Judge first must find that the person is removable.313 Usually this could be shown if the person is present in violation of the law314 (i.e., an undocumented immigrant), or if a legal resident is convicted of a certain type of criminal offense, such as an aggravated felony or a crime relating to a controlled substance.315 The deportation cases that I have discussed in Section III.C where the Court failed to apply Chevron arose when these grounds of removal were contested. But that is not the end of a typical case in Immigration Court. In fact, it is often only the beginning. In fact, in many cases removability is conceded immediately.

Once removability is established, the proceeding moves to a second stage at which the non-citizen may ask for relief from removal.316 Potential relief takes many forms, but it includes protection from persecution abroad through asylum317 and withholding of removal,318 as well as discretionary cancellation of removal for long-time residents who have not been convicted of an aggravated felony.319 Often, there is no dispute over the non-citizen’s threshold removability, but eligibility for asylum is hotly contested. A great deal of the appellate immigration litigation at the circuit courts concerns eligibility for relief from removal, and Chevron deference has often played a central role.320

On the question of whether Chevron should matter with regard to claims of relief, I can offer two different answers. The first answer is descriptive—what is the Court actually doing? The second answer is prescriptive—what should the Court do?

In cases concerning relief from removal the Supreme Court has been more willing to apply Chevron, dating back to its 1987 decision in Cardoza-Fonseca, and followed by Aguirre-Aguirre (1999), Negusie (2009), and Martinez Gutierrez (2012).321 There are plausible arguments that may explain the Court’s apparent inclination that Chevron should apply to relief from removal, but not to grounds of removability. Grounds of removal constitute the legal justification for forcibly expelling someone from the country. They are the legal regulation of the government’s power over the individual. Since these statutes are the direct justification for violating individual liberty, their application and interpretation require special judicial attention. By contrast, relief from removal is more akin to eligibility for an immigration benefit or seeking an admission. This distinction may make judges more comfortable with permitting some range of executive discretion. The Court has long held that a person making such an application is not entitled to due process.322 The justices may perceive that deferring to congressional delegation is justified.

Before 2018, it would have been correct to say that the Court consistently applies meaningful Chevron deference in relief from removal cases. But this must be tempered by the Court’s decision in Pereira v. Sessions.323 This case concerned eligibility for cancellation of removal, which is a form of relief adjudicated in immigration court.324 The lower court had found the statute ambiguous,325 and the circuits had split how the statute should be understood.326 Nevertheless, writing for an 8-1 majority, Justice Sotomayor found the statute “clear and unambiguous” and thus “the Court need not resort to Chevron deference, as some lower courts have done.”327 The majority issued a soft anti-Chevron decision, in that it renders Chevron irrelevant even in a case where many other judges found a statute unclear. This is notable because this was a relief from removal case, the kind of case in which the Court had previously appeared comfortable applying meaningful deference. It also drew a pointed rebuke from Justice Alito:

Here, a straightforward application of Chevron requires us to accept the Government’s construction of the provision at issue. But the Court rejects the Government’s interpretation in favor of one that it regards as the best reading of the statute. I can only conclude that the Court, for whatever reason, is simply ignoring Chevron.328

Justice Alito is right; the Court was effectively ignoring Chevron. But that just begs the questions: Was this a case where Chevron should have applied?

I am not convinced that using deference in relief from removal cases is more justified than in grounds of removal cases. When there is a dispute about relief from removal, what is ultimately at stake is the same as in criminal grounds of removal cases: will the person be deported? Moreover, in asylum cases, the non-citizen is arguing that she will be subject to persecution if deported. Unlike an applicant for a visa (the topic at issue in Cuellar de Osorio) a person who is seeking asylum in a removal hearing is inside the United States and is entitled to due process.329 The matters under review in court are legal definitions, not matters of discretion.330 Congress explicitly preserved judicial review on questions of law in these cases, which undermines the argument that deference would respect an implicit congressional delegation to the executive branch.331 Thus, I argue that the Supreme Court should treat the interpretation of asylum eligibility as it does grounds of removability. It is worth noting that then-Judge Gorsuch issued his broadside against Chevron in an immigration case that was not about a ground of deportation, but instead involved eligibility for an immigration benefit.332 This departs from the pattern observable in Supreme Court cases, but I would argue that Gorsuch got it right.

Nevertheless, there is a consistent pattern in the way the Supreme Court has applied or not applied Chevron in these cases, which suggests that the justices perceive claims for relief to be better candidates for deference than grounds of removal. While my main argument in this Article aims to describe, explain and defend the Court’s practice—on relief from removal I am making the more ambitious claim that the Court should change its approach.

D.     The Lenity Question

Immigrants have often argued that grounds of deportation based on criminal convictions must be interpreted narrowly (and thus in their favor) according to the rule of lenity, an issue that has also emerged in some tax law contexts.333 Lenity requires that an ambiguous criminal statute be interpreted in the manner most favorable to the defendant.334 Justice Scalia, as well as the Sixth Circuit’s Judge Sutton, have argued that lenity should be invoked for dual use statutes, since they have a criminal law application.335 Several scholars have also weighed in, favoring lenity when deportation is premised on a criminal conviction.336

The Supreme Court itself appeared to endorse the argument for lenity with regard to immigration aggravated felonies in its 2004 decision in Leocal v. Ashcroft, albeit only in a footnote.337 However, the Court (in an opinion by Justice Scalia) cited this footnote approvingly in a case concerning immigration detention,338 and has said in at least one other context that ambiguities over grounds of removal should be construed in favor of the immigrant.339 The Court has also indicated support for lenity in a case involving dual use statutes outside the immigration context.340 The Court has more recently avoided the question.341

In the criminal context, the rule of lenity ensures that people have “fair warning of the boundaries of criminal conduct” while also reinforcing the primacy of courts in interpreting the law.342 In Judge Sutton’s words, “[w]hen a single statute has twin applications, the search for the least common denominator leads to the least liberty-infringing interpretation.”343 There are different ways to conceive of how lenity, if applicable, would interact with Chevron deference when an agency seeks to impose, in a civil context, a statute that also has a criminal application. One approach would suggest that deference does not apply to dual use statutes.344 This might be thought of as a Step Zero formulation, because it articulated a limitation on when Chevron should apply. Another option would be to apply Chevron’s analytical framework, but to then hold that the rule of lenity applies to Chevron’s second step—the question of whether the BIA’s interpretation is reasonable.345 This seems consistent with the manner in which the Court has tended to first attempt to resolve statutory ambiguity, and only if that fails, invoke lenity.346 On the other hand, it is difficult to see how either of these formulations would lead to different results. Either way, lenity’s demand for a narrow interpretation of the statute would trump any contrary interpretation from the agency. Some have argued that lenity does not really fit in either step, and may not be appropriate in non-criminal cases.347

The rule of lenity is usually cited in reference to criminal grounds of removal. However, Das argues for something quite similar in reference to statutes authorizing the detention of immigrants.348 Based on the normative importance of liberty, she argues for “a presumption in favor of physical liberty” when interpreting statutes.349 She notes that, much like lenity, this could lead either to not applying Chevron or to constraining the agency’s actions when a statute is considered ambiguous under step one.350 These are compelling arguments, but they are different than the thesis of this Article because they dictate how to resolve substantive interpretive questions of law, whereas the question with Chevron is a matter of separation of powers, e.g., who should be the primary decision-maker.

Despite the Court’s past statements that appear favorable to the use of lenity in immigration cases, the Court has recently avoided re-affirming its application when presented with two opportunities to do so. In Torres, the BIA decided that jurisdictional elements of the aggravated felony definition should not count for the categorical approach.351 Mr. Torres argued that the Court should apply lenity rather than Chevron because deportation is especially grave, and because the aggravated felony definition has a criminal application.352 As we have seen in Section III.D, the government asked for deference, and the Court decided the case in favor of the government—but, did not mention Chevron deference, nor lenity.353 In 2017, the Court decided Esquivel-Quintana, in which the parties raised essentially the same arguments about lenity and Chevron. This was the case, when it was in the lower courts, in which Judge Sutton had issued his widely-cited opinion opposing deference and calling for lenity. The Court this time acknowledged the issues, but dispensed with them in a single sentence, holding that the statutory provision was not ambiguous, so there was no need to resolve the Chevron-lenity dispute.354

The lenity question is unresolved and likely to re-emerge. I do not seek to resolve this question in this Article, and it is not necessary in order to make the case that Chevron does not (and recently, has not) applied in cases involving physical liberty. While lenity is often proposed as an alternative to Chevron, the Court could choose to reject both. As we have already seen in Section III.B, the Court in removal cases has consistently avoided deference, but it has not consistently resolved the statutory provisions in favor of the immigrant as the rule of lenity would likely require. Even in the criminal context, the Court has long noted that lenity has limits, and only applies when it cannot otherwise resolve ambiguity in a statute.355 It stands to reason that if the Court is growing more confident in its ability to find meaning in superficially ambiguous statutes, the rule of lenity will ebb in importance just as Chevron deference has declined.

Limiting Chevron deference re-asserts the judicial role in interpreting statutes, but it does not provide instructions on how the judiciary should approach a statutory interpretation. One could coherently argue that the severity of deportation calls for narrow interpretations. Or one could argue that lenity should apply in dual use statutes only which would keep the rule of lenity tethered to the criminal law arena. While these are coherent theoretical arguments, it is harder to make the case that this is what the Supreme Court has been doing. Rather, the clear pattern is that the Court has not applied Chevron, and instead has re-asserted the judicial role in immigration cases that involve grounds for deportation and detention. But that does not mean the Court will always accept the statutory interpretation that is most favorable to the immigrant. The government can still win these cases—it just is not receiving deference as of late.

The rule of lenity is clearly appealing for an immigrant fighting deportation, or for anyone fighting the government over a statutory interpretation. If lenity is indeed an application of due process that constrains the power of the federal government, then there is indeed a compelling constitutional case for it. But some important doubts have been raised about whether this constitutional argument is somewhat overstated.356 As Jill E. Family explains in a recent article, the intersection of Chevron and lenity is opaque.357 As she explains, lenity is really about how to interpret a statute, while Chevron is about who should do the interpretation.358

The claim to lenity skirts the central problem that courts face when reviewing deportation cases, which is the deportability of the non-citizen. Instead of focusing on the stakes of immigration enforcement, lenity asks courts considering a deportation case to instead focus on the fact that the statutes are also used in a criminal context in other cases. Applied rigidly and in isolation, lenity could lead to a situation in which judges defer to the executive branch on detention questions, and as a default on most grounds of deportation, but then apply a highly pro-immigrant canon of interpretation if the ground of deportation happens to be a dual use statute. That would not seem to be a coherent or satisfactory approach. It also avoids the possibility that the Court is forging a somewhat hybrid approach in removal cases, treating these cases as neither purely criminal nor purely administrative.359 For these reasons, the Court should focus on what is at stake in deportation and detention cases, and to focus on the role the judiciary should play given those stakes.

V.     Conclusion

Deportation truly is different, at least from more typical issues that arise in administrative law. But it is fundamentally similar to a question that has long been a concern in Anglo-American law, and especially in constitutional law: How should law regulate government intrusions on physical liberty? Maintaining separation of powers, especially a fully independent review by the judiciary, is a well-established part of the answer to that question. Chevron deference in the context of immigration enforcement undermines this safeguard. Cases involving physical liberty are different, giving the judiciary a unique and sacred role in a democracy based on checks and balances.

Unlike arguments for lenity, avoiding Chevron deference in deportation cases “would not lead to any guaranteed results either pro-immigrant or anti-immigrant,” as Jill E. Family writes.360 Moreover, as I have tried to show here, it would not amount to a broadside attack on Chevron in all cases. Even in the supposedly exceptional realm of immigration, there are many situations in which the Supreme Court has applied Chevron with full force, precisely as administrative law textbooks would anticipate. Rather than see immigration cases as unique, courts should approach them with a focus on what is at stake for the people involved.

If Chevron applies in cases of deportation and detention, a single branch of government would be able to both execute the law against individuals, and at the same time issue authoritative interpretations of the law under which physical liberty is to be violated. The application of Chevron deference would be wrong in this context, and while the Supreme Court has never articulated this as a rule, it has been correct in practice to avoid applying Chevron in these cases. It is time for the Court to state the rule.


  1. [1]. See, e.g., Peter J. Henning, Gorsuch Nomination Puts Spotlight on Agency Powers, N.Y. Times (Feb. 6, 2017), https://www.nytimes.com/2017/02/06/business/dealbook/gorsuch-nomination-puts-spotlight-on-agency-powers.html; Steven Davidoff Solomon, Should Agencies Decide Law? Doctrine May Be Tested at Gorsuch Hearing, N.Y. Times (Mar. 14, 2017), https://www.nytimes.com/
    2017/03/14/business/dealbook/neil-gorsuch-chevron-deference.html.

  2. [2]. Gutierrez-Brizuela v. Lynch, 834 F.3d 1142, 1155 (10th Cir. 2016) (Gorsuch, J., concurring).

  3. [3]. Philip Rucker, Bannon: Trump Administration is in Unending Battle for ‘Deconstruction of
    the Administrative State
    , Wash. Post (Feb. 23, 2017), https://www.washingtonpost.com/news/
    powerpost/wp/2017/02/23/bannon-trump-administration-is-in-unending-battle-for-deconstruction-of-the-administrative-state; see also Matt Ford, Judge Gorsuch Goes to Washington, Atlantic (Mar. 20, 2017), https://www.theatlantic.com/politics/archive/2017/03/judge-gorsuch-goes-to-washington/
    520230; Henry Gass, Gorsuch Hearings: Should Agencies—or Courts—Decide the Law?, Christian Sci. Monitor (Mar. 22, 2017), https://www.csmonitor.com/USA/Justice/2017/0322/Gorsuch-hearings-Should-agencies-or-courts-decide-the-law.

  4. [4]. See, e.g., Ian Mason, Neil Gorsuch is Ready to Take on Administrative State, Breitbart (Nov. 17, 2017), https://www.breitbart.com/big-government/2017/11/17/gorsuch-ready-administrative-state.

  5. [5]. Maria Cantwell, Cantwell Statement on Judge Neil Gorsuch’s Nomination to U.S. Supreme Court, Maria Cantwell: U.S. Senator Wash. (Mar. 30, 2017), https://www.cantwell.senate.gov/news/
    press-releases/cantwell-statement-on-judge-neil-gorsuchs-nomination-to-us-supreme-court.

  6. [6]. See discussion infra Part III.

  7. [7]. See Jill E. Family, Immigration Law Allies and Administrative Law Adversaries, 32 Geo. Immigr. L.J. 99, 122–26 (2017); Sarah Madigan, Revisiting Deference to Agencies in Criminal Deportation Cases, Reg. Rev. (Nov. 16, 2017), https://www.theregreview.org/2017/11/16/madigan-revisiting-deference-criminal-deportation (summarizing arguments for immigration-specific exceptions to Chevron). See generally Gabriel J. Chin et al., Chevron and Citizenship, 52 U.C. Davis L. Rev. 145 (2018) (questioning the application of Chevron in citizenship adjudication).

  8. [8]. See, e.g., Attorney General, Memorandum for the Executive Office for Immigration Review: Renewing Our Commitment to the Timely and Efficient Adjudication of Immigration Cases to Serve the National Interest 1–2 (Dec. 5, 2017) (calling on Immigration Judges to work toward “an end to unlawfulness in our immigration system” and stating that the manner in which immigration cases are adjudicated directly impacts sovereign interests).

  9. [9]. See, e.g., Attorney General, Remarks to the Executive Office for Immigration Review (Oct. 12, 2017) (criticizing liberal interpretations of asylum law and “dirty immigration lawyers”); Jeff Sessions, Attorney General Jeff Sessions Delivers Remarks on Violent Crime to Federal, State and Local Law Enforcement, U.S. Dep’t of Justice (Apr. 28, 2017), https://www.justice.gov/opa/
    speech/attorney-general-jeff-sessions-delivers-remarks-violent-crime-federal-state-and-local-law; Jeff Sessions, Attorney General Jeff Sessions Delivers Remarks Before Media Availability in El Paso, Texas, U.S. Dep’t of Justice (Apr. 20, 2017), https://www.justice.gov/opa/speech/attorney-general-jeff-sessions-delivers-remarks-media-availability-el-paso-texas; Jeff Sessions, Attorney General Jeff Sessions Announces The Department of Justice’s Renewed Commitment to Criminal Immigration Enforcement, U.S. Dep’t of Justice (Apr. 11, 2017), https://www.justice.gov/opa/pr/attorney-general-jeff-sessions-announces-department-justice-s-renewed-commitment-criminal.

  10. [10]. 8 U.S.C. § 1252 (2012).

  11. [11]. Id. § 1252(a)(2)(D).

  12. [12]. See discussion infra Section III.C.

  13. [13]. See generally Kent Barnett & Christopher J. Walker, Chevron in the Circuit Courts,
    116 Mich. L. Rev. 1 (2017) (discussing how the Chevron deference doctrine is applied in administrative law cases).

  14. [14]. See generally David S. Rubenstein & Pratheepan Gulasekaram, Immigration Exceptionalism, 111 Nw. U. L. Rev. 583 (2017) (discussing the exceptional treatment the Supreme Court tends to give immigration law).

  15. [15]. See generally Patrick Glen, Response to Walker on Chevron Deference and Mellouli v. Lynch, Yale J. on Reg.: Notice & Comment (June 20, 2015), http://yalejreg.com/nc/response-to-walker-on-chevron-deference-and-mellouli-v-lynch-by-patrick-glen (discussing the possibility
    of a “deportation-is-different” explanation for the Court’s reluctance with regard to Chevron); Michael Kagan, Chevron’s Immigration Exception, Revisited, Yale J. on Reg.: Notice & Comment (June 10, 2016), http://yalejreg.com/nc/chevron-s-immigration-exception-revisited-by-michael-kagan (discussing the role of Chevron in immigration cases and advocating for a “deportation-is-different” theory); Chris Walker, The “Scant Sense” Exception to Chevron Deference in Mellouli v. Lynch, Yale J. on Reg.: Notice & Comment (June 2, 2015), http://yalejreg.com/nc/the-scant-sense-exception-to-chevron-deference-in-mellouli-v-lynch-by-chris-walker (discussing the possibility that the Roberts Court may be reluctant to give deference in certain deportation cases).

  16. [16]. See discussion infra Section IV.A.

  17. [17]. This appears to be the case with Bureau of Prisons sentencing cases. See Barber v. Thomas, 560 U.S. 474, 502 (2010) (“There is no indication that BOP has exercised the sort of interpretive authority that would merit deference underChevron.”); see also Lopez v. Terrell,
    654 F.3d 176, 182 (2d Cir. 2011) (“[W]here the regulation identified by the [BOP] does not speak to the statutory ambiguity at issue, Chevron deference is inappropriate.”).

  18. [18]. See discussion infra Section IV.D; infra text accompanying notes 91–92, 268–74.

  19. [19]. Barnett & Walker, supra note 13, at 2.

  20. [20]. Pereira v. Sessions, 138 S. Ct. 2105, 2121 (2018) (Kennedy, J., concurring).

  21. [21]. Chevron, U.S.A., Inc. v. Nat. Res. Def. Council, Inc., 467 U.S. 837, 842 (1984) (“When a court reviews an agency’s construction of the statute which it administers, it is confronted with two questions.”).

  22. [22]. Id.

  23. [23]. Id. at 843.

  24. [24]. Id. at 844.

  25. [25]. See Ian Bartrum, The Constitutional Canon as Argumentative Metonymy, 18 Wm. & Mary Bill Rts. J. 327, 329 (2009) (“[A] canonical text takes on its own metonymic meanings—sometimes quite apart from its literal textual meaning—within the practice of constitutional law.”).

  26. [26]. See Gary Lawson & Stephen Kam, Making Law Out of Nothing at All: The Origins of the Chevron Doctrine, 65 Admin. L. Rev. 1, 63 (2013).

  27. [27]. Id. at 39–41 (discussing General Motors Corp. v. Ruckelshaus, 742 F.2d 1561 (D.C. Cir. 1984)).

  28. [28]. See id. at 6.

  29. [29]. See Connor N. Raso & William N. Eskridge, Jr., Chevron as a Canon, Not a Precedent:
    An Empirical Study of What Motivates Justices in Agency Deference Cases
    , 110 Colum. L. Rev. 1727,
    1732–33 (2010) (“[A] simple ideological account does not explain why arch-conservative Justice Scalia followed Chevron to allow the agency discretion to adopt a very liberal rule for the ADEA. But it is possible that his willingness to go along with a liberal result in this case might be part of a larger conservative strategy in the general run of cases.”).

  30. [30]. Skidmore v. Swift & Co., 323 U.S. 134, 140 (1944).

  31. [31]. See generally Raso & Eskridge, supra note 29 (finding that justices apply Chevron differently in different contexts).

  32. [32]. See Barnett & Walker, supra note 13, at 12; William N. Eskridge, Jr. & Lauren E. Baer, The Continuum of Deference: Supreme Court Treatment of Agency Statutory Interpretations from Chevron to Hamdan, 96 Geo. L.J. 1083, 1124–25 (2008); Michael Herz, Chevron is Dead; Long Live Chevron, 115 Colum. L. Rev. 1867, 1870 (2015). But see Natalie Salmanowitz & Holger Spamann, Does the Supreme Court Really Not Apply Chevron When It Should? (September 2, 2018) (working paper), available at https://ssrn.com/abstract=3243095 (presenting new empirical research findings indicating that the Court may invoke Chevron more consistently than previous research indicated).

  33. [33]. See Raso & Eskridge, supra note 29, at 1766 (“Chevron and the other formal deference regimes have the following characteristics in practice: They are flexible rules of thumb or presumptions deployed by the Justices episodically and not entirely predictably, rather than binding rules that the Justices apply more systematically.”).

  34. [34]. Peter L. Strauss, “Deference” Is Too Confusing—Let’s Call Them “Chevron Space” and “ Skidmore Weight, 112 Colum. L. Rev. 1143, 1144–45 (2012).

  35. [35]. Herz, supra note 32, at 1867.

  36. [36]. Michigan v. EPA, 135 S. Ct. 2699, 2712 (2015) (Thomas, J., concurring).

  37. [37]. See Bowles v. Seminole Rock & Sand Co., 325 U.S. 410, 414 (1945).

  38. [38]. See Perez v. Mortg. Bankers Ass’n, 135 S. Ct. 1199, 1210–13 (2015) (Scalia, J., concurring) (arguing that even if Chevron remains, Auer deference should be reversed); id. at 1231 (Thomas, J. concurring) (arguing the Seminole Rock/Auer “line of precedents undermines our obligation to provide a judicial check on the other branches”); id. at 1210 (Alito, J., concurring) (“The opinions of Justice SCALIA and Justice THOMAS offer substantial reasons why theSeminole Rockdoctrine may be incorrect.”).

  39. [39]. See Aaron L. Nielson, D.C. Circuit Review—Reviewed: A New Step for Chevron?, 36 Yale J. on Reg.: Notice & Comment (June 16, 2017), http://yalejreg.com/nc/d-c-circuit-review-reviewed-a-new-step-for-chevron; cf. Auer v. Robbins, 519 U.S. 452, 457–58 (1997); Adam J. White, Scalia and Chevron: Not Drawing Lines, But Resolving Tensions, 36 Yale J. on Reg: Notice & Comment (Feb. 23, 2016), http://yalejreg.com/nc/scalia-and-chevron-not-drawing-lines-but-resolving-tensions-by-adam-j-white.

  40. [40]. Gutierrez-Brizuela v. Lynch, 834 F.3d 1142, 1152 (Gorsuch, J., concurring) (“Chevronseems no less than a judge-made doctrine for the abdication of the judicial duty.”).

  41. [41]. See generally Christopher J. Walker, Toward a Context-Specific Chevron Deference, 81 Mo. L. Rev. 1095 (2016) (discussing the possibility that the Court will embrace Chief Justice Roberts’ context-specific Chevron approach following Justice Scalia’s death).

  42. [42]. Id. at 1103–05; see City of Arlington v. FCC, 569 U.S. 290, 318 (2013) (Roberts, C.J., dissenting).

  43. [43]. City of Arlington, 569 U.S. at 313 (quoting Free Enterprise Fund v. Pub. Co. Accounting Oversight Bd., 561 U.S. 477, 499 (2010)).

  44. [44]. Walker, supra note 41, at 1103.

  45. [45]. See Nicholas R. Bednar & Kristin E. Hickman, Chevron’s Inevitability, 85 Geo. Wash. L. Rev. 1392, 1439–41 (2017); Note, The Rise of Purposivism and the Fall of Chevron: Major Statutory Cases in the Supreme Court, 130 Harv. L. Rev. 1227, 1227–30 (2017) [hereinafter The Rise of Purposivism].

  46. [46]. See discussion infra Section II.C.

  47. [47]. Cf. SAS Inst. Inc. v. Iancu, 138 S. Ct. 1348, 1358 (2018) (“[W]hether Chevron should remain is a question we may leave for another day.”).

  48. [48]. Pereira v. Sessions, 138 S. Ct. 2105, 2129 (2018) (Alito, J., dissenting) (citations omitted).

  49. [49]. See Bednar & Hickman, supra note 45, at 1446–49.

  50. [50]. See Barnett & Walker, supra note 13, at 4–6.

  51. [51]. Id. at 6.

  52. [52]. Antonin Scalia, Judicial Deference to Administrative Interpretations of Law, 1989 Duke L.J. 511, 512–17.

  53. [53]. See United States v. Mead Corp., 533 U.S. 218, 226–27, 229 (2001).

  54. [54]. Scalia, supra note 52, at 513 (alteration in original) (quoting Marbury v. Madison,
    5 U.S. 137, 177 (1803)).

  55. [55]. See Thomas W. Merrill, Article III, Agency Adjudication, and the Origins of the Appellate Review Model of Administrative Law, 111 Colum. L. Rev. 939, 950–51 (2011).

  56. [56]. Id.

  57. [57]. Compare K Mart Corp. v. Cartier, Inc., 486 U.S. 281, 322 (1988) (Scalia, J., concurring in part and dissenting in part) (“[O]ne of the most important reasons we defer to an agency’s construction of a statute [is] its expert knowledge of the interpretation’s practical consequences.”), with Scalia, supra note 52, at 521 (criticizing technical expertise as insufficient to justify judicial deference).

  58. [58]. Lawson & Kam, supra note 26, at 11.

  59. [59]. Scalia, supra note 52, at 514.

  60. [60]. Chevron U.S.A., Inc. v. Nat. Res. Def. Council, Inc., 467 U.S. 837, 865–66 (1984).

  61. [61]. See Elena Kagan, Presidential Administration, 114 Harv. L. Rev. 2245, 2326–31, 2369 (2001) (noting that presidents should use the power of regulatory agencies to achieve policy goals because they can be subject to political accountability through elections).

  62. [62]. See generally Kent Barnett et al., Administrative Law’s Political Dynamics, 71 Vand. L.
    Rev. 1463 (2018) (describing administrative law’s political dynamics and effect on partisanship in judicial decision making).

  63. [63]. Glob. Tel*Link v. FCC, 866 F.3d 397, 408 (D.C. Cir. 2017); see also Aaron Nielson, D.C. Circuit Review – Reviewed: A New Step for Chevron?, Yale J. on Reg.: Notice & Comment (June 16, 2017), http://yalejreg.com/nc/d-c-circuit-review-reviewed-a-new-step-for-chevron.

  64. [64]. United States v. Mead Corp., 533 U.S. 218, 226–27, 229 (2001) (“We hold that administrative implementation of a particular statutory provision qualifies forChevrondeference when it appears that Congress delegated authority to the agency generally to makerules carrying the force of law ... . Congress ... may not have expressly delegated authority or responsibility to implement a particular provision or fill a particular gap.”); see also Thomas W. Merrill & Kristin E. Hickman, Chevron’s Domain, 89 Geo. L.J. 833, 863–64 (2001) (noting that congressional intent theory is the primary foundation for Chevron).

  65. [65]. See Merrill & Hickman, supra note 64, at 873, 912; Cass R. Sunstein, Chevron Step Zero, 92 Va. L. Rev. 187, 191 (2006).

  66. [66]. See Evan J. Criddle, Fiduciary Foundations of Administrative Law, 54 UCLA L. Rev. 117, 146 (2006).

  67. [67]. See Edward H. Stiglitz, Delegating for Trust, 166 U. Pa. L. Rev. 633, 637 (2018).

  68. [68]. See Scalia, supra note 52, at 515.

  69. [69]. Id.

  70. [70]. Id.

  71. [71]. See Barnett & Walker, supra note 13, at 33–34.

  72. [72]. See Strauss, supra note 34, at 1145.

  73. [73]. Id. at 1144.

  74. [74]. See The Rise of Purposivism, supra note 45, at 1227.

  75. [75]. Id.

  76. [76]. See City of Arlington v. FCC, 569 U.S. 290, 312–17 (Roberts, C.J., dissenting).

  77. [77]. Id. at 309 (Breyer, J., concurring).

  78. [78]. Id.

  79. [79]. Walker, supra note 41, at 1112.

  80. [80]. See David S. Rubenstein, “Relative Checks”: Towards Optimal Control of Administrative Power, 51 Wm. & Mary L. Rev. 2169, 2219–21 (2010).

  81. [81]. Nat’l Cable & Telecomms. Ass’n v. Brand X Internet Servs., 545 U.S. 967, 974 (2005).

  82. [82]. Auer v. Robbins, 519 U.S. 452, 458 (1997).

  83. [83]. See Adrian Vermeule, Bureaucracy and Distrust: Landis, Jaffe, and Kagan on the Administrative State, 130 Harv. L. Rev. 2463, 2488 (2017) (arguing that there are plural, conflicting justifications for the administrative state which cannot be fully reconciled with each other).

  84. [84]. United States v. Booker, 543 U.S. 220, 226–27 (2005).

  85. [85]. Id. at 239.

  86. [86]. Id. at 237.

  87. [87]. See F. Andrew Hessick & Carissa Byrne Hessick, The Non-Redelegation Doctrine,
    55 Wm. & Mary L. Rev. 163, 172–90 (2013).

  88. [88]. Gutierrez-Brizuela v. Lynch, 834 F.3d 1142, 1149–58 (10th Cir. 2016) (Gorsuch, J., concurring).

  89. [89]. Id. at 1152.

  90. [90]. Id.

  91. [91]. Id. at 1151; see also Aditya Bamzai, The Origins of Judicial Deference to Executive Interpretation, 126 Yale L.J. 908, 985, 1000 (2017) (detailing that “section 706 of the APA that a ‘reviewing court shall decide all relevant questions of law’ and ‘interpret constitutional and statutory provisions’” and “[Chevron] cannot be squared with the text of section 706 of the APA”); John F. Duffy, Administrative Common Law in Judicial Review, 77 Tex. L. Rev. 113, 193–99 (1998) (explaining that the Administrative Procedure Act and Chevron are at odds).

  92. [92]. Gutierrez-Brizuela, 834 F.3d at 1154 (Gorsuch, J., concurring) (citing Michigan v. EPA, 135 S. Ct. 2699, 2712 (2015) (Thomas, J., concurring)).

  93. [93]. Id. at 1154–55.

  94. [94]. Id. at 1156–57.

  95. [95]. Id. at 1156.

  96. [96]. Id. at 1155.

  97. [97]. See Raso & Eskridge, supra note 29, at 1735 (“The better path for reform is to simplify the deference regimes and tie them more tightly to their policy rationales, in the manner that the Court has done for substantive canons of statutory construction.”).

  98. [98]. Cf. id. (“We argue that scholars are being unrealistic when they demand that the Supreme Court adopt and consistently apply formal deference regimes that will ‘constrain’ the Justices in future cases. The Justices will not follow such regimes—and sooner or later lower court judges will not either.”).

  99. [99]. Nat’l Cable & Telecomms. Ass’n v. Brand X Internet Servs., 545 U.S. 967, 980 (2005).

  100. [100]. Michigan v. E.P.A., 135 S. Ct. 2699, 2712 (2015) (Thomas, J., concurring) (citations omitted).

  101. [101]. Barnett & Walker, supra note 13, at 4 (citing Herz, supra note 32).

  102. [102]. I have explained the logic underlying this methodology elsewhere. See generally Michael Kagan, Loud and Soft Anti-Chevron Decisions, 53 Wake Forest L. Rev. 37 (2018) (explaining the difference between loud and soft Chevron).

  103. [103]. See Walker, supra note 41, at 1096.

  104. [104]. See supra text accompanying note 60.

  105. [105]. See Kent Barnett, Why Bias Challenges to Administrative Adjudication Should Succeed,
    81 Mo. L. Rev. 1023, 1035–39 (2017) (noting Court’s ignoring of Chevron when it would have appeared to apply in Texas Department of Housing & Community Affairs v. Inclusive Communities Project, Inc., 135 S. Ct. 2507 (2015)).

  106. [106]. Id.

  107. [107]. Pereira v. Sessions, 138 S. Ct. 2105, 2121–29 (2018) (Alito, J., dissenting).

  108. [108]. Esquivel–Quintana v. Sessions, 137 S. Ct. 1562, 1567 (2017).

  109. [109]. See infra Section III.D.

  110. [110]. Esquivel–Quintana, 137 S. Ct. at 1572.

  111. [111]. See The Rise of Purposivism, supra note 45, at 1243–45.

  112. [112]. Cf. Raymond M. Kethledge, Ambiguities and Agency Cases: Reflections After (Almost) Ten Years on the Bench, 70 Vand. L. Rev. En Banc 315, 319–20 (2017) (arguing that statutes, if interpreted diligently, are not ambiguous). “That a statute is complicated does not mean it is ambiguous. It just means that the judge needs to work harder to determine—in the sense of ascertain—the statute’s meaning . . . . In my own opinions as a judge, I have never yet had occasion to find a statute ambiguous. In my view, statutory ambiguities are less like dandelions on an unmowed lawn than they are like manufacturing defects in a modern automobile: they happen, but they are pretty rare, given the number of parts involved.” Id.

  113. [113]. See Thomas W. Merrill, Judicial Deference to Executive Precedent, 101 Yale L.J. 969, 970 (1992) (“[T]he Chevron framework is used in only about half the cases that the Court perceives as presenting a deference question.”).

  114. [114]. Id.

  115. [115]. Eskridge & Baer, supra note 32, at 1091.

  116. [116]. See Barnett & Walker, supra note 13, at 72–73.

  117. [117]. Herz, supra note 32, at 1870.

  118. [118]. Cf. Michael Coenen & Seth Davis, Minor Courts, Major Questions, 70 Vand. L. Rev. 777, 799–800 (2017) (arguing that the major cases exception should be used only at the Supreme Court level).

  119. [119]. See, e.g., Eskridge & Baer, supra note 32, at 1097 (reporting different empirical patterns in Chevron’s application in different areas of law); Raso & Eskridge, supra note 29, at 1776.

  120. [120]. See Matthew J. Lindsay, Disaggregating “Immigration Law, 68 Fla. L. Rev. 179, 182–83 (2016).

  121. [121]. See id. at 185 n.23.

  122. [122]. See Landon v. Plasencia, 459 U.S. 21, 32 (1982).

  123. [123]. See supra Section II.A; infra Section III.C.

  124. [124]. See supra Section II.A; infra Section III.C.

  125. [125]. See Lindsay, supra note 120, at 185; see also Michael Kagan, Shrinking the Post-Plenary Power Problem, 68 Fla. L. Rev. F. 59, 61–62 (2016) (explaining that immigration law may really be many different bodies of law that share an impact on non-citizens).

  126. [126]. 8 C.F.R. § 1003.0 (2018).

  127. [127]. Id. § 1003.1(b).

  128. [128]. Id. § 1241.1.

  129. [129]. 8 U.S.C. § 1252(a) (2012).

  130. [130]. See Peter L. Strauss, When the Judge Is Not the Primary Official with Responsibility to Read: Agency Interpretation and the Problem of Legislative History, 66 Chi.-Kent L. Rev. 321, 347 (1990).

  131. [131]. See Shruti Rana, Chevron Without the Courts?: The Supreme Court’s Recent Chevron Jurisprudence Through an Immigration Lens, 26 Geo. Immigr. L.J. 313, 325 (2012) (questioning whether immigration agencies have a legitimate claim to special expertise).

  132. [132]. See, e.g., Fregozo v. Holder, 576 F.3d 1030, 1034, 1036 (9th Cir. 2009).

  133. [133]. See Board of Immigration Appeals: Procedural Reforms to Improve Case Management, 67 Fed. Reg. 54,877, 54,878 (Aug. 26, 2002) (codified at 8 C.F.R. pt. 3).

  134. [134]. See Eric M. Fink, Liars and Terrorists and Judges, Oh My: Moral Panic and the Symbolic Politics of Appellate Review in Asylum Cases, 83 Notre Dame L. Rev. 2019, 2020–21 (2008); see also, e.g., Benslimane v. Gonzales, 430 F.3d 828, 829–30 (7th Cir. 2005)(“This tension between judicial and administrative adjudicators ... . isdue to the fact that the adjudication of these cases at the administrative level has fallen below the minimum standards of legal justice. Whether this is due to resource constraints or to other circumstances beyond the Board’s and the Immigration Court’s control, we do not know, though we note that the problem is not of recent origin.” (citations omitted)).

  135. [135]. See Jill E. Family, Beyond Decisional Independence: Uncovering Contributors to the Immigration Adjudication Crisis, 59 U. Kan. L. Rev. 541, 552, 564 (2011) (describing growth of the immigration adjudication backlog and the resulting “immigration adjudication crisis.”); Immigration Court Filings Take Nose Dive, While Court Backlog Increases, TRACImmigration (Oct. 30, 2017), http://trac.syr.edu/immigration/reports/487.

  136. [136]. See, e.g., Abulashvili v. Attorney Gen., 663 F.3d 197, 208–09 (3d Cir. 2011); Dia v. Ashcroft, 353 F.3d 228, 250–51 (3d Cir. 2003).

  137. [137]. Jaya Ramji-Nogales et al., Refugee Roulette: Disparities in Asylum Adjudication, 60 Stan. L. Rev. 295, 296 (2007).

  138. [138]. Benslimane v. Gonzales,430 F.3d 828, 830 (7th Cir. 2005).

  139. [139]. David Zaring, Reasonable Agencies, 96 Va. L. Rev. 135, 183–84 (2010) (finding that agencies appearing before the D.C. Circuit fewer than ten times from 2000 to 2004 prevailed 80% of the time, compared to 68% for agencies appearing before that court more than ten times).

  140. [140]. See Michael M. Hethmon, Tsunami Watch on the Coast of Bohemia: The BIA Streamlining Reforms and Judicial Review of Expulsion Orders, 55 Cath. U. L. Rev. 999, 1008 (2006).

  141. [141]. See Adam B. Cox, Deference, Delegation, and Immigration Law, 74 U. Chi. L. Rev. 1671, 1683–84 nn. 41–42 (2007).

  142. [142]. See Merrill, supra note 113, at 996.

  143. [143]. 8 C.F.R. § 1003.1(d)(ii) (2008).

  144. [144]. See Stephen H. Legomsky, Restructuring Immigration Adjudication, 59 Duke L.J. 1635, 1665 (2010).

  145. [145]. See In re J-F-F-, 23 I & N Dec. 912, 913 (A.G. 2006) (“The Executive Office for Immigration Review, which includes the Board and Immigration Judges, is subject to the direction and regulation of the Attorney General.”).

  146. [146]. Id.

  147. [147]. Id.; see also In re D-J-, 23 I & N Dec. 572, 575 (A.G. 2003).

  148. [148]. See Alberto R. Gonzales & Patrick Glen, Advancing Executive Branch Immigration Policy Through the Attorney General’s Review Authority, 101 Iowa L. Rev. 841, 894–95 (2016). But see Bijal Shah, The Attorney General’s Disruptive Immigration Power, 102 Iowa L. Rev. Online 129, 132–34 (2017) (arguing that the Attorney General’s unique role in adjudication would make expansive use of political decision-making problematic).

  149. [149]. See Dara Lind, Jeff Sessions is Exerting Unprecedented Control Over Immigration Courts—by Ruling on Cases Himself, Vox (May 21, 2018, 1:06 PM), https://www.vox.com/policy-and-politics/
    2018/5/14/17311314/immigration-jeff-sessions-court-judge-ruling.

  150. [150]. See Memorandum from James R. McHenry III, Director, Exec. Office for Immigration Review, to The Office of the Chief Immigration Judge, All Immigration Judges, All Court Administrators, and All Immigration Court Staff (Jan. 17, 2018) (setting case completion targets for immigration court docket management).

  151. [151]. Board of Immigration Appeals: Procedural Reforms to Improve Case Management,
    67 Fed. Reg. 54,878, 54,893 (Aug. 26, 2002) (codified at 8 C.F.R. pt. 1003); see also Stephen H. Legomsky, Deportation and the War on Independence, 91 Cornell L Rev. 369, 372–76 (2006) (describing reforms of the BIA and problems posed for independent immigration adjudication).

  152. [152]. See generally Legomsky, supra note 144 (arguing that immigration adjudication should be transferred to an Article III court so as to provide it the required independence).

  153. [153]. 8 U.S.C. § 1103(a)(1) (2012).

  154. [154]. I.N.S. v. Cardoza-Fonseca, 480 U.S. 421 (1987).

  155. [155]. Id. at 453.

  156. [156]. Id. at 446.

  157. [157]. Id. at 448.

  158. [158]. See Merrill, supra note 113, at 986 (“By the end of the next Term, however, the Court was again applying the Chevron doctrine (irregularly, as ever) to questions of law, and Cardoza-Fonseca quietly dropped from sight.”).

  159. [159]. Id.

  160. [160]. Cardoza-Fonseca, 480 U.S. at 445–48; id. at 453–55 (Scalia, J., concurring).

  161. [161]. Brief for the Petitioner, I.N.S. v. Elias-Zacarias, 502 U.S. 478 (1992) (No. 90-1342), 1991 WL 11003946, at *23.

  162. [162]. See generally Elias-Zacarias, 502 U.S. at 478 (making no mention of Chevron or of deference).

  163. [163]. I.N.S. v. Aguirre-Aguirre, 526 U.S. 415, 424–25 (1999).

  164. [164]. Id. at 424.

  165. [165]. Id. (citing 8 U.S.C. § 1103(a)(1) (1994 ed., Supp. III)).

  166. [166]. Id. at 425–26.

  167. [167]. See Negusie v. Holder, 555 U.S. 511, 517 (2009).

  168. [168]. See Christopher J. Walker, The Ordinary Remand Rule and the Judicial Toolbox for Agency Dialogue, 82 Geo. Wash. L. Rev. 1553, 1579 (2014).

  169. [169]. Negusie, 555 U.S. at 517; Gonzales v. Thomas, 547 U.S. 183, 187 (2006); I.N.S. v. Orlando Ventura, 537 U.S. 12, 16–17 (2002).

  170. [170]. I.N.S. v. Abudu, 485 U.S. 94, 96 (1988).

  171. [171]. See id.

  172. [172]. Scialabba v. Cuellar de Osorio, 134 S. Ct. 2191, 2213 (2014).

  173. [173]. Id. at 2203 (citations omitted).

  174. [174]. Id.

  175. [175]. Id. at 2207.

  176. [176]. Id. at 2214 (Roberts, C.J., concurring).

  177. [177]. The Court also ignored Chevron when an immigration case focused on constitutional questions, but this does not seem to deviate from the normal understanding of the doctrine.
    See generally Kerry v. Din, 135 S. Ct. 2128 (2015) (making no reference to Chevron in the opinion).

  178. [178]. But see Pereira v. Sessions, 138 S. Ct. 2105, 2113 (2018) (avoiding Chevron deference by finding clear meaning in a statute lower courts had found ambiguous). For further discussion of Pereira, see infra Section IV.C.

  179. [179]. I.N.S. v. Cardoza-Fonseca, 480 U.S. 421 (1987).

  180. [180]. Brief for the Petitioner, I.N.S. v. Cardoza-Fonseca, 480 U.S. 421 (1987) (No. 85-782), 1986WL 727528, at *9.

  181. [181]. Cardoza-Fonseca, 480 U.S. at 447.

  182. [182]. See supra note 150 and accompanying text.

  183. [183]. I.N.S. v. Elias-Zacarias, 500 U.S. 915 (1991).

  184. [184]. I.N.S. v. Elias-Zacarias, 502 U.S. 478 (1992).

  185. [185]. I.N.S. v. Aguirre-Aguirre, 526 U.S. 415 (1999).

  186. [186]. Id. at 424 (“Because the Court of Appeals confronted questions implicating ‘an agency’s construction of the statute which it administers,’ the court should have applied the principles of deference described in Chevron.”).

  187. [187]. Negusie v. Holder, 555 U.S. 511 (2009).

  188. [188]. Brief for the Respondent at 10–11, Negusie v. Holder 555 U.S. 511 (2008) (No. 07-499), 2008 WL 3851621, at 10.

  189. [189]. Negusie, 555 U.S. at 517 (“When the BIA has not spoken on ‘a matter that statutes place primarily in agency hands,’ our ordinary rule is to remand to ‘giv[e] the BIA the opportunity to address the matter in the first instance in light of its own expertise.” (alteration in original) (quoting I.N.S. v. Orlando Ventura, 537 U.S. 12, 16–17 (2002)).

  190. [190]. Judulang v. Holder, 565 U.S. 42 (2011).

  191. [191]. Id. at 52 n.7.

  192. [192]. Id. at 52.

  193. [193]. Holder v. Martinez Gutierrez, 566 U.S. 583 (2012).

  194. [194]. Brief for the Petitioner, Holder v. Martinez Gutierrez, 566 U.S. 583 (2012) (Nos. 10-1542 & 10-1523), 2011 WL 5544816, at *33.

  195. [195]. Martinez Gutierrez, 566 U.S. at 597–98.

  196. [196]. Scialabba v. Cuellar de Osorio, 134 S. Ct. 2191 (2014).

  197. [197]. Id. at 2203 (“Principles of Chevron deference apply when the BIA interprets the immigration laws.”).

  198. [198]. Pereira v. Sessions, 138 S. Ct. 2105 (2018).

  199. [199]. Brief for the Respondent, Pereira v. Sessions, 138 S. Ct. 2105 (2018) (No. 17-459), 2018 WL 1557067, at *18–19.

  200. [200]. Juliet Stumpf, The Crimmigration Crisis: Immigrants, Crime, and Sovereign Power, 56 Am. U. L. Rev. 367, 376 (2006).

  201. [201]. While this offense normally carries a ten-year maximum sentence, the Act enhances the penalty to a 15-year minimum sentence. See 18 U.S.C. § 922(g) (2012); id. § 924(b), (e)(1), (e)(2)(B)(ii).

  202. [202]. 8 U.S.C. § 1101(a)(43)(G) (defining “burglary offense” as an “aggravated felony”);
    id. § 1227(a)(2)(A)(iii) (“Any alien who is convicted of an aggravated felony at any time after admission is deportable.”); id. § 1226(c) (prohibiting release on bond for non-citizens with aggravated felony convictions).

  203. [203]. See Taylor v. United States, 495 U.S. 575, 580 (1990).

  204. [204]. See Model Penal Code § 221.1 (Am. Law Inst. 1985).

  205. [205]. See, e.g., Cal. Penal Code § 459 (2017); Nev. Rev. Stat. § 205.060 (2017).

  206. [206]. Taylor, 495 U.S. at 598–602.

  207. [207]. Id. at 599.

  208. [208]. Id. at 601–02.

  209. [209]. See Descamps v. United States, 570 U.S. 254, 257 (2013).

  210. [210]. See id. at 264–65; Moncrieffe v. Holder, 569 U.S. 184, 190–91 (2013); Mellouli v. Lynch, 135 S. Ct. 1980, 1987 (2015).

  211. [211]. Descamps, 570 U.S. at 258–59.

  212. [212]. Id. at 276.

  213. [213]. Moncrieffe, 569 U.S. at 184.

  214. [214]. Id. at 188.

  215. [215]. Id. at 192.

  216. [216]. Id. at 193–94; 8 U.S.C. § 1101(a)(43) (2012).

  217. [217]. 21 U.S.C. § 841(b)(4); Moncrieffe, 569 U.S. at 193–94.

  218. [218]. See Moncrieffe, 569 U.S. at 190 (contrasting the Georgia statute with its federal counterpart).

  219. [219]. Id. (internal quotation marks omitted).

  220. [220]. See Shannon M. Grammel, Note, Chevron Meets the Categorical Approach, 70 Stan. L. Rev. 921, 935–38 (2018) (outlining the rationale behind the categorical approach).

  221. [221]. Chairez-Castrejon, 26 I. & N. Dec. 349, 354 (BIA 2014), superseded in part by In re Chairez-Castrejon, 26 I. & N. Dec. 819 (BIA 2016).

  222. [222]. Chairez-Castrejon, 26 I. & N. Dec. at 354.

  223. [223]. See infra Table 2, notes 243–65.

  224. [224]. See, e.g., Brief for the Respondent, Moncrieffe v. Holder, 569 U.S. 184 (2013) (No. 11-702), 2012 WL 3803440, at *9–12; Brief for the Petitioner, Gonzales v. Duenas-Alvarez, 549 U.S. 183 (2007) (No. 05-1629), 2006 WL 3064108, at *19 n.12 (noting that the BIA had not issued a published decision on the issue at hand). But see Brief for Respondent, Carachuri-Rosendo v. Holder,
    560 U.S. 563 (2010) (No. 09-60), 2010 WL 723015, at *11–14 (summarizing argument without asking for deference, despite a published en banc BIA decision).

  225. [225]. Brief for the Respondents, Lopez v. Gonzales, 549 U.S. 47 (2006) (Nos. 05-547,
    05-7664), 2006 WL 2474082, at *32 n.26 (“While the Board is not entitled to deference in its construction of ... a criminal statute that it has not been charged with administering, the Board’s construction of 8 U.S.C. 1101(a)(43), and particularly its judgment, borne of hands-on experience, about the inadministrability of imposing the hypothetical-federal-felony approach on the INA’s aggravated felony provision, merit deference.”).

  226. [226]. Compare Brief for the Respondent, Nijhawan v. Holder, 557 U.S. 29 (2009) (No. 08-495), 2009 WL 815242, at *14, *4550 (arguing Chevron deference should be applied), with Torres v. Lynch, 136 S. Ct. 1619 (2016) (making no mention of Chevron or deference of any kind), and Nijhawan v. Holder, 557 U.S. 29 (2009) (same).

  227. [227]. Esquivel-Quintana v. Sessions, 137 S. Ct. 1562, 1568 (2017).

  228. [228]. Id. at 1572–73.

  229. [229]. See id. at 1567.

  230. [230]. Id.

  231. [231]. Id. at 1572.

  232. [232]. Compare Pelayo-Garcia v. Holder, 589 F.3d 1010, 1014 (9th Cir. 2009) (a statute criminalizing sex between a 21-year-old and a person under 15 is not categorically sexual abuse of a minor), with United States v. Alvarez-Gutierrez, 394 F.3d 1241, 1243 (9th Cir. 2005) (finding that a Nevada statute criminalizing sex between an 18-year-old and a person under 16 does constitute sexual abuse of a minor).

  233. [233]. Cf. Esquivel-Quintana v. Lynch, 810 F.3d 1019, 1027 (6th Cir. 2016) (Sutton, J., concurring in part and dissenting in part) (“Either reading has much to commend it ... .”), rev’d on other grounds, Esquivel-Quintana v. Sessions, 137 S. Ct. 1562 (2017).

  234. [234]. Mellouli v. Lynch, 135 S. Ct. 1980, 1984 (2015); see also 8 U.S.C. § 1227(a)(2)(B)(i) (2012) (“Any alien who at any time after admission has been convicted of a violation of ... any law or regulation of a State, the United States, or a foreign country relating to a controlled substance (as defined in section 802 of title 21), other than a single offense involving possession for one’s own use of 30 grams or less of marijuana, is deportable.”).

  235. [235]. Mellouli, 135 S. Ct. at 1988–89.

  236. [236]. Id. at 1989.

  237. [237]. Compare 8 U.S.C. § 1227(a)(2)(B) (classifying controlled substance violations generally as a removable offense), with 8 U.S.C. § 1101(a)(43)(B) (defining “illicit trafficking in a controlled substance” as an aggravated felony).

  238. [238]. Mellouli, 135 S. Ct. at 1989.

  239. [239]. Id.

  240. [240]. Id. at 1988.

  241. [241]. Moncrieffe v. Holder, 569 U.S. 184, 190 (2013).

  242. [242]. Roberts: ‘My Job is to Call Balls and Strikes and Not to Pitch or Bat, CNN (Sept. 12, 2005,
    4:58 PM), http://www.cnn.com/2005/POLITICS/09/12/roberts.statement/index.html.

  243. [243]. Leocal v. Ashcroft, 543 U.S. 1 (2004).

  244. [244]. Brief for the Respondents at 5, Leocal v. Ashcroft, 543 U.S. 1 (2004) (No. 03-583), 2004 WL 1617398, at *5 (noting that the BIA had followed circuit court case law, rather than issued its own interpretation of the statute).

  245. [245]. Lopez v. Gonzales, 549 U.S. 47 (2006).

  246. [246]. Brief for the Respondents, Lopez v. Gonzales, 549 U.S. 47 (2006) (No. 05-547), 2006 WL 2474082, at *32 n.26.

  247. [247]. Gonzales v. Duenas-Alvarez, 549 U.S. 183 (2007).

  248. [248]. Brief for the Petitioner, Gonzales v. Duenas-Alvarez, 549 U.S. 183 (2007) (No. 05-1629), 2006 WL 3064108, at *19 n.12.

  249. [249]. Nijhawan v. Holder, 557 U.S. 29 (2009).

  250. [250]. Brief for the Respondent, Nijhawan v. Holder, 557 U.S. 29 (2009) (No. 08-495), 2009 WL 815242, at *14–15, *45–50.

  251. [251]. Carachuri-Rosendo v. Holder, 560 U.S. 563 (2010).

  252. [252]. Brief for the Respondent, Carachuri-Rosendo v. Holder, 560 U.S. 563 (2010) (No. 09-60), 2010 WL 723015.

  253. [253]. Moncrieffe v. Holder, 133 S. Ct. 1678 (2013).

  254. [254]. Brief for the Respondent, Moncrieffe v. Holder, 133 S. Ct. 1678 (2013) (No. 11-702), 2012 WL 3803440.

  255. [255]. Mellouli v. Lynch, 135 S. Ct. 1980 (2015).

  256. [256]. Brief for the Respondent, Mellouli v. Holder, 134 S. Ct. 2873 (2014) (No. 13-1034), 2014 WL 6613094, at *45–52.

  257. [257]. Mellouli, 135 S. Ct. at 1989.

  258. [258]. Id. (“Because it makes scant sense, the BIA's interpretation . . . is owed no deference under the doctrine described in Chevron.”).

  259. [259]. Torres v. Lynch, 136 S. Ct. 1619 (2016).

  260. [260]. Brief for the Respondent, Torres v. Lynch, 136 S. Ct. 1619 (2016) (No. 14-1096), 2015 WL 5626637, at *14.

  261. [261]. Esquivel-Quintana v. Sessions, 137 S. Ct. 1562 (2017).

  262. [262]. Brief for the Respondent, Esquivel-Quintana v. Sessions, formerly Esquivel-Quintana v. Lynch, 137 S. Ct. 1562 (2017) (No. 16-54), 2017 WL 345128, at *36–54.

  263. [263]. Brief for the Respondent, Esquivel-Quintana, 137 S. Ct. 1562, 1572 (2017) (No. 16-54) (“We have no need to resolve whether the rule of lenity or Chevron receives priority in this case because the statute, read in context, unambiguously fore­closes the Board’s interpretation. Therefore, neither the rule of lenity nor Chevron applies.”).

  264. [264]. Sessions v. Dimaya, 138 S. Ct. 1204 (2018).

  265. [265]. Brief for the Petitioner, Sessions v. Dimaya, formerly, Lynch v. Dimaya, 138 S. Ct. 1204 (2018) (No. 15-1498), 2016 WL 6768940, at *24.

  266. [266]. Esquivel-Quintana v. Lynch, 810 F.3d 1019, 1027 (6th Cir. 2016) (Sutton, J., concurring in part and dissenting in part); see also Whitman v. United States, 135 S. Ct. 352, 353–54 (2014) (Scalia, J., statement regarding denial of certiorari).

  267. [267]. Esquivel-Quintana, 810 F.3d at 1028.

  268. [268]. 8 U.S.C. § 1101(a)(43) (2012).

  269. [269]. Id. § 1227(a)(2).

  270. [270]. Id. § 1326(b)(2); id. § 1327.

  271. [271]. See supra notes 255–58.

  272. [272]. Esquivel-Quintana, 810 F.3d at 1024 (discussing Babbitt v. Sweet Home Chapter of Communities for a Great Or., 515 U.S. 687, 704 (1995)), overruled by Esquivel-Quintana v. Sessions, 137 S.Ct. 1562 (2017). But see Whitman v. United States, 135 S. Ct. 352, 353–54 (2014) (arguing that the Court’s willingness to defer in Sweet Home “contradicts the many cases before and since holding that, if a law has both criminal and civil applications, the rule of lenity governs its interpretation in both settings”).

  273. [273]. Esquivel-Quintana, 810 F.3d at 1024.

  274. [274]. Sessions v. Dimaya, 138 S. Ct. 1204, 1217 (2018).

  275. [275]. Id. at 1234 (internal quotation marks omitted).

  276. [276]. Id. at 1231 (Gorsuch, J., concurring).

  277. [277]. Abramski v. United States, 134 S. Ct. 2259, 2274 (2014).

  278. [278]. See Touby v. United States, 500 U.S. 160, 167 (1991) (affirming the power to criminally punish the manufacture of designer drugs when the controlled substance was proscribed by expedited procedure when the administrative power was subject to “multiple specific restrictions” and judicial review was available). But see United States v. Booker, 543 U.S. 220, 230 (2005) (mandatory sentencing factors must be proved to a jury).

  279. [279]. Whitman v. United States, 135 S. Ct. 352, 353 (2014).

  280. [280]. Esquivel-Quintana v. Lynch, 810 F.3d 1019, 1027 (2016) (Sutton, J., concurring in part and dissenting in part) (citation omitted), rev’d on other grounds sub nom., Esquivel-Quintana v. Sessions, 137 S. Ct. 1562 (2017).

  281. [281]. See, e.g., Jordan v. De George, 341 U.S. 223, 231 (1951) (conducting a void for vagueness examination of a ground of deportation because “[d]espite the fact that this is not a criminal statute, we shall nevertheless examine the application of the vagueness doctrine to this case. We do this in view of the grave nature of deportation.”).

  282. [282]. See Gutierrez-Brizuela v. Lynch, 834 F.3d 1142, 1149–56 (10th Cir. 2016) (Gorsuch, J., concurring) (questioning whether Judge Sutton’s rationale against deference can be limited to dual use statutes).

  283. [283]. Glen, supra note 15 (discussing the possibility of a “deportation-is-different” explanation for the Court’s reluctance with regard to Chevron); see also Walker, supra note 15 (discussing the possibility that the Roberts Court may be reluctant to give deference in certain deportation cases).

  284. [284]. See Padilla v. Kentucky, 559 U.S. 356, 365 (2010) (“We have long recognized that deportation is a particularly severe ‘penalty.’” (quoting Fong Yue Ting v. United States, 149 U.S. 698, 740 (1893))); Jordan, 341 U.S. at 231 (“[D]eportation is a drastic measure and at times the equivalent of banishment or exile ... . It is the forfeiture for misconduct of a residence in this country. Such a forfeiture is a penalty.” (quoting Fong Haw Tan v. Phelan, 333 U.S. 6, 10 (1948))). See generally Peter L. Markowitz, Deportation is Different, 13 U. Pa. J. Const. L. 1299 (2011) (describing why deportation is a distinct area of the law).

  285. [285]. Scott v. Illinois, 440 U.S. 367, 373 (1979).

  286. [286]. Ex Parte Bollman, 8 U.S. 75, 82 (1807).

  287. [287]. The Federalist No. 78 (Alexander Hamilton) (McLean ed., 1810) (internal quotation marks omitted).

  288. [288]. Id.

  289. [289]. See Gerstein v. Pugh, 420 U.S. 103, 120 (1975) (“These adversary safeguards [of a criminal trial] are not essential for the probable cause determination required by the Fourth Amendment.”).

  290. [290]. Id.

  291. [291]. Id. at 116.

  292. [292]. Id. at 117–18.

  293. [293]. Alina Das, Unshackling Habeas Review: Chevron Deference and Statutory Interpretation in Immigration Detention Cases, 90 N.Y.U. L. Rev. 143, 146–48 (2015).

  294. [294]. Compare Rodriguez v. Robbins, 804 F.3d 1060 (9th Cir. 2015) (failing to mention Chevron in a case concerning pre-removal mandatory detention), rev’d sub nom. Jennings v. Rodriguez, 138 S. Ct. 830 (2018), with Lora v. Shanahan, 804 F.3d 601, 609 (2d Cir. 2015) (applying Chevron to interpretation of the mandatory detention statute).

  295. [295]. Das, supra note 293, at 149.

  296. [296]. Zadvydas v. Davis, 533 U.S. 678, 682 (2001).

  297. [297]. Brief for the Petitioners at 44, Reno v. Ma, 533 U.S. 678 (2001) (Nos. 99-7791, 00-38), 2000 WL 1784982, at *43–44.

  298. [298]. Zadvydas, 533 U.S. at 697.

  299. [299]. Id. at 690.

  300. [300]. Clark v. Martinez, 543 U.S. 371 (2005).

  301. [301]. Id. at 382.

  302. [302]. Id. at 402 (Thomas, J., dissenting).

  303. [303]. Demore v. Kim, 538 U.S. 510, 539 (2003).

  304. [304]. Brief for the Petitioners, Demore, 538 U.S. 510 (2003) (No. 01-1491), 2002 WL 31016560, at *9.

  305. [305]. See Michael Kagan, Jennings v. Rodriguez Might Not Be About Immigration After All,
    36 Yale J. on Reg.: Notice & Comment (Mar. 2, 2018), http://yalejreg.com/nc/jennings-v-rodriguez-might-not-be-about-immigration-after-all.

  306. [306]. Brief for the Petitioners, Jennings v. Rodriguez, 138 S. Ct. 830 (2018) (No. 15-1204), 2016 WL 5404637, at *18, *52, *54.

  307. [307]. See Jennings v. Rodriguez, 138 S. Ct. 830 (2018).

  308. [308]. See Family, supra note 7, at 103 (noting that a potential alliance between immigration advocates and conservative critics of Chevron could pose a threat to the entire doctrine).

  309. [309]. See, e.g., Gutierrez-Brizuela v. Lynch, 834 F.3d, 1142, 1155 (Gorsuch, J., concurring) (arguing that administrative agencies have the power).

  310. [310]. Id. at 1156.

  311. [311]. See supra Section II.A.

  312. [312]. See Scott v. Illinois, 440 U.S. 367, 373–74 (1979) (denying right to appointed counsel to defendant sentenced to a fine because “the Sixthand Fourteenth Amendments to the United States Constitutionrequire only that no indigent criminal defendant be sentenced to a term of imprisonment unless the State has afforded him the right to assistance of appointed counsel in his defense”); Duncan v. Louisiana, 391 U.S. 145, 160–61 (1968) (finding no right to a jury trial when the defendant is charged only with petty offenses).

  313. [313]. 8 U.S.C. § 1229a(c) (2012).

  314. [314]. Id. § 1227(a)(1)(B).

  315. [315]. Id. § 1227(a)(2) (setting out criminal grounds of removal).

  316. [316]. Id. § 1229a(c)(4).

  317. [317]. 8 C.F.R. § 208.14 (2018).

  318. [318]. 8 U.S.C. § 1231(b)(3); 8 C.F.R § 208.16.

  319. [319]. 8 U.S.C. § 1229b.

  320. [320]. See, e.g., Reyes v. Lynch, 842 F.3d 1125, 1129 (9th Cir. 2016) (holding, in an asylum eligibility case, that “[w]e conclude that the BIA’s articulation of its ‘particularity’ and ‘social distinction’ requirements for demonstrating membership in a ‘particular social group’ are entitled to Chevron deference”).

  321. [321]. See supra Section III.C.

  322. [322]. See Landon v. Plasencia, 459 U.S. 21, 32 (1982).

  323. [323]. See generally Pereira v. Sessions, 138 S. Ct. 2105 (2018) (considering eligibility for cancellation of removal).

  324. [324]. Id. at 2109.

  325. [325]. Id. at 2113.

  326. [326]. Id. at 2114.

  327. [327]. Id. at 2113.

  328. [328]. Id. at 2121 (Alito, J., dissenting).

  329. [329]. Landon v. Plasencia, 459 U.S. 21, 32 (1982) (“Our cases have frequently suggested that a continuously present resident alien is entitled to a fair hearing when threatened with deportation ... .”).

  330. [330]. See 8 U.S.C. § 1252 (2012) (prohibiting judicial review on matters of discretion); see also Diallo v. Holder, 715 F.3d 714, 716 (8th Cir. 2013) (“[T]here is no constitutionally protected liberty interest in discretionary relief from removal.”).

  331. [331]. See supra note 330 and accompanying text.

  332. [332]. Gutierrez-Brizuela v. Lynch, 834 F.3d 1142, 1144 (10th Cir. 2016).

  333. [333]. See Kristin E. Hickman, Of Lenity, Chevron, and KPMG, 26 Va. Tax Rev. 905, 920 (2006) (discussing application of lenity and Chevron in tax cases).

  334. [334]. See United States v. Thompson/Center Arms Co., 504 U.S. 505, 518 (1992); Crandon v. United States, 494 U.S. 152, 168 (1990).

  335. [335]. See Whitman v. United States, 135 S. Ct. 352, 353 (2014) (denying cert.); Esquivel-Quintana v. Lynch, 810 F.3d 1019, 1027–28 (6th Cir. 2016) (Sutton, J., concurring in part and dissenting in part).

  336. [336]. See Rebecca Sharpless, Zone of Nondeference: Chevron and Deportation for a Crime, 9 Drexel L. Rev. 323, 350 (2018); Brian G. Slocum, The Immigration Rule of Lenity and Chevron Deference, 17 Geo. Immigr. L.J. 515, 525–26 (2003). But see David S. Rubenstein, Putting the Immigration Rule of Lenity in its Proper Place: A Tool of Last Resort After Chevron, 59 Admin. L. Rev. 479, 488–90 (2007) (arguing that lenity should be relevant in immigration cases only if deference to the agency fails to resolve all ambiguity, and that lenity should not apply with the same force that it does in criminal cases).

  337. [337]. Leocal v. Ashcroft, 543 U.S. 1, 11 n.8 (“Because we must interpret the statute consistently, whether we encounter its application in a criminal or noncriminal context, the rule of lenity applies.”).

  338. [338]. Clark v. Martinez, 543 U.S. 371, 380 (2005).

  339. [339]. See, e.g., I.N.S. v. St. Cyr, 533 U.S. 289, 320 (2001).

  340. [340]. See, e.g., Babbitt v. Sweet Home Chapter of Communities for a Great Or., 515 U.S. 687, 704 (1995) (discussing the Endangered Species Act); United States v. Thompson/Center Arms Co.,
    504 U.S. 505, 518 (1992) (discussing the National Firearms Act).

  341. [341]. See David Hahn, Note, Silent and Ambiguous: The Supreme Court Dodges Chevron and Lenity in Esquivel-Quintana v. Sessions, Minn. L. Rev. De Novo (Nov. 29, 2017), http://www.minnesota
    lawreview.org/2017/11/silent-and-ambiguous.

  342. [342]. Crandon v. United States, 494 U.S. 152, 158 (1990).

  343. [343]. Esquivel-Quintana v. Lynch, 810 F.3d 1019, 1028 (6th Cir. 2016) (Sutton, J., concurring in part and dissenting in part).

  344. [344]. See, e,g., Whitman v. United States, 135 S. Ct. 352, 353 (2014) (Scalia, J., statement respecting denial of certiorari).

  345. [345]. See Slocum, supra note 336, at 575; Rubenstein, supra note 336, at 517–19.

  346. [346]. See United States v. Thompson/Center Arms Co., 504 U.S. 505, 518 (1992); Crandon,
    494 U.S. at 168; see also Maracich v. Spears,133 S. Ct. 2191, 2209 (2013); Kasten v. Saint–Gobain Performance Plastics Corp.,563 U.S. 1, 16 (2011); Chapman v. United States, 500 U.S. 453, 463 (1991).

  347. [347]. Rubenstein, supra note 336, at 505 –10.

  348. [348]. Das, supra note 293, at 202–05.

  349. [349]. Id. at 205.

  350. [350]. Id.

  351. [351]. Id.

  352. [352]. Brief for the Petitioner at 18–20, Torres v. Lynch, 136 S. Ct. 1619 (2016) (No. 14-1096), 2015 WL 4967191, at *18–20.

  353. [353]. See supra Section III.D.

  354. [354]. See supra Section III.D.

  355. [355]. See United States v. Castleman, 572 U.S. 157, 173 (2014) (holding that lenity only applies in cases of “grievous ambiguity”); Chapman v. United States, 500 U.S. 453, 463 (1991); United States v. Wiltberger, 18 U.S. 76, 77 (1820) (“Though penal laws are to be construed strictly; yet the intention of the legislature must govern in the construction of penal, as well as other statutes, and they are not to be construed so strictly as to defeat the obvious intention of the legislature.”).

  356. [356]. See Hickman, supra note 333, at 935.

  357. [357]. Family, supra note 7, at 116–17.

  358. [358]. Id.

  359. [359]. See generally César Cuauhtémoc García Hernández, Crimmigration Law 3 (2015) (explaining the emergence of “crimmigration law” which is the “intersection of criminal law and procedure with immigration law and procedure”).

  360. [360]. Family, supra note 7, at 117.

*

B.A. Northwestern University, J.D. University of Michigan Law School. Professor of Law, The University of Nevada, Las Vegas, William S. Boyd School of Law.



I am grateful to Gil Kahn for ideas, input, and interest that made this Article possible. My thanks are due to Gabriel “Jack” Chin, Jill E. Family, César Cuauhtémoc García Hernández, David Rubenstein, Rebecca Sharpless, and Christopher J. Walker for helpful suggestions. All errors are mine.