I. Introduction

Chevron deference is at the height of its powers in refugee and asylum cases, with the highest possible human consequences. Important curtailments on eligibility for asylum that began under President Obama and accelerated under President Trump have largely survived judicial challenge because courts have been willing to extend maximum deference to the Executive Branch on the interpretation of refugee law.[1] Even more dramatic limitations on asylum instituted by the Trump Administration may be affirmed under Chevron if the Biden Administration does not change course. The heightened importance of Chevron in refugee and asylum cases stands in contrast to the diminished importance that this doctrine has played in other types of immigration cases, in which the Supreme Court in particular has consistently ignored government requests for deference when considering legal grounds for deporting and detaining immigrants.[2] The role of deference in refugee cases illustrates the potential for immigration questions to scramble the politics surrounding Chevron deference.[3] Immigrant rights advocates have joined conservatives in attacking Chevron and its progeny.[4] Meanwhile, Democratic senators attacked President Trump’s judicial nominees for lacking fealty to Chevron deference.[5] The fact that Trump-era immigration restrictions might end up owing their durability to Chevron deference is at least a little ironic since in its judicial nominations, the Trump Administration had been portrayed as an arch‑opponent of the Chevron doctrine.[6]

At the moment this Article is being finalized in early 2021, the transition from Trump to Biden has put the future of American immigration policy into considerable flux. But despite this turmoil, some clear patterns have emerged in how the Supreme Court uses Chevron in immigration cases. The Supreme Court has avoided meaningful application of Chevron deference in two key types of immigration cases: grounds of removal and immigration detention.[7] The Court has never explicitly explained this as a rule, but the pattern is clear, and it is consistent with the general principle that judicial deference is inappropriate when physical liberty is at stake.[8] Or as several authors have put it: “Deportation is different.”[9] And yet, Chevron deference has exerted a powerful and consistent influence on a closely related area of law that, in practice, effectively determines whether tens of thousands of people are deported every year: eligibility for asylum.[10] Recent decisions by President Trump’s Attorneys General to narrow asylum protections, followed by new regulations limiting asylum, rely on this deference and have been testing how far the courts are actually willing to go with it.[11]

One of the earliest invocations of Chevron deference by the Supreme Court was an asylum case, I.N.S. v. Cardoza-Fonseca, and it remains a seminal case in the field of asylum law.[12] But in many ways, the full importance of Chevron in asylum cases has become clearer in recent years at the circuit court level. This can be seen in cases debating the nexus clause of the refugee definition, a provision that effectively means that asylum-seekers will be denied protection and deported even when there is no doubt that they are in grave danger.[13] The government has prevailed in much of this litigation by relying on Chevron deference.[14] Close reading hints that judges in some of these cases have had doubts about the interpretations of the asylum statute offered by the Board of Immigration Appeals but have affirmed removal orders because of deference to the agency.[15] In other words, these are cases in which the Chevron revolution is really happening and with the highest possible stakes.[16] Lives are literally at risk.

This Article explores whether Chevron deference really makes sense in refugee cases. It joins several others that have reassessed the role of Chevron in immigration cases. In particular, Shoba Sivaprasad Wadhia and Christopher J. Walker have recently argued for a limitation on Chevron deference based on a process question.[17] They argue that justifications for deference are particularly weak in the context of immigration adjudication.[18] Instead, they argue that Chevron would be appropriate when the Attorney General interprets the Immigration and Nationality Act through rulemaking.[19] This Article addresses the problem by focusing primarily on the substantive subject matter, rather than the process by which an agency interprets a statute. Whether adopted through adjudication or rulemaking, should courts really defer to the Executive Branch on the interpretation of asylum law?

This question is urgent.[20] The efforts of President Trump’s Attorneys General to narrow eligibility for asylum may determine the fates of tens of thousands of asylum-seekers fleeing gang violence in Central America.[21] It was clear that the Department of Justice intended to defend these efforts to restrict asylum through heavy reliance on Chevron.[22] In perhaps the most important change to the American understanding of what it means to be a refugee, Attorney General William Barr issued a decision purporting to eliminate family-based persecution as a basis for asylum.[23] Barr acknowledged openly that this change went against a well-developed body of circuit court caselaw,[24] but he emphasized the congressionally delegated power given to him to change interpretations of the law under Chevron and Brand X.[25] As if to answer the possible complaint that Chevron is ill-fitted to statutory interpretations issued through administrative adjudication, during the last weeks of the Trump Administration, the Department of Justice codified its new restrictive view of asylum law in regulations.[26] Even if the incoming Biden Administration undoes these new rules, they are a sign that as immigration policy has become more volatile, presidents are more likely to push their authority to radically reshape asylum law through the administrative process, especially if the courts remain highly deferential in this area of law.

During the Trump years, some courts have used arbitrary‑and-capricious review to restrain the political instability that may result from a straightforward application of Chevron deference. Most notably, as I explain in Part VIII, the D.C. Circuit held that one of the Trump Administration’s key decisions limiting asylum eligibility would pass under Chevron, but nevertheless failed arbitrary-and-capricious review because it significantly departed from well-established policy, without adequate explanation.[27] Other circuits have declined to reach this conclusion, but they have done so by interpreting new policies in narrow terms, thus minimizing the degree to which they change the law and allowing them to survive arbitrary-and-capricious review.[28] This new interest in arbitrary-and-capricious review, which has been encouraged by Chief Justice Roberts and the Supreme Court, will be an important trend to watch.[29] It seems to signal an uneasiness with the legal instability that results from deference doctrines like Chevron.

So why does the Supreme Court seem so comfortable with Chevron deference in asylum cases when it has been reluctant to defer to the government on other kinds of deportation cases? More to the point, is this deference justified? The applicability of Chevron deference may determine whether people in danger of mortal harm receive protection. It also offers broader lessons about the relative strength of various justifications for this central doctrine in modern administrative law. In particular, it shines a light on the political accountability rationale, which is often offered as a justification for deference to the Executive Branch. Heavy reliance on the political branches to interpret a statute in a highly partisan policy arena, like immigration, is likely to lead to significant instability in statutory interpretation and may undermine important jurisprudential principles.

The purpose of this Article will be two-fold.

First, it will explore how Chevron deference appears to have had a significant impact in asylum law. It will identify the reasons why asylum applications may be different from other cases in which deportation is at stake. There are formalistic differences, beginning with the fact that asylum is a form of relief from removal, not a ground for removal. There may also be different concerns at play in this area of law, which may make agency expertise or political accountability more important.

Second, the Article will reassess whether deference in asylum cases is warranted. This assessment builds, in part, on the theory that the Supreme Court has had good reason to not apply deference in deportation and detention cases.[30] Asylum cases are in every practical sense about deportation, and if anything, the stakes for the immigrant are even higher in a case where there is a fear of severe human rights violations. This is not the case in all immigration cases that reach the federal courts.[31] There is also good reason to doubt that the Attorney General has actually handled this area of law in a manner that supports a claim for deference. As Judge Posner wrote, “Deference is earned; it is not a birthright.”[32] And yet, as we will see, there are cogent arguments justifying more deference in asylum cases than in other kinds of deportation cases. These arguments rest to a great extent on the premise that greater political accountability is a good thing when interpreting a statute. Yet that proposition in effect encourages politicization of immigration adjudication, a phenomenon that is already happening, has proven to be highly controversial, and may pose serious problems.

Part II summarizes, based on previous research, patterns in the Supreme Court’s use of Chevron deference in immigration cases, showing how asylum cases are treated differently than deportation and detention cases. Part III offers a brief primer for the uninitiated on the basic vocabulary of asylum and refugee law, especially as it relates to the definition of a “particular social group” and the requirement that a refugee must be in danger for a certain reason. Part IV looks at deference in circuit court asylum caselaw, showing that Chevron has proven to be extremely influential in these cases. Part V assesses possible rationales for deference in asylum cases and concludes that political accountability and presidential prerogatives over foreign affairs offer the strongest justification for deference in asylum law. Part VI highlights the dilemma posed by politicization of administrative adjudication as a means of interpreting statutes, including assessing the role of arbitrary-and-capricious review as an additional limitation on executive authority. Part VII asks the central question: Is politicization a virtue or a vice? Part VIII examines the inclination of some courts, including the D.C. Circuit, to impose arbitrary-and-capricious review on top of Chevron deference, in order to reduce the legal instability that comes with enhanced political accountability.

II. Patterns of Chevron Deference in Immigration Appeals

Chevron deference involves two famous steps when an administrative agency interprets a statute that it administers.[33] The first asks whether the intent of Congress is clear from the statute. 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.”[34] The most prominent justification for this deference is that it respects congressional intent to delegate interpretation of a law to the agency responsible for implementing it.[35] But the Supreme Court has offered other rationales as well. One is political accountability. The central idea here is that interpreting statutes often requires making policy choices, and these policy choices are better left to the Executive Branch than to the courts.[36] Another justification for deference is the proposition that agencies have technical expertise that helps them interpret complex statutes.[37]

Deference by a court to an executive agency on a matter of statutory interpretation ought to require a compelling justification. Section 706 of the Administrative Procedure Act assigns courts the responsibility to resolve “all relevant questions of law” and “constitutional and statutory provisions.”[38] Justice Scalia, a prominent proponent of Chevron deference for much of his career on the Supreme Court, warned early on that “[i]t is not immediately apparent why a court should ever accept the judgment of an executive agency on a question of law.”[39] He wrote that in 1989. Today, it would seem even more urgent to examine whether there is a good reason for deference because the Supreme Court seems increasingly unsure about the doctrine. Some Justices have directly questioned its constitutionality.[40] Even before these doubts surfaced explicitly, it had become clear that the Supreme Court has been extremely inconsistent in its application of Chevron deference.[41] As Justice Kennedy wrote in an immigration case in his last term, “[I]t seems necessary and appropriate to reconsider . . . the premises that underlie Chevron and how courts have implemented that decision.”[42]

Chevron deference would seem to have particularly strong foundations in immigration law, a field in which deference to the Executive Branch predated modern administrative law.[43] Congress has explicitly stated in the Immigration and Nationality Act that the Attorney General’s determination on questions of law “shall be controlling.”[44] In most cases, the Attorney General exercises this power through adjudication, either by decisions he makes himself, or through decisions by the Board of Immigration Appeals (BIA).[45] In a 1999 withholding case, the Supreme Court said, “It is clear that principles of Chevron deference are applicable to this statutory scheme.”[46] In 2014, the Court was more emphatic in a case concerning eligibility for a family-sponsored visa: “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.”[47]

While the Supreme Court has talked as if Chevron’s application in immigration law is a simple and settled matter, the Court’s actions paint a different picture. There have been at least four Supreme Court immigration cases decided since 1999 in which the Court has meaningfully applied Chevron deference.[48] But none of these involved grounds of removal or immigration detention.[49] In a recent study, I found that in many decisions “concerning the BIA’s interpretation of criminal grounds of removal the Supreme Court has simply failed to even mention the existence of Chevron.”[50] Meanwhile, in other criminal grounds‑for‑removal cases, the Court mentioned Chevron but avoided actually giving any deference.[51] In an immigration case, Justice Alito complained that the Court was ignoring Chevron as if it had been overruled “in a secret decision that has somehow escaped my attention.”[52] It seems that Chevron matters a lot in some immigration cases but not in others.

Administrative law scholars have long questioned whether the Supreme Court is consistent in applying Chevron deference and whether consistency should even be expected. In an oft-cited study, William N. Eskridge Jr. and Lauren E. Baer found broad inconsistency across more than 1,000 Supreme Court cases where Chevron should have been applicable.[53] Because of this inconsistency, Eskridge and others argued that “scholars are being unrealistic when they demand that the Supreme Court adopt and consistently apply formal deference regimes.”[54] Instead, Chevron’s analytical framework offers “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.”[55] The Court itself has hinted that it does not expect itself to be rigorously consistent.[56] More recently, Natalie Salmanowitz and Holger Spamann have raised doubts about the empirical basis for this skepticism about Supreme Court consistency.[57] In a replication study, they found—contrary to Eskridge and Baer—that the Court had actually been quite consistent with Chevron.[58] They suggested that the flaw in the original study was that it ignored whether the parties asked the Court to apply deference.[59] When the cases in which no one asked for deference in the briefs are removed from the analysis, the Court appeared to invoke Chevron more reliably.[60]

My own research is narrower, focusing only on immigration cases, but it may offer a bridge between these two opposing views of whether the Supreme Court is consistent in applying Chevron. First, my study took note of whether the government asked for Chevron deference and found that this factor did not explain the Court’s failure to apply Chevron in certain types of immigration cases.[61] My research anticipated the type of nuanced analysis advocated by Salmanowitz and Spamann, but nevertheless found the following:

In at least seven decisions . . . concerning the BIA’s interpretation of criminal grounds of removal the Supreme Court has simply failed to even mention the existence of Chevron. . . . In some cases the Department of Justice did not ask for deference, usually because there was no published BIA decision at issue. 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. 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, [yet] the Court still ignored Chevron entirely in its decision.[62]

I also argue that Eskridge and his colleagues had erred in concluding that the Court’s inconsistency was essentially idiosyncratic.[63] If that were a correct description, then it would be reasonable to conclude that the Supreme Court applies Chevron, at best, as a loose set of guiding principles, not as a binding rule. But what if it is not random? If clear patterns can be detected, they may help us better understand the situations in which the Justices feel comfortable with deference, and those where they do not.[64] In other words, we should not be afraid to look for patterns and significance in the Court’s failure to apply Chevron in cases where it would seem to be relevant.[65]

Much like Salmanowitz and Spamann, I agree that we need to pay attention not just to the abstract normative question about Chevron’s applicability in a particular case, but to a longer list of factors that would make the Court’s failure to apply deference more or less noteworthy.[66] As the Justices seem to be growing more doubtful about Chevron deference, we should pay attention to both loud anti-Chevron decisions and soft anti-Chevron decisions.[67] The loud ones are decisions where the Court explicitly announced a limitation on Chevron.[68] The major cases exception, for matters of “deep economic and political significance,” in King v. Burwell would be an obvious example.[69] But we should also pay attention to soft decisions, “where the Supreme Court failed to apply Chevron when it ostensibly should have mattered or applied it in such a way as to render the doctrine irrelevant.”[70] These cases present themselves in two different ways, both evident in the Supreme Court’s immigration jurisprudence. Sometimes, the Court simply fails to even mention Chevron, as if it doesn’t even exist.[71] Other times, the Court mentions Chevron but seems to give it no real force.[72]

The mere fact that in a single case the Court ignores Chevron or does not seem to apply it with full force does not mean much on its own. Moreover, given that Chevron deference has become a fraught topic, the Justices might be inclined to avoid it if they can find another way to decide a case.[73] But that does not mean that Chevron has no relevance, in the sense that the Court might still turn to it when it really might matter. To paraphrase a line often attributed to Sigmund Freud, sometimes inconsistency is just inconsistency.[74] Yet if the Court avoids relying on Chevron over and over again in a specific type of case while relying on it heavily and regularly in another type of case, then it’s not really inconsistency at all. Instead, the pattern indicates the circumstances in which the Supreme Court finds Chevron most applicable and those for which it finds it inappropriate or at least more fraught.[75] Such “soft” cases allow the Supreme Court to quietly test and refine the appropriate boundaries of the doctrine, without prematurely stating a rule.

When this kind of nuanced, factor-sensitive analysis is conducted in immigration cases, we do not find idiosyncrasy. Instead, we find a pattern. While the Supreme Court has applied Chevron deference in many types of immigration cases, it has quite consistently avoided meaningful deference in cases concerning grounds of removal and detention.[76] The typical version of this case involves a legal immigrant who is convicted of a state criminal offense, leading to a question of law about whether that conviction is one of the removable offenses listed in the Immigration and Nationality Act.[77] One theory offered to explain these cases is that some of these grounds of removal are “dual-use” statutes, since the same ground of removal may be a civil ground for deportation in Immigration Court, and also an element of a crime.[78] For example, a conviction that counts as an aggravated felony would allow for the removal of a legal resident.[79] But if a person reentered the United States after an aggravated felony, it would constitute an element of a criminal offense as well.[80] Perhaps the reason these deportation cases are treated differently for Chevron purposes is that Chevron does not apply to elements of a criminal offense and thus cannot apply to a dual-use statute.[81]

The dual-use theory is elegant, but it does not explain the full pattern for two reasons. First, the Supreme Court has also avoided meaningful Chevron deference in cases concerning grounds of removal that are not dual-use.[82] This was the case in Mellouli v. Lynch, for instance, which involved removal for a purported violation of the Controlled Substances Act but not for an aggravated felony.[83] In that case, the Court mentioned Chevron but gave no real deference because, it said, the BIA’s interpretation of the statute made “scant sense.”[84] Second, and perhaps more important, the Court also avoids Chevron deference in cases involving immigration detention. This was true in two cases that are more than a decade old—Zadvydas v. Davis[85] and Clark v. Martinez.[86] It was also true in detention cases decided in 2018 and 2019, Jennings v. Rodriguez[87] and Nielsen v. Preap.[88] Thus, it is no longer enough to say that deportation is different. Instead, it seems that deportation and detention are both different for Chevron purposes. I have explained this as a physical liberty exception to Chevron, on the theory that detention and deportation both involve deprivations of physical liberty, and nondeferential judicial review is especially sacred in this circumstance.[89]

All of this still leaves Chevron deference very much intact in a wide range of immigration matters. One variety would be visa applications, which involve people wanting to come to the United States or sponsor their relatives to come.[90] Another concerns discretionary relief.[91] Another—and the most important for this Article—are claims for relief from removal, which includes asylum cases. Asylum law played a role in Chevron’s early history with the 1987 decision in INS. v. Cardoza-Fonseca.[92] The Court ignored Chevron (despite the Solicitor General requesting it) in a subsequent asylum case in 1992, INS. v. Elias-Zacarias.[93] But since then the Court has been more consistent. One of the most frequently cited invocations of Chevron’s applicability in immigration is an asylum case, INS. v. Aguirre-Aguirre,[94] while another asylum case, Negusie v. Holder, became a vehicle for the Court to reinforce the ordinary remand rule, which is closely related to the command that courts defer to the agency on interpretive matters.[95]

In sum, the Supreme Court has consistently avoided meaningful application of Chevron deference in cases concerning grounds of deportation and detention. By contrast, it has usually applied deference in other types of immigration cases. The next parts will describe the outsized influence Chevron has had on asylum cases in the circuit courts.

III. A Brief Primer on Asylum

I offer here a brief overview of refugee law in order to provide context for the role of Chevron in asylum cases. In order to qualify as a “refugee,” and thus in order to win asylum, a noncitizen must show that she is outside her country of nationality “and is unable or unwilling to avail himself or herself of the protection of[] that country because of persecution or a well-founded fear of persecution on account of race, religion, nationality, membership in a particular social group, or political opinion.”[96] The refugee cases at issue in this Article concern the definition of “membership in a particular social group,” which has long been one of the hardest to interpret phrases in the refugee definition.

There are many high-stakes questions about the United States’ asylum system. The Trump Administration sought to change a number of policies that are currently the subject of litigation,[97] but I will not attempt to address most of them in this Article, in part because much of the related litigation is unresolved as I write this. These issues include the question of whether asylum-seekers may be forced to wait in Mexico while their applications are pending,[98] whether passing through a third country before reaching the southern border may be a bar to asylum,[99] and whether entering the country illegally may be a bar to asylum.[100] I will instead focus solely on the legal definition of a refugee—the core eligibility criteria for asylum. The refugee definition is routinely contested in petitions for review of orders of removal filed by immigrants in the circuit courts of appeal. In these appeals, the Department of Justice has often relied on Chevron deference to defend limitations on eligibility for asylum.[101]

The requirement that a refugee must fear persecution on account of an enumerated protected ground (race, religion, nationality, membership in a particular social group, or political opinion) is one of the harshest limitations in asylum law. It means that a person who is genuinely in danger of severe harm (having “a well-founded fear of persecution”) may be denied protection and deported because she would not be persecuted for the right reason.[102] Race, religion, nationality, or political opinion all have their potential interpretive pitfalls, but asylum cases based on these grounds have tended to be more straightforward. When a person flees imprisonment because he has participated in antigovernment demonstrations, the crux of the case will be whether the evidence is strong enough that he is genuinely in danger, but there typically will not be a dispute that this type of danger would qualify.[103] When a person makes an asylum claim because he says he has converted from one religion to another and that his government would torture him for doing do so, there may be doubt about whether he is manufacturing the conversion to generate an asylum claim, but usually there will not be doubt that this kind of claim, if credible, should succeed.[104] But when a person flees gender-based violence or threats from a criminal group, it is not necessarily enough to show that she is genuinely in danger, nor that she has already been a victim of grievous abuse.[105] These cases depend on the far more ambiguous category “membership in a particular social group.”[106] The results are often strikingly harsh because it leads judges to effectively say: We know you will be killed or raped, but not for the right reason. A recent decision by the Eleventh Circuit illustrates this stark reality:

The Immigration Judge (IJ) denied relief because, although Perez-Zenteno was beaten and brutally raped and her daughter kidnapped, she failed to prove that she was persecuted on account of membership in a statutorily protected group. The social group offered was neither sufficiently particular nor socially distinct. . . . Because we too agree that Perez-Zenteno has failed to establish membership in a particular social group, as defined by Congress, and because no nexus has been shown, we hold that the petition must be denied.[107]

It is probably no surprise that the law in such cases would be hotly contested. In these cases, Chevron’s fullest potential power may be seen in action, with the gravest consequences.

Defining “membership in a particular social group” has been the subject of legal to-and-fro since at least the 1990s, affecting cases of severe gender-based violence and also violence by criminal gangs who target families and children, among others. The Board of Immigration Appeals (BIA) has long struggled to develop a coherent interpretation of “particular social group” and to apply it consistently.[108] From the mid-1980s until 2006, the BIA defined a social group by fundamental or immutable characteristics.[109] The BIA reached this first understanding by applying a canon of construction to the statute. Since “particular social group” appears in a list of other protected grounds, the principle of ejusdem generis called for interpreting this category to be analogous to race, religion, nationality, and political opinion.[110] This interpretive approach was well-received by circuit courts, but the BIA nevertheless began to move away from it.[111] Eventually, in 2014, the BIA added two new criteria. Now, in addition to being defined by a fundamental or immutable characteristic, a particular social group must be “socially distinct” and possess “particularity.”[112] This new framework has largely been accepted by the circuit courts, although not without some efforts to limit its impact or to require further case-by-case adjudication.[113] As Fatma Marouf noted, the BIA itself has struggled to provide consistent and coherent guidance about how the criteria should be applied.[114]

In 2018, Attorney General Jeff Sessions issued a major decision, Matter of A-B-.[115] Attorney General Sessions largely affirmed the BIA’s new test for a particular social group but rejected a precedent decision that allowed asylum claims from women fleeing domestic violence.[116] Attorney General Sessions rejected the somewhat convoluted way the BIA had previously analyzed domestic violence cases but did not offer direction about the kind of asylum claims that might succeed.[117] In fact, by discarding a convoluted but constraining BIA precedent, Attorney General Sessions may have opened the door for some immigration judges to use a simpler and broader definition of asylum eligibility. The Attorney General rejected a social group defined as “married women who are unable to leave the relationship,” but some immigration judges have since decided that domestic violence victims might instead be persecuted simply because they are part of the particular social group of “women,” full stop.[118] The Attorney General also appeared to raise the standard for asylum when people flee harm from nonstate actors. Whereas the well‑established rule was that the asylum-seeker must show that her government was merely unable or unwilling,[119] Matter of A-B- says that “[t]he applicant must show that the government condoned the private actions ‘or at least demonstrated a complete helplessness to protect the victims.’”[120] As I will discuss in Part VIII, the D.C. Circuit and, to a lesser extent, other circuits have seen this “complete helplessness” standard as new and significantly different.[121]

In a number of respects, Matter of A-B- was as much an example of political rhetoric as it is a precedent decision offering a legal interpretation. Attorney General Sessions offered a lengthy decision expressing skepticism about asylum claims based on fears of criminal actors, but this may be entirely dicta.[122] In line with Sessions’s public rhetoric on asylum and immigration,[123] he seemed to lecture asylum-seekers:

[T]here are alternative proper and legal channels for seeking admission to the United States other than entering the country illegally and applying for asylum in a removal proceeding. . . . Aliens seeking an improved quality of life should seek legal work authorization and residency status, instead of illegally entering the United States and claiming asylum.[124]

The restrictive view of asylum eligibility in Matter of A-B- was in sync with the Trump Administration’s generally hostile and restrictive view of these migrants.[125] Narrowing eligibility for asylum makes it easier to deport these asylum-seekers. It also bolsters public rhetoric arguing that they never had valid asylum claims anyway, although that claim depends on what should even be considered a valid asylum case.[126] This complementarity might be a compelling argument for giving the Executive Branch space to interpret ambiguous statutes in line with the President’s policy agenda. But it might equally be a reason for concern that adjudication has been politicized, which might be a reason for federal courts to scrutinize decisions more rigorously.[127] I will discuss this concern in more detail in Part VII.

Attorney General William Barr also issued a precedent decision on the definition of membership in a particular social group.[128] Since 1985, the BIA had embraced “family” as a quintessential example of a particular social group.[129] The full reach of this understanding was often contested, so the BIA raised caution that the validity of a family group may depend on “the nature and degree of the relationships involved and how those relationships are regarded by the society in question.”[130] The BIA reaffirmed the essential validity of family-based refugee claims as recently as 2017.[131] In July 2019, however, Attorney General Barr overruled that decision and issued a holding at odds with decades of established law, finding that “in the ordinary case, a nuclear family will not, without more, constitute a ‘particular social group.’”[132] The Attorney General acknowledged that several circuit courts held that family ties can in fact define a particular social group.[133] The Ninth Circuit, for example, found that “there is nothing in the statute itself, nor in the BIA’s interpretation of the relevant provisions, to suggest that membership in a family is insufficient, standing alone, to constitute a particular social group in the context of establishing eligibility for asylum.”[134]

The Attorney General argued that the circuit courts had gotten this wrong and that some of their decisions had been undermined by later caselaw.[135] But he did not rely on this claim alone. He argued that he had the authority to replace the circuit courts’ interpretations with his own, first because Congress had explicitly delegated interpretation of the Immigration and Nationality Act to the Attorney General,[136] and second because of Chevron deference.[137] He stressed Brand X deference because it requires Chevron deference “even in cases where the courts of appeals might have interpreted the phrase differently in the first instance.”[138] Barr’s reason for preemptively, and perhaps a bit defensively, citing Brand X is clear, given that he was critiquing courts that have the potential to invalidate his decision. As of early October 2019, no circuit had issued a decision reviewing the Attorney General’s decision in Matter of L-E-A-.[139] However, the Fifth Circuit—which had actually never reviewed family as a social group—already noted in dicta: “Matter of L-E-A- is at odds with the precedent of several circuits.”[140] If Attorney General Barr’s new interpretation had any prospect for surviving judicial scrutiny, it would seem to be only because Chevron deference would be applied with maximal force. Meanwhile, on December 11, 2020, the Executive Office of Immigration Review (which supervises the Immigration Courts) and the Department of Homeland Security published sweeping new regulations, after a notice-and-comment period, codifying the most restrictive reading of Matter of A-B-.[141] The new regulations were set to go into force on January 11, 2021, just days before the inauguration of a new president, but were enjoined by a district court because they were approved by Chad Wolf, who was improperly installed as Acting Secretary of Homeland Security.[142]

IV. Deference in Circuit-Level Asylum Cases

Measuring the true practical impact of Chevron requires a search for a very particular sort of case. As a rule, requiring deference from the judiciary, such as Chevron deference, should make it easier for the government to win when an agency’s legal interpretation is challenged. But the government would win many cases when its interpretation of legislation is challenged anyway, even under de novo review. Moreover, even under a robust application of Chevron, the government should not always win because Chevron does not call for an affirmance when the statute is clear or the agency’s interpretation of it is unreasonable. Chevron’s true impact thus appears in a very specific kind of decision. It comes in cases where the agency is not clearly wrong, but not clearly right, either. Chevron matters in cases where the court might have ruled the other way if not for deference. Courts do not say this very often. In fact, I do not know of any ideal-type example in immigration law in which a court directly says, “I think the agency is wrong, and I would have ruled the opposite way under de novo review, but only because of the deference required I will affirm.” But some recent cases litigating the boundaries of the asylum definition hint at this.

A case where Chevron appeared to have been decisive is the 2018 Third Circuit decision, S.E.R.L. v. Attorney General.[143] The petitioner was a Honduran woman whose daughter had already been granted asylum because the daughter had been abducted, raped, and stalked by two men, including her stepfather.[144] The mother feared they would abuse her too.[145] The Immigration Judge found her account of the facts to be credible, although the IJ quibbled some about how much the mother had actually been abused in the past.[146] Regardless, the crux of the case was whether “S.E.R.L.'s proposed particular social group—immediate family members of Honduran women unable to leave a domestic relationship—lacked the requisite particularity and social distinction” to be considered a “particular social group” for purposes of asylum.[147] The dual requirement that an asylum applicant show that the proposed group is particular and socially distinct was, at that time, a departure from the long-standing test used by the BIA, known as the Acosta test, which had required that a particular social group be “a group of persons all of whom share a common, immutable characteristic.”[148]

S.E.R.L. was decided in 2018, but by that time the circuit courts had been battling with the Board of Immigration Appeals about the definition of a particular social group for more than a decade. The BIA had taken steps to transform the Acosta test by 2006.[149] But its initial attempts did not fare well in all circuits, including in the Third Circuit.[150] Much of this concerned the BIA’s attempts to impose a “social visibility” requirement, which several circuit courts had found to be inconsistently and unclearly defined in the Board’s own caselaw.[151] The Board persisted, issuing two precedent decisions in 2014 that refined the criteria.[152] In these cases, the Board replaced its failed “social visibility” test with a new “socially distinct” criteria. Now, an applicant for asylum must show that she is persecuted on account of membership in a group that is “(1) composed of members who share a common immutable characteristic, (2) defined with particularity, and (3) socially distinct within the society in question.”[153] The BIA also clarified that “particularity” meant that the group must have “definable boundaries.”[154] “Social distinction” meant not literal “visibility” but rather “whether the people of a given society would perceive a proposed group as sufficiently separate or distinct.”[155]

For purposes of this Article, the important question is the role that Chevron played in the circuit courts’ treatment of the BIA’s evolving approach. In S.E.R.L., the Third Circuit followed a textbook Chevron two-step analysis. The particular social group requirement was not clearly defined in the statute.[156] The BIA’s new interpretation is reasonable and so it was affirmed.[157] But the most telling part of the decision may be the way the court treated the petitioner’s arguments against the Board’s new standard: primarily, that the BIA had not meaningfully refined an interpretation that had already been rejected; that it was acting inconsistently in different cases; and that it had acted against well-tested canons of construction.[158] The court said, “Those critiques raise legitimate concerns.”[159] The court examined the critiques extensively.[160] But the court concluded: “[N]otwithstanding our concerns, we conclude that the requirements are reasonable and warrant Chevron deference.”[161] This language—notwithstanding our concerns—is as strong an indication as we will ever normally find that Chevron deference was decisive to the outcome. The court is signaling openly that it is sympathetic to critiques of the government’s new interpretation of the statute, but it is good enough for Chevron Step Two. Or, put another way, had this been de novo review, or maybe even Skidmore deference, the court might have ruled the other way.

The Third Circuit’s doubts about the Board’s interpretation do not seem to be unique. The Ninth Circuit affirmed the same BIA interpretation under a Chevron analysis.[162] The Ninth Circuit did not communicate concerns about it as openly as the Third Circuit, but it found the BIA’s interpretation to be reasonable only after attaching to it a number of caveats about the degree to which it was really new.[163] The Seventh Circuit has apparently withheld judgment.[164]

Two caveats are worth noting. First, it should come as no particular surprise that the circuit courts apply Chevron deference with more consistency and rigor than does the Supreme Court.[165] In some circuits, Chevron deference in immigration cases has a superficially separate life from Supreme Court caselaw, such that Chevron is not even cited in cases that clearly invoke the doctrine that we know as Chevron deference. This is the case in the Eighth Circuit, for example. Sometimes, the circuit will specifically state that it is applying Chevron.[166] But other times, it cites to its own caselaw.[167] This is to a large extent a formality, but it indicates symbolically the degree to which Chevron in the circuits is different from the Supreme Court. A second caveat is this: The government does not necessarily need Chevron to win in asylum cases. In nearly all of these cases, once the court affirms the Board’s general framework of what may constitute a particular social group, it then examines whether the specific proposed group in that case could qualify.[168] The Eleventh Circuit recently decided a case in which it was unsure whether Chevron applied to a single‑member Board decision but resolved in the alternative that the proposed group would fail under any interpretation of the statute.[169]

Given this backdrop, it is not surprising that Matter of A-B- has fared well under Chevron analysis in the circuit courts.[170] And yet, while Matter of A-B- has generally been affirmed under Chevron, it has also been somewhat minimized. For example, the Fifth Circuit affirmed Matter of A-B-, but only after stating that it does not categorically exclude all domestic-violence-based asylum claims and stating that “A-B- [does] not constitute a change in policy” because it relied on “standards firmly established in BIA precedents.”[171] The Ninth Circuit also affirmed Matter of A-B- but dismissed much of the Attorney General’s decision as rhetorical rather than a legal text.[172] I will return to this tendency of courts to affirm while also minimizing administrative action when I address arbitrary-and-capricious review in Part VIII.

Chevron deference in asylum cases is notable for two main reasons. First, on this subject, the circuit courts are acting in line with Supreme Court tendencies, not in contrast to them. As we have already seen, the Supreme Court has avoided Chevron in grounds-of-removal cases, but it has more consistently applied it in asylum cases. Second, Chevron seems to be deciding high-stakes cases at a time when its standing at the Supreme Court is in doubt. The Third Circuit stated this openly, in one of the several passages of its decision that seemed to signal misgiving about the legal interpretation that it affirmed: “The Chevron doctrine of deference to federal agencies is open to question, but it is the law, and it allows the BIA to change its statutory interpretation and still be entitled to full deference from Article III courts.”[173]

V. Why Defer in Asylum Cases?

So far, my main point has been observational. Chevron does not seem to matter much in cases about immigration detention and grounds of removal, but it matters a lot in other immigration cases and has been decisive in some extremely high-stakes circuit cases about the boundaries of asylum law. The next question is normative: Is this pattern defensible?

The Supreme Court has said that Chevron deference should apply in all immigration cases.[174] But given the Court’s divergent behavior, a baseline question is whether the Supreme Court has effectively stated its own rule in overly broad terms. Applying Chevron in all immigration cases is not what the Supreme Court has actually done. In my previous study of this issue, I argued that nondeference in grounds of removal cases is entirely appropriate because of a basic separation of powers concern:

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.[175]

Of course, separation of powers is the foundation for a broader attack on Chevron writ large.[176] For present purposes, I need not take a position on this broader attack. If Chevron is invalid in its entirety, then it is not applicable in immigration cases of any kind and we need not go any further. My argument is narrower. Separation-of-powers concerns should be especially heightened when deportation or detention is at stake, so that even if the Supreme Court does not overturn Chevron entirely, it is right to make an exception in these matters.

What, then, do we make of asylum cases? Why should Chevron play such an outsized role in high-stakes asylum cases when it does not in other deportation cases? There is a straightforward argument that they should be treated the same: Asylum cases are fundamentally about deportation, in that deportation is what will happen if the government wins in a petition for review in the federal courts, affirming a denial of asylum. Moreover, these are deportation cases in which a person may be subject to grave harm. An argument may be made that, if anything, given the high stakes, there should be less deference in asylum cases.

The interpretation of the refugee definition involves the implementation of an international treaty.[177] The Supreme Court has long recognized that statutory eligibility criteria for asylum is a means of implementing treaty obligations, specifically obligations under the 1951 United Nations Convention Relating to the Status of Refugees.[178] Maureen Sweeney argues that, instead of Chevron, interpretation of the refugee definition should be governed by the rule in Murray v. Schooner Charming Betsy, which held that courts should avoid statutory interpretations that clash with international law.[179] However, this argument may not fully resolve the question of whether Chevron should apply. In the seminal case where it acknowledged the role of international law in asylum cases, the Supreme Court did not see any reason to question the applicability of Chevron deference.[180] The Charming Betsy doctrine could be invoked in the course of applying Chevron’s normal analysis, for instance, by resolving statutory ambiguity at Step One. That would presumably constrain the amount of deference afforded to the Executive Branch without rejecting the Chevron doctrine outright.

There are several arguments that may justify Chevron deference in asylum law, more so than in other immigration cases. These include formalistic differences between grounds-of-removal and asylum eligibility, special reasons for deference on matters of foreign affairs, reinforcing political accountability by deferring to political branches on policy choices, and expertise. I conclude that the political-accountability rationale is the most persuasive, though it raises countervailing concerns about the use of administrative adjudication to make high-stakes policy choices.

A. Formalism and Proof Burdens

Treating asylum eligibility different from grounds of removal flows naturally from the structure of removal (deportation) proceedings. A standard removal case in Immigration Court begins with the Department of Homeland Security bearing the burden of proof to show that the respondent is removable from the United States.[181] The cases where the Supreme Court has avoided meaningful Chevron deference were decided at this stage. But once removability is established, the respondent has the opportunity to ask for relief from removal, of which asylum is one variety.[182] The respondent bears the burden of proof in applications for relief from removal, which operate in a manner loosely analogous to an affirmative defense in a criminal case.[183]

Does the fact that the respondent, rather than the government, bears the burden of proof justify Chevron deference in some way? Allocation of the burden of proof is a different issue than allocation of responsibility for interpreting a statute, and it is by no means self-evident that one should flow from another. It is important to remember in this context that the questions that end up adjudicated in federal appellate cases under Chevron are questions about legal eligibility. Decisions that are discretionary are generally outside the courts’ jurisdiction entirely, rendering Chevron irrelevant.[184] But the fact that the respondent bears the burden may yet indicate something important. The reason the burdens of proof in deportation cases are allocated this way reflect the fact that, on claims of relief, the respondent is trying to claim a benefit to which she is not by default entitled. By contrast in removability, the government is seeking to use force against someone (to deport them). Asylum in a removal proceeding only arises when the government has already proven that a person may be removed from the country and she seeks relief from removal. Perhaps these differences make deference in relief-from-removal cases a less direct threat to separation of powers than in ground‑of‑removal cases.

The Supreme Court has, in one important and very recent case, ignored Chevron in adjudicating eligibility for relief from removal. Pereira v. Sessions concerned cancellation of removal, for which calculating the length of residence in the United States is an important criterion.[185] The clock on residence in the United States stops when a removal proceeding begins—known as the stop time rule. But in Pereira, the Court found that time does not stop if the Notice to Appear that initiated removal proceedings was not completed properly. The lower court had found the statute ambiguous, leading to a Chevron analysis.[186] The Supreme Court found clarity in the statute that the lower court could not see, so “the Court need not resort to Chevron deference, as some lower courts have done.”[187] This led to Justice Alito’s complaint that the Court was ignoring Chevron without overruling it[188] and to Justice Kennedy’s call for the Court to revisit the validity of Chevron entirely.[189] The fact that this discussion among the Justices occurred in a relief-from-removal case suggests that this formal distinction between removability and relief might not be a correct description of the Court’s pattern any longer. But for now, Pereira is just one data point, and it appears to be an exception. Moreover, in Pereira, the majority did at least mention Chevron and purport to make a Step-One decision, which it does not always do in removability cases.[190]

It is important to note in this context that the fact that a respondent bears the burden of proof and the government gets Chevron deference on the law makes for a very steep hill to climb for anyone fighting deportation through an asylum application. The cases where circuit courts actually wrestle with the way the BIA has construed asylum law are relatively exceptional. Circuit courts can also affirm asylum denials by deference to findings of fact, which are upheld if backed by substantial evidence. Circuit courts often rely on this, even in cases that could raise legal disputes about the definition of a particular social group.[191]

B. Expertise

A longstanding, but somewhat troubled, argument for deference is that executive agencies have an advantage over courts in terms of technical expertise on an area of law.[192] The D.C. Circuit has said that deference to an agency is often “particularly strong when the [agency] is evaluating scientific data within its technical expertise.”[193] On the other hand, Justice Scalia argued that Skidmore deference is sufficient for courts to take due account of technical explanations offered by experts, arguing that Chevron is better justified as a recognition of congressional intent to trust executive agencies to make policy choices.[194] However, these two rationales are not mutually exclusive. Congress may want to rely on agencies to make policy choices because they have “‘unique expertise,’ often of a scientific or technical nature.”[195] Thus, to defer to an agency’s expertise may be a way to implement congressional intent.

Immigration cases rarely involve any kind of scientific expertise.[196] They almost always raise classic problems of fact and law, which would seem to dilute any claims that an executive body has a relative advantage compared with courts. That is especially so when the agency that would get deference operates through adjudication, using a decision-making process that mirrors those of a court, with briefing, examination of a record, and reasoned written decisions.[197] Deference sometimes might be justified by the complexity of a statutory scheme.[198] But the theory that the BIA has more expertise than federal courts on this particular statutory scheme cannot be accepted without a caveat: The circuit courts of appeal decide a lot of asylum cases.[199] The circuit courts, thus, do not necessarily lack expertise on the relevant questions of law. We know that judges tend to become less deferential as they encounter more and more cases in a particular area of law.[200] The immigration cases where the Supreme Court has seemed most reluctant to defer have involved interpretations of criminal law and constitutional due process.[201] It would not be surprising for an Article III court to feel that no other branch of government has an advantage in resolving such problems. It could be that federal judges are more willing to defer on the interpretation of asylum law, which is famously amorphous, than on the interpretation of federal criminal law. But on the whole, there is good reason to be skeptical that expertise is a fully convincing reason for disparate treatment between asylum and grounds of deportation.

C. Foreign Affairs

The application of Chevron in immigration cases flows naturally from the fact that immigration law is a species of administrative law. But the fact that immigration touches on foreign affairs is sometimes offered as an additional reason for granting Chevron deference.[202] In a case about denying a visa to a would-be immigrant who was still outside the United States, the Supreme Court said: “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.”[203]

This rationale should not be accepted mechanically, however. Not all immigration questions invoke foreign policy concerns in a meaningful way.[204] But perhaps, this concern does offer a reason to see asylum and refugee cases differently from removability cases.

Removability cases often involve the question of whether a state criminal conviction counts as a ground of removal under federal immigration law.[205] This problem calls for the categorical approach, assessing whether there is a match between a federal “generic” definition of a crime and a particular state’s definition of an analogous crime. These cases focus on interpreting American criminal law and spotlight the federal system of American government, stressing the interplay between federal law and the laws of the several states. Yet they do not implicate foreign affairs in any particularly obvious way, other than the fact that the respondent is a foreign national. All of the relevant law and facts are American. If these cases concern foreign affairs, then perhaps so do a wide variety of common contracts and torts cases that happen to involve foreign nationals. Asylum cases are different. In an asylum case, all the relevant facts are about events in a foreign country. The legal questions are about how to assess the conduct of foreign actors. These concerns may make the executive’s authority over foreign affairs a more salient concern.

Foreign-affairs concerns offer a possible reason to distinguish asylum from cancellation-of-removal cases, like Pereira v. Sessions. Both cancellation and asylum are forms of relief from removal, but their substantive concerns are very different. While asylum focuses on persecution in a foreign country, cancellation of removal is exclusively domestic in orientation.[206] One form of cancellation of removal is for long-term legal residents and is granted to eligible people as a matter of discretion.[207] Another form of cancellation benefits long-term undocumented residents and makes its central eligibility criteria a showing of “extremely unusual hardship” to an American citizen or legal resident, should the respondent be removed from the country.[208] This focus on hardship for Americans (and U.S. legal residents) turns the spotlight far away from foreign affairs. In a typical cancellation‑of‑removal case, the factual inquiry will focus on the medical or educational needs of a disabled U.S.-citizen child or parent,[209] while an asylum case would focus on whether a police force in a foreign country is corrupt or repressive.

While there is a close conceptual fit between foreign policy concerns and asylum law, the case should not be overstated. First, asylum law is still law. It is not a matter of discretion, and if Congress had meant it to be tied directly to the foreign policy concerns of the day, it might have made asylum more discretionary. We have an example of such an immigration status in the form of temporary protected status, which is declared by the Attorney General for particular groups of immigrants.[210] But it is not something for which an individual can petition, unless their nationality has already been designated.[211] A similar example would be the President’s authority to exclude a class of foreigners from entering the United States, which implicates national security directly.[212] By contrast, asylum law sets out statutory criteria for eligibility drawn directly from an international treaty, the 1951 Refugee Convention.[213] In a sense, the key foreign policy choice was to incorporate this treaty into enforceable domestic immigration law. Once Congress did this, it was debatable whether shifting foreign policy concerns remained an interpretive concern. This nexus to an international treaty is an argument for invoking the Charming Betsy doctrine, which would constrain the Executive Branch’s interpretive leeway and may narrow the scope of deference available.[214]

The administrative means by which asylum law is interpreted and administered also weakens the foreign policy rationale for deference. The interpretation of the refugee definition is not handled by the state department or the White House. Instead, it is interpreted through adjudication, primarily by the Board of Immigration Appeals, which is supposed to exercise independent judgment.[215] It is not particularly clear why the BIA should be regarded as having any foreign affairs expertise, nor why its adjudications should be seen as a vehicle for foreign policy. Arguably, decisions by the Attorney General, who can overrule the BIA, may be more clearly tied to the President’s agenda. But this system of adjudication would be put under considerable strain if the Attorney General appears to be implementing a policy agenda rather than engaging in neutral adjudication.[216] Moreover, the reliance on adjudication and the use of precedential decisions cuts down on executive flexibility, which is different from other immigration cases where the President is free to make quick decisions related to foreign policy.[217]

The history of the American asylum system indicates substantial efforts to shield asylum adjudication from foreign policy and other political concerns.[218] A key innovation of the 1980 Refugee Act, which is the foundation of our asylum system, was the elimination of ideological limitations on refugee policy.[219] The new Act made questions of law, not foreign policy, the central eligibility criteria and eliminated the Attorney General’s discretion about whether to withhold removal when a person showed she was in danger of persecution under international law.[220] Circuit courts have, at times, criticized the BIA when it has appeared overly dependent on state department assessments.[221] The asylum system was reshaped in 1990 by the settlement in American Baptist Churches v. Thornburgh, known as the ABC settlement. In the ABC settlement, the government agreed to issue new regulations, making clear that

foreign policy and border enforcement considerations are not relevant to the determination of whether an applicant for asylum has a well-founded fear of persecution [and] the fact that an individual is from a country whose government the United States supports or with which it has favorable relations is not relevant to the determination of whether an applicant for asylum has a well-founded fear of persecution.[222]

The ensuing regulations established a specialized asylum corps, broadened the evidence that should be considered in asylum adjudication, and reduced the role of the state department.[223] The Supreme Court has since avoided deciding whether the United States can deny a refugee claim solely on foreign policy grounds.[224] All of these measures, with the ABC settlement being most explicit, suggest that foreign-affairs-motivated interference in asylum adjudication is actually not a good thing—and, thus, not a good reason to defer to the outcomes.

D. Policy Choice and Political Accountability

The theory of Chevron’s Step Two is that when there are two or more reasonable interpretations of a statute, the question depends at least partly on a policy choice.

[A]n 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.[225]

The Third Circuit alluded to such a policy choice in explaining why the Board of Immigration Appeals abandoned an interpretation of “particular social group” that had been widely accepted by the courts:

Eventually, the BIA determined that the Acosta test had proven to be over-inclusive and unworkable, in part because it encompassed virtually any past acts or experiences, since the past cannot be changed and is, by definition, immutable. Thus, in 1999, the BIA began supplementing the Acosta test with additional requirements.[226]

The lynchpin of this shift was the assessment that the prior test had been over-inclusive. That is, at its core, a policy judgment. A reasonable person could just as easily conclude that the asylum definition is meant to be inclusive.

Once it is clear that an interpretive question depends on a value judgment about the best policy orientation, it may make sense to allow this judgment to be made by a political branch of government. Then-Professor Elena Kagan argued that this is often a reason to allow presidential influence over a decision that does not require particular expertise.[227] Presidential influence over such decisions establishes transparency and “an electoral link between the public and the bureaucracy.”[228] But there are downsides. For one thing, just because a statute is difficult to interpret does not mean that the right answer cannot be found nor that all interpretations are equally valid. Also, as with foreign affairs, the peculiar structure of immigration adjudication adds complexity to the policy-choice rationale. The BIA is supposed to be insulated from politics and the election of a new president does not immediately change its membership.[229] This severely weakens the argument for direct political accountability as a reason for deference to the BIA. But here again, the fact that the Attorney General has preeminence becomes very important. This makes asylum law potentially quite responsive to political shifts in a new Administration through the appointment of a new Attorney General. As Wadhia and Walker argue, the political accountability theory of Chevron could be more convincing in the immigration context if the Attorney General engages the public through a notice-and-comment process.[230]

VI. The Dilemma of Politicized Administrative Adjudication

One of Chevron’s key virtues is that it constrains political partisanship in the federal courts.[231] One way that Chevron arguably accomplishes this is by focusing on the political accountability of the other branches of government and shifting responsibility for the difficult choices to them. However, there is an elephant in the room. The interpretations of statutory law that have generated the circuit court decisions on the meaning of a particular social group have been promulgated through adjudication of individual cases, in which certain cases are designated as precedents.[232] As Wadhia and Walker thoroughly explore, this is an awkward mechanism by which to apply the theoretical virtues of Chevron deference.[233] Moreover, immigration adjudication in particular has been the source of considerable critique and controversy.[234]

Administrative adjudication is always a strange animal, especially when adjudication is entrusted to a law enforcement agency, as it is in immigration cases with the Department of Justice. This puts an agency, and potentially a single official like the Attorney General, in a position where he must serve “as both judge and civil servant.”[235] This creates tension between the expectation that adjudication will be impartial and the fact that the Executive Branch is inherently political.[236] Among administrative adjudicators, immigration judges stand out in that they work for a prosecuting agency. The Attorney General defends orders of removal in federal appellate courts, much as a district attorney would defend criminal convictions on appeal.[237]

Some scholars have argued that there was a longstanding assumption that presidents would not assert political control over adjudication.[238] Catherine Kim noted that even advocates of presidential control over administrative agencies have often seen adjudication as distinct.[239] The idea that deference is warranted because it leaves space for policy choice, and thus for political accountability, suggests that it is actually a good thing for immigration adjudication to be driven by partisan or policy preferences. But such tendencies are rarely treated as a good thing.[240]

Kim argues that political interference in immigration adjudication may be in tension with the rule in the canonical cases of Londoner v. City of Denver and Bi-Metallic Investment Co. v. State Board of Equalization of Colorado.[241] These cases highlight the importance of a fair hearing when an individual is singled out for hardship by a government action, which is certainly the case in a deportation proceeding.[242] Full and fair hearings are seen as essential in deportation and asylum cases; courts have criticized immigration judges who have indicated otherwise.[243] Lower courts have said that in deportation proceedings “[a] neutral judge is one of the most basic due process protections.”[244]

When agencies act as adjudicators, they often appear to be mirroring the procedures of the judiciary, which is a natural response to demands that they show impartiality. This is certainly the case in immigration adjudication, where the adjudicative body is called a court, although it is part of the Department of Justice, and it uses an adversarial process with motions, briefs, and judge‑issued decisions that, in function, look much like a trial level court.[245] The BIA and the Attorney General designate certain decisions as precedents, much the same way as circuit courts of appeal choose to “publish” some of their opinions to be binding precedents.[246] Courts do this to constrain themselves through the rule of stare decisis, on the theory that a judicial decision is an application of rules of law, not political discretion.[247] By constraining its own actions through precedent, the BIA limits the potential for politics to influence policy and thus undermines one of the supposed virtues of Executive Branch rulemaking.

If the virtue of Executive Branch decision-making is political accountability, then imitating the judiciary through the use of precedent would seem counterproductive. Reliance on precedent in the Executive Branch constrains the President’s agenda.[248] A new Administration that wants to change policy through adjudication has options, as President Trump’s Attorneys General have shown. The most straightforward mechanism would be for a new Attorney General to refer a new case to him or herself, and to issue a new precedent decision, reversing the old precedent. That is what Attorney General Barr did with Matter of L-E-A-. But there is another potential route. A new administration might be able to discard an old rule with which it disagrees by simply declining to ask a reviewing court to apply Chevron deference.[249] Some scholars have questioned whether agencies should actually be allowed to waive Chevron deference because this can become a means to circumvent regular decision-making processes.[250] However, for now, this appears to be a potentially viable route.

The constitutional appropriateness of non-Article III adjudication rests on the Supreme Court’s longstanding view that Congress may delegate to the Executive Branch adjudication of “public rights” cases that deal with the relationship between the government and individuals subject to its authority.[251] Christopher J. Walker notes that “political control over agency adjudication that implicates core life, liberty, or property interests potentially raises due process concerns.”[252] One response to this problem would be to transfer as much adjudication as possible to Article III courts, but this is an ambitious proposition.[253] Walker suggests that a more modest solution might be to strip adjudicatory decisions of Chevron deference, which would, in effect, mean that agency heads—like the Attorney General—would be restrained from using adjudication of individual cases to make major policy decisions.[254] He observes that this might be a way of implementing Chief Justice Roberts’s view that the public‑rights doctrine should be limited to administrative adjudicators who are “adjuncts” to the federal courts, in that their role would be limited mostly to findings of fact.[255] This approach would seem to reverse the rule in Chenery II that adjudication can be a means of establishing generally applicable rules.[256] President Trump’s former White House Counsel Don McGahn called for just that.[257]

Wadhia and Walker argue that deference on interpretation of immigration law is more justified when Administrations use the rulemaking process rather than adjudication.[258] This might open up asylum law to more public engagement through the notice‑and‑comment process.[259] The Trump Administration engaged rulemaking to restrict asylum eligibility for people who transit through third countries before seeking protection in the United States,[260] and in its final days, on the entire scope of asylum eligibility.[261] Some judges faulted the Trump Administration for not going through notice and comment with earlier high-stakes changes in asylum policy.[262]

Going through notice and comment rather than relying on adjudication could strengthen arguments in favor of deference by utilizing a process more open to the public. However, favoring rulemaking over adjudication would put stress on a different question: Should statutory interpretation be a fundamentally political exercise, or should it be independent, analytical, and fundamentally judicial?

VII. Is Politicization a Virtue or Vice?

Much seems to depend on whether politicization of statutory interpretation is really a good thing. In theory, a benefit of having an agency head closely supervise adjudication is that it encourages consistency and control across disparate adjudicators.[263] However, in the immigration context, the Attorney General’s personal involvement in adjudication is not necessary to establish consistent rules since the Board of Immigration Appeals can issue precedent decisions of its own that are binding on all Immigration Courts. What the Attorney General’s involvement adds is politics. The Attorney General is a cabinet-level appointee of the President. By contrast, the BIA is actually designed to be explicitly insulated from politics.[264] If political accountability is a good reason for courts to defer to an agency, then the intervention of the Attorney General should make the argument for deference stronger. The same could be said for making major changes to asylum eligibility through rulemaking. If the real goal is to enhance political accountability for statutory interpretation, this kind of intervention in asylum law would be healthy. If the voters do not approve, in theory they can express their will through the ballot. But to make statutory interpretation a partisan, rather than judicial, endeavor may be fraught with significant downsides.

Trump-era interventions by the Attorney General in the Immigration Courts have made increasingly urgent the question of whether reinterpretation of asylum law, among other major immigration law issues, by political appointees is a good thing. In addition to narrowing asylum eligibility criteria, President Trump’s Attorneys General have asserted control over the way immigration judges (IJs) manage their dockets procedurally. Catherine Kim and Amy Semet’s empirical study of immigration court adjudication suggests that IJs are more likely to order removal under the Trump Administration that in prior administrations, regardless of which President originally appointed the IJ.[265] Trump’s Attorneys General curtailed immigration judges’ leeway to grant continuances.[266] They limited IJs’ authority to terminate cases[267] and to administratively close them.[268] Attorney General Jeff Sessions spoke of the immigration courts as a tool of the Trump Administration’s overall enforcement strategy.[269] There have been claims that the Administration is using ideology as a factor in hiring new immigration judges.[270] These interventions attracted considerable criticism.[271]

During the Trump Administration, aggressive interventions in immigration and asylum adjudication by the Attorney General put new pressure on the already difficult question about the role of policy preferences in asylum cases. While the Attorney General has long had the power to adjudicate high-stakes cases rather than leave them to the BIA, not every Attorney General has done so to the same degree and there is considerable debate about whether it is a good thing.[272] Regardless, it is clear that President Trump’s Attorneys General are using this authority to its fullest extent.[273] It has long been clear that immigration judges are mere “employees” of the Department of Justice, but the trappings and procedures of the immigration courts blurred this classification, allowing the immigration courts to look and function much like a regular court.[274] But recent changes have challenged that, including a number of new requirements that immigration judges manage their dockets in particular ways and the decision to remove tools that the judges had to delay close cases without issuing a removal order. These interventions have raised public concern about the neutrality of the immigration courts.[275]

Of particular relevance here are the interventions of Attorney General Jeff Sessions in asylum law, which were rhetorical as well as substantive. While he was Attorney General, Sessions made two speeches to immigration judges in the Executive Office for Immigration Review (EOIR) in which he used rhetoric deriding asylum claims, and asylum-seekers’ attorneys, in terms that are difficult to imagine coming from a judge in an Article III court. In one, he said that “caselaw . . . has expanded the concept of asylum well beyond Congressional intent” and that “[w]e also have dirty immigration lawyers who are encouraging their otherwise unlawfully present clients to make false claims.”[276] In another, he spoke to a group of new immigration judges, complaining that “[g]ood lawyers, using all of their talents and skill, work every day—like water seeping through an earthen dam—to get around the plain words of the INA to advance their clients’ interests. Theirs is not the duty to uphold the integrity of the act.”[277]

Part of these remarks reflects a view that asylum law had been interpreted too broadly and is the same as the reason the BIA began narrowing the definition of a particular social group. But while this may be a defensible (if debatable) legal view, the assertion that lawyers who argue the opposite position are acting against the “integrity” of the law and may be “dirty” is a step beyond the collegiality normally expected in the legal profession. If this were said by a judge in another context—if, for instance, a judge hearing criminal cases said that defense lawyers are “dirty” and that arguments favoring defendants endangered the integrity of the law—there would be clear arguments that due process was violated. Circuit courts have reminded immigration judges that they “must assiduously refrain from becoming advocates for either party.”[278] Due process is violated when an immigration judge appears to have prejudged the merits of asylum claims.[279]

Before the election of Donald Trump, Alberto Gonzales and Patrick Glen argued that the President should make greater use of the Attorney General’s authority to reshape immigration law by issuing precedential decisions.[280] They argued that this mechanism was more legitimate than executive actions such as President Obama’s Deferred Action for Childhood Arrivals program (DACA), “thus may be a less controversial method by which to advance immigration policy.”[281] The Trump Administration has followed their recommendations and has successfully illustrated the potential reach of the Attorney General’s power to change the law if courts do not intervene. But the promise that this mechanism would defuse controversy has not been borne out. Instead, asylum adjudication has been inserted into the realm of presidential immigration law, where key policies are no more stable than the political fortunes of a particular president or presidential candidate.[282] By January 2020, all of the leading Democratic candidates for President had promised to reverse Attorney General Jeff Sessions’s decisions limiting asylum eligibility.[283] The most likely result from such politicization is instability and inconsistency as policies shift with the election cycles.

The experience of the Trump presidency suggests that favoring rulemaking over adjudication would only partially answer the concerns raised by this politicization. On asylum law, the Trump Administration started with adjudication, with cases like Matter of A-B-, and ended with rulemaking, especially the December 11, 2020 publication of the rule that immigrant advocates dubbed the “death to asylum” regulation.[284] It seems unlikely that the incoming Biden Administration would be under any less pressure to reverse the policies just because a notice‑and‑comment process had taken place. It also seems entirely foreseeable that in future presidential elections, rival candidates will continue to promise to tighten or loosen asylum policies, which will in turn mean that interpretation of asylum law may change every four years.

The instability that comes with politicization is troubling. It is one thing for different presidents to exercise discretionary powers differently. But on refugee law, the instability from one presidential administration to the next is about the interpretation of a nondiscretionary statutory text, one that is drawn from an international treaty. The Federalist Papers warned against such instability in the law:

The internal effects of a mutable policy are still more calamitous. It poisons the blessing of liberty itself. It will be of little avail to the people, that the laws are made by men of their own choice, if the laws be so voluminous that they cannot be read, or so incoherent that they cannot be understood; if they be repealed or revised before they are promulgated, or undergo such incessant changes that no man, who knows what the law is to-day, can guess what it will be to-morrow. Law is defined to be a rule of action; but how can that be a rule, which is little known, and less fixed?[285]

This instability might unsettle some judges and make some more hesitant to defer the legal interpretations of an ever more volatile Executive Branch. It would seem to invoke Justice Scalia’s warning that statutes should not be rendered chameleons.[286] These are reasons to either reject Chevron deference in the context of asylum law or to at least limit the range of deference afforded to the Executive Branch.

VIII. Is Arbitrary-and-Capricious Review a Limitation on Chevron?

In 2020, the D.C. Circuit seemed to respond to concerns about instability in asylum law. In Grace v. Barr, plaintiffs challenged the application of Matter of A-B- in credible fear interviews, which determine whether asylum-seekers at the border will be permitted to access the full asylum adjudication system.[287] The D.C. Circuit found that Chevron applies easily to the refugee definition because its key concept of “persecution” is not clearly defined in the statutory text.[288] But while Grace found that Matter of A-B- offered an interpretation of the statute that was permissible under Chevron, it nevertheless failed arbitrary-and-capricious review.[289] According to the D.C. Circuit, Attorney General Sessions substantially departed from the longstanding understanding of asylum eligibility for people who fled persecution by nonstate actors.[290] Previously in such cases, an asylum-seeker could win, in part, by showing merely that her government was “unable or unwilling” to protect her, but under the new interpretation she would have to show her government “condoned the private actions ‘or at least demonstrated a complete helplessness to protect the victims.’”[291] The D.C. Circuit found this was a significant change, made without sufficient explanation. Thus, while the Attorney General’s interpretation was not unreasonable for Chevron purposes, it was an arbitrary-and-capricious change in policy direction.[292]

Grace suggests that arbitrary-and-capricious review may be a significant check on the Executive over and above Chevron deference. This possibility is intriguing and not entirely new. The potential for such an analysis is strongly suggested by the Supreme Court’s decision in Encino Motorcars v. Navarro.[293] In that case, the Department of Labor abandoned a “decades-old” interpretive approach to the regulation of car dealerships and gave “little explanation” for the change.[294] The Court found this failure to “provide a reasoned explanation for the change” rendered the new rule arbitrary.[295] The Encino Motorcars decision by Justice Kennedy says that “Chevron deference is not warranted where the regulation is procedurally defective,” in this case because of the failure to give reasons.[296] Yet the Supreme Court in Encino Motorcars never says clearly if the Department of Labor’s rule could have survived under Chevron but for the failure to give reasons. As a result, the precise interplay between arbitrariness review and Chevron is left somewhat unclear.[297] At least one author has argued that there is nothing of importance in Encino Motorcars since an arbitrary decision is per se unlawful, which makes Chevron irrelevant.[298] But the D.C. Circuit’s decision in Grace suggests something more significant. In Grace, the court struck down a change in interpretation as arbitrary after squarely holding that it could pass Chevron. Grace is thus clearer than Encino Motorcars that arbitrary-and-capricious review is fully independent from Chevron and, in some contexts, more difficult for the branch to navigate.

The proposition that arbitrary-and-capricious review meaningfully limits Chevron deference seems to depend on two possible features of arbitrary-and-capricious review. First, it would need to be far less deferential than often assumed. This assumption seems to be consistent with Chief Justice Roberts’s vigorous application of arbitrary-and-capricious review in the census and DACA (Deferred Action for Childhood Arrivals), both high-profile, Trump-era cases that were entangled with intense immigration-related political struggles.[299] The robust form of arbitrary-and-capricious review evident in these cases has been seen as a departure from the norm.[300] Second, these cases suggest that arbitrary-and-capricious review may be interested in different concerns than Chevron deference. Rather than simply being a different rung on the ladder of higher or lower forms of deference, this robust form of arbitrary-and-capricious review is suspicious of sudden changes in policy. Grace, like the census and DACA cases, involved a significant change in past policy and practice. Thus, while Chevron asks only if a statutory interpretation is reasonable when taken in isolation, arbitrary‑and-capricious review may demand compelling rationales whenever an administration changes policy direction. If this approach holds, arbitrary-and-capricious review might become an important, small-c conservative constraint on Executive power, pushing toward stability and against the turbulence of highly politicized policymaking.

The D.C. Circuit appears, for now, to be standing alone, but only to a degree. Other circuits have been reluctant to strike down Matter of A-B- under arbitrary-and-capricious review.[301] But in so doing, they, too, indirectly imposed a form of constraint on the Executive Branch’s latitude to change policy too quickly. The D.C. Circuit found a problem under arbitrary-and-capricious review because it saw Matter of A-B- as a major change.[302] Courts that have reached the opposite conclusion have tended to minimize the impact of Matter of A-B-, and thus reasoned that it does not represent an arbitrary change in policy. For instance, the Fifth Circuit said, “[B]ecause A-B- did not change any policy relating to asylum and withholding of removal claims, we reject [the] argument that A-B- constituted an arbitrary and capricious change in policy.”[303] Through this approach, arbitrary‑and‑capricious review seems to push courts toward minimizing the impact of new administrative interpretations of the law, even when the government prevails.

IX. Conclusion

The stakes are incredibly high in asylum cases; claimants are typically at risk for physical harm of the gravest kind. Circuit courts have signaled that they might not always affirm limitations on asylum eligibility were they not required to apply deference. And while the Supreme Court decides a few of these cases, it has seemed much more comfortable invoking Chevron deference in asylum cases than in other kinds of immigration cases, especially compared to cases concerning grounds of removal and detention.

There are compelling reasons to see asylum cases as more amenable to deference than other legal questions that determine whether a person will be deported. This focuses on the nexus to foreign affairs and the important role asylum policy plays in migration policy generally. Certainly, in a democracy there is value in giving voters a greater say in policy choices. If political accountability is indeed a good reason for courts to defer, then politicization of asylum adjudication is not really a problem. But that seems far too simplistic. Courts are thus likely to be torn between their inclination to allow space for policymaking and a well-established commitment to neutral adjudication. There are statutory and historical reasons to think that politics is actually not meant to play a major role in asylum cases.[304] The central problem is that political accountability and neutral adjudication are both potentially positive virtues in an administrative system, but it is difficult to bring them both to bear at the same time.

It might be thought that deferring to the political branches is preferred when a question has become more politically fraught. After all, electoral politics is our primary mechanism for resolving such policy disagreements. If this argument holds, then deference makes sense. At the same time, there is a vast difference between discretionary choices over policy direction and the interpretation of a statutory text. When an interpretive choice becomes more politically divisive, letting the Executive Branch interpret a statute will render the meaning of the text as unstable as the electoral college. The meaning given to a statute would shift dramatically, not because Congress enacted a new law but simply because a new Attorney General interpreted an old statute differently than the last one. That instability about the meaning of a statutory text would not serve the rule of law well. While asylum policy is certainly a hot political issue, the refugee definition is a pure question of law. While Congress has indeed delegated legal interpretation of immigration law to the Attorney General overall, the history of the U.S. asylum system has been to increasingly insulate the process from politics as much as possible and to disavow foreign policy concerns as an influence on the legal interpretation of asylum eligibility. Politicizing this process, and then deferring to political choices, would insert considerable instability into a high-stakes area of law and would make resolution of a question of law depend ultimately on who won the last presidential election, not on legal analysis.

Courts have many ways to defuse this dilemma. The Supreme Court could certainly disavow Chevron entirely or in any case touching on asylum. But that would be the most far-reaching approach. Courts can also continue to recognize Chevron’s role in asylum cases, while at the same time applying their own interpretive analysis to the refugee definition at Chevron Step One so as to limit the space for the Executive Branch to change asylum policy abruptly. But courts have been willing to grant considerable deference to the Attorney General on asylum law in the past. If that continues, then the scope of American asylum law may shift radically from one president to the next, even if the statutory text remains exactly the same.


  1. See discussion infra Part IV.

  2. See Michael Kagan, Chevron’s Liberty Exception, 104 Iowa L. Rev. 491, 512, 518, 524–25, 535 (2019).

  3. Compare Brief of Amicus Curiae Senator Sheldon Whitehouse in Support of Respondent at 2, 19, Kisor v. Wilkie, 139 S. Ct. 2400 (2019) (No. 18-15) (argument from a Democratic senator that attempts to undermine deference to administrative agencies are “part of a larger strategy to disable public interest regulation”), with Brief for Amici Curiae The National Immigrant Justice Center and the American Immigration Lawyers Association in Support of Petitioner at 1–2, 26–30, Kisor v. Wilkie, 139 S. Ct. 2400 (2019) (No. 18-15) [hereinafter Brief of NIJC/AILA] (argument from immigrant advocacy organizations that the Supreme Court is right to reconsider aspects of deference to agencies and that deference poses heightened problems in immigration cases).

  4. See Brief of NIJC/AILA, supra note 3, at 20, 26–30 (arguing against Auer deference).

  5. 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 [https://perma.cc/2QCN-R5DX]; Dianne Feinstein, Opinion, Why I’m Voting ‘No’ on Brett Kavanaugh’s Supreme Court Nomination, L.A. Times (Sept. 16, 2018, 4:15 AM), https://www.latimes.com/opinion/op-ed/la-oe-feinstein-kavanaugh-hearings-20180916-story.html [https://perma.cc/56TF-AFD7].

  6. See, e.g., Michael McConnell, Kavanaugh and the “Chevron Doctrine, Hoover Inst. (July 30, 2018), https://www.hoover.org/research/kavanaugh-and-chevron-doctrine [https://perma.cc/PM5D-TLAF]; Joshua A. Geltzer, Opinion, Trump’s Supreme Court Might Overturn a Doctrine, but That Won’t Destroy the 'Administrative State, L.A. Times (Aug. 5, 2018, 4:05 AM), https://www.latimes.com/opinion/op-ed/la-oe-geltzer-kavanaugh-administrative-state-20180805-story.html [https://perma.cc/6N6P-XL5Z]; Henning, supra note 5; 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 [https://perma.cc/533C-7G8S].

  7. See Kagan, supra note 2, at 524–25, 534–35; Michael Kagan, Chevron Goes Missing in an Immigration Case. Again., Yale J. on Regul.: Notice & Comment (Mar. 19, 2019) [hereinafter Kagan, Chevron Goes Missing], https://www.yalejreg.com/nc/chevron-goes-missing-in-an-immigration-case-again/ [https://perma.cc/N7FH-ALP7].

  8. Kagan, supra note 2, at 532; Kagan, Chevron Goes Missing, supra note 7.

  9. Peter L. Markowitz, Deportation Is Different, 13 U. Pa. J. Const. L. 1299, 1332 (2011); e.g., Patrick Glen, Response to Walker on Chevron Deference and Mellouli v. Lynch, Yale J. on Regul.: Notice & Comment (June 10, 2015), https://www.yalejreg.com/nc/response-to-walker-on-chevron-deference-and-mellouli-v-lynch-by-patrick-glen/ [https://perma.cc/6U2U-G4RE] (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 Regul.: Notice & Comment (June 10, 2016), https://www.yalejreg.com/nc/chevron-s-immigration-exception-revisited-by-michael-kagan/ [https://perma.cc/F2MY-9CK5] (discussing the role of Chevron in immigration cases and advocating for a “deportation is different” theory); Christopher J. Walker, The “Scant Sense” Exception to Chevron Deference in Mellouli v. Lynch, Yale J. on Regul.: Notice & Comment (June 2, 2015), https://www.yalejreg.com/nc/the-scant-sense-exception-to-chevron-deference-in-mellouli-v-lynch-by-chris-walker/ [https://perma.cc/932E-BKZL] (discussing the possibility that the Roberts Court may be reluctant to give deference in certain deportation cases).

  10. See Kagan, supra note 2, at 520, 537–39 (showing the Supreme Court has been consistently more willing to apply Chevron in relief from removal cases, especially asylum); The Difference Between Asylum and Withholding of Removal, Am. Immigr. Council (Oct. 6, 2020), https://www.americanimmigrationcouncil.org/research/asylum-withholding-of-removal [https://perma.cc/G6BF-KJSR]. But see Pereira v. Sessions, 138 S. Ct. 2105, 2121, 2129 (2018) (Alito, J., dissenting) (complaining that the Court had ignored Chevron in an immigration case that did not concern deportation or detention).

  11. See discussion infra Part III.

  12. INS. v. Cardoza-Fonseca, 480 U.S. 421, 425, 488 (1987); see Kagan, supra note 2, at 517.

  13. See Deborah E. Anker, Law of Asylum in the United States § 5:1 (2020 ed.) (“The refugee definition requires that persecution or fear of persecution be ‘for reasons of’ or, under U.S. law, ‘on account of’ one of the iterated grounds: race, religion, nationality, membership in a particular social group, or political opinion.”).

  14. See, e.g., Reyes v. Lynch, 842 F.3d 1125, 1129, 1133 (9th Cir. 2016) (affirming the BIA’s “particularity” and “social distinction” requirements for membership in a “particular social group” under Chevron); S.E.R.L. v. Att’y Gen. U.S., 894 F.3d 535, 549–50 (3d Cir. 2018).

  15. See, e.g., S.E.R.L., 894 F.3d at 550 (noting that critiques of the BIA interpretation “raise legitimate concerns . . . [but] notwithstanding our concerns, we conclude that the requirements are reasonable and warrant Chevron deference”).

  16. Contra Michael Herz, Chevron Is Dead, Long Live Chevron, 115 Colum. L. Rev. 1867, 1867 (2015) (“Despite all the attention, . . . the ‘Chevron revolution’ never quite happens.”).

  17. Shoba Sivaprasad Wadhia & Christopher J. Walker, The Case Against Chevron Deference in Immigration Adjudication, 70 Duke L.J. 1197, 1223–24 (2021).

  18. Id. at 1201, 1217–23. Wadhia and Walker take inspiration for their argument from a broader critique of Chevron deference in administrative adjudication offered by Kristin E. Hickman and Aaron L. Nielson. Id. at 1201; see also Kristin E. Hickman & Aaron L. Nielson, Narrowing Chevron’s Domain, 70 Duke L.J. 931, 964, 981 (2021).

  19. Wadhia & Walker, supra note 17, at 1234, 1239–40.

  20. See Maureen A. Sweeney, Enforcing/Protection: The Danger of Chevron in Refugee Act Cases, 71 Admin. L. Rev. 127, 130 (2019) (“This question of deference can mean the difference between lifesaving protection and deportation back to danger.”).

  21. See Katie Benner & Caitlin Dickerson, Sessions Says Domestic and Gang Violence Are Not Grounds for Asylum, N.Y. Times (June 11, 2018), https://www.nytimes.com/2018/06/11/us/politics/sessions-domestic-violence-asylum.html [https://perma.cc/87X4-ULBV]. See generally Fatma Marouf, Becoming Unconventional: Constricting the ‘Particular Social Group’ Ground for Asylum, 44 N.C. J. Int’l L. 489 (2019).

  22. See, e.g., S.E.R.L. v. Att’y Gen. U.S., 894 F.3d 535, 549 (3d Cir. 2018).

  23. Matter of L-E-A- (L-E-A- II), 27 I. & N. Dec. 581, 596 (Att’y Gen. 2019).

  24. Id. at 589 (“I recognize that a number of courts of appeals have issued opinions that recognize a family-based social group as a ‘particular social group’ under the asylum statute.”).

  25. Id. at 592.

  26. Procedures for Asylum and Withholding of Removal; Credible Fear and Reasonable Fear Review, 65 Fed. Reg. 80274 (Dec. 11, 2020) (to be codified at 8 C.F.R. pts. 1003, 1208, 1235).

  27. See infra Part VIII.

  28. Id.

  29. E.g., Dep’t of Homeland Sec. v. Regents of Univ. of Cal., 140 S. Ct. 1891, 1912–13 (2020).

  30. See Kagan, supra note 2, at 532.

  31. Cf. Scialabba v. Cuellar de Osorio, 573 U.S. 41, 55–56, 75 (2014) (applying Chevron deference in the denial of a family-sponsored visa application).

  32. Kadia v. Gonzales, 501 F.3d 817, 819, 821 (7th Cir. 2007).

  33. 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.”).

  34. Id. at 842–43.

  35. See United States v. Mead Corp., 533 U.S. 218, 226–27, 229 (2001) (“We hold that administrative implementation of a particular statutory provision qualifies for Chevron deference when it appears that Congress delegated authority to the agency generally to make rules carrying the force of law . . . . Congress . . . may not have expressly delegated authority or responsibility to implement a particular provision or fill a particular gap.”).

  36. See Chevron, 467 U.S. at 865–66; Elena Kagan, Presidential Administration, 114 Harv. L. Rev. 2245, 2326–28, 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).

  37. See Chevron, 467 U.S. at 865 (“Judges are not experts in the field . . . .”). But see Kagan, supra note 2, at 501–02 (summarizing critiques of the technical expertise justification for deference).

  38. 5 U.S.C. § 706.

  39. Antonin Scalia, Judicial Deference to Administrative Interpretations of Law, 1989 Duke L.J. 511, 512–13 (1989).

  40. See Michigan v. EPA, 135 S. Ct. 2699, 2712 (2015) (Thomas, J., concurring); Gutierrez-Brizuela v. Lynch, 834 F.3d 1142, 1149, 1152 (10th Cir. 2016) (Gorsuch, J., concurring) (“Chevron seems no less than a judge-made doctrine for the abdication of the judicial duty.”).

  41. See Kent Barnett & Christopher J. Walker, Chevron in the Circuit Courts, 116 Mich. L. Rev. 1, 19 (2017); 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); Herz, supra note 16, at 1870.

  42. Pereira v. Sessions, 138 S. Ct. 2105, 2120–21 (2018) (Kennedy, J., concurring).

  43. See generally Hiroshi Motomura, Immigration Law After a Century of Plenary Power: Phantom Constitutional Norms and Statutory Interpretation, 100 Yale L.J. 545, 551–53 (1990).

  44. 8 U.S.C. § 1103(a)(1).

  45. See Justin Chasco, Comment, Judge Alberto Gonzales? The Attorney General’s Power to Overturn Board of Immigration Appeals’ Decisions, 31 S. Ill. U. L.J. 363, 375 (2007).

  46. INS. v. Aguirre-Aguirre, 526 U.S. 415, 419, 424 (1999) (“Under the immigration laws, withholding is distinct from asylum, although the two forms of relief serve similar purposes.”).

  47. Scialabba v. Cuellar de Osorio, 573 U.S. 41, 56–57 (2014) (plurality opinion) (citations omitted) (quoting Aguirre-Aguirre, 526 U.S. at 425).

  48. See Kagan, supra note 2, at 520–21 tbl.1.

  49. Id. at 519–21.

  50. Id. at 524.

  51. Id. at 525–26.

  52. Pereira v. Sessions, 138 S. Ct. 2105, 2121, 2129 (2018) (Alito, J., dissenting).

  53. Eskridge, Jr. & Baer, supra note 41, at 1124–25; see also 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, 1765–66 (2010).

  54. Raso & Eskridge, Jr., supra note 53, at 1735.

  55. Id. at 1766.

  56. See King v. Burwell, 135 S. Ct. 2480, 2488 (2015) (Chevron deference is a method that the court says it “often” applies).

  57. Natalie Salmanowitz & Holger Spamann, Does the Supreme Court Really Not Apply Chevron When It Should?, 57 Int’l Rev. L. & Econ. 81, 83–85 (2019).

  58. Id. at 81, 89 (“Our reexamination of this study finds that the fraction of such cases is far lower, and indeed closer to zero.”).

  59. Id. at 83.

  60. Id. at 85–86, 89.

  61. See Kagan, supra note 2, at 524–25, 527–28 tbl.2.

  62. Id. at 524–25 (footnotes omitted).

  63. Raso & Eskridge, Jr., supra note 53, at 1766. (“Idiosyncrasy in deployment (or not) of deference regimes is tolerated within the Court.”).

  64. See Michael Kagan, Loud and Soft Anti-Chevron Decisions, 53 Wake Forest L. Rev. 37, 40 (2018).

  65. See Kent Barnett, Why Bias Challenges to Administrative Adjudication Should Succeed, 81 Mo. L. Rev. 1023, 1036–37 (2016).

  66. Kagan, supra note 64, at 50 (noting factors such as: “(1) whether the Supreme Court itself acknowledged lack of statutory clarity; (2) whether lower court judges were divided on the statutory meaning, providing an objective indication that the statute’s meaning was subject to reasonable disagreement; (3) whether lower courts disagreed with the agency’s interpretation, similarly indicating room for reasonable disagreement; (4) whether the lower court decision under review applied Chevron; and (5) whether the government asked for deference to the agency’s interpretation”).

  67. Id. at 47–48.

  68. Id.

  69. King v. Burwell, 135 S. Ct. 2480, 2488–89 (2015).

  70. Kagan, supra note 64, at 40, 48.

  71. Id. at 49.

  72. Id.

  73. I am indebted to David Rubenstein for this insight.

  74. Freud might not have actually said “sometimes a cigar is just a cigar,” though the line is typically attributed to him. See Sometimes a Cigar Is Just a Cigar, Quote Investigator, https://quoteinvestigator.com/2011/08/12/just-a-cigar/ [https://perma.cc/4NJR-ZKG3] (last updated Aug. 12, 2011).

  75. See Kagan, supra note 64, at 54–55.

  76. See Kagan, supra note 2, at 533–35, 537–39.

  77. See, e.g., Torres v. Lynch, 136 S. Ct. 1619, 1623 (2016).

  78. See Esquivel-Quintana v. Lynch, 810 F.3d 1019, 1027–28 (6th Cir. 2016) (Sutton, J., concurring in part and dissenting in part) (proposing a dual-application argument for avoiding Chevron deference), rev’d sub nom. Esquivel-Quintana v. Sessions, 137 S. Ct. 1562 (2017).

  79. See 8 U.S.C. § 1101(a)(43) (defining an aggravated felony); id. § 1227(a)(2)(iii) (aggravated felony as a ground for removal).

  80. Id. § 1326(b)(2); see also id. § 1327.

  81. See Esquivel-Quintana, 810 F.3d at 1027–28 (Sutton, J., concurring in part and dissenting in part).

  82. See Kagan, supra note 2, at 529–30.

  83. Mellouli v. Lynch, 135 S. Ct. 1980, 1983–84 (2015).

  84. Id. at 1989.

  85. Zadvydas v. Davis, 533 U.S. 678, 682, 689, 700–01 (2001).

  86. Clark v. Martinez, 543 U.S. 371, 402 (2005) (Thomas, J., dissenting).

  87. Jennings v. Rodriguez, 138 S. Ct. 830, 836 (2018).

  88. See Nielsen v. Preap, 139 S. Ct. 954 (2019) (making no mention of Chevron deference in a case concerning immigrant detention where the parties had argued extensively about Chevron’s application), rev’g Preap v. Johnson, 831 F.3d 1193, 1199 (9th Cir. 2016).

  89. See Kagan, supra note 2, at 532.

  90. See, e.g., Scialabba v. Cuellar de Osorio, 573 U.S. 41, 55–56 (2014).

  91. See, e.g., Judulang v. Holder, 565 U.S. 42, 52 n.7, 53 (2011); Holder v. Martinez Gutierrez, 566 U.S. 583, 586, 591 (2012).

  92. INS v. Cardoza-Fonseca, 480 U.S. 421, 446–48 (1987); see also Thomas W. Merrill, Judicial Deference to Executive Precedent, 101 Yale L.J. 969, 985 (1992).

  93. INS v. Elias-Zacarias, 502 U.S. 478, 482 (1992); Brief for Petitioner, INS. v. Elias‑Zacarias, 502 U.S. 478 (1992) (No. 90-1342), 1991 WL 11003946, at *23.

  94. INS v. Aguirre-Aguirre, 526 U.S. 415, 423–25 (1999).

  95. Negusie v. Holder, 555 U.S. 511, 517 (2009).

  96. 8 U.S.C. § 1158(b)(1); id. § 1101(a)(42)(A).

  97. See generally Nat’l Immigrant Just. Ctr., A Timeline of the Trump Administration’s Efforts to End Asylum (2019) https://immigrantjustice.org/sites/default/files/uploaded-files/no-content-type/2019-08/Asylum_Timeline_August2019.pdf [https://perma.cc/5TJA-3P5S].

  98. Id. at 4.

  99. Id. at 3.

  100. Id. at 6.

  101. See infra Part IV.

  102. 8 U.S.C. § 1101(a)(42)(A); see also Anker, supra note 13, § 5:1.

  103. See, e.g., Matter of A-E-M-, 21 I. & N. Dec. 1157, 1159–60 (B.I.A. 1998) (noting that in a political asylum case, the level of risk may be assessed to be less when family members are not targeted).

  104. See Michael Kagan, Refugee Credibility Assessment and the “Religious Imposter” Problem, 43 Vand. J. Transnat’l L. 1179, 1182 & n.5 (2010).

  105. See Anker, supra note 13, § 5:40.

  106. See id.

  107. Perez-Zenteno v. U.S. Att’y Gen., 913 F.3d 1301, 1304 (11th Cir. 2019).

  108. See Marouf, supra note 21, at 489–90 (tracing the evolution of BIA jurisprudence).

  109. Id. at 489.

  110. See Matter of Acosta, 19 I. & N. Dec. 211, 233 (B.I.A. 1985).

  111. See discussion infra Section V.D.

  112. See Marouf, supra note 21, at 490.

  113. See Anker, supra note 13, § 5:43.

  114. Marouf, supra note 21, at 490–91.

  115. Matter of A-B-, 27 I. & N. Dec. 316, 316 (Att’y Gen. 2018), overruling Matter of A-R-C-G-, 26 I. & N. Dec. 338 (B.I.A. 2014).

  116. Id. at 330–31, 333.

  117. Marouf, supra note 21, at 492.

  118. Id. at 513–14.

  119. Matter of Acosta, 19 I. & N. Dec. 211, 222 (B.I.A. 1985).

  120. Matter of A-B-, 27 I. & N. Dec. 316, 337 (Att’y Gen. 2018) (quoting Galina v. INS, 213 F.3d 955, 958 (7th Cir. 2000)).

  121. See discussion infra Part VIII.

  122. See Matter of A-B-, 27 I. & N. Dec. at 343 (“No country provides its citizens with complete security from private criminal activity, and perfect protection is not required.”).

  123. See, e.g., Richard Gonzales, Sessions Says ‘Zero Tolerance’ for Illegal Border Crossers, Vows to Divide Families, NPR (May 7, 2018, 8:17 PM), https://www.npr.org/sections/thetwo-way/2018/05/07/609225537/sessions-says-zero-tolerance-for-illegal-border-crossers-vows-to-divide-families [https://perma.cc/UA9V-KSVA].

  124. Matter of A-B-, 27 I. & N. Dec. at 345.

  125. See, e.g., Dara Lind, Trump Wants to Make Asylum Seekers’ Stay in the US Harder—and Shorter, Vox (Apr. 30, 2019, 12:50 PM), https://www.vox.com/2019/4/30/18523990/trump-asylum-border-new [https://perma.cc/EGF7-HLCN].

  126. See, e.g., The White House, Our Nation’s Weak Asylum Laws Are Encouraging an Overwhelming Increase in Illegal Immigration (Nov. 1, 2018), https://trumpwhitehouse.archives.gov/briefings-statements/nations-weak-asylum-laws-encouraging-overwhelming-increase-illegal-immigration/ [https://perma.cc/7WZU-QCRG] (citing low final grant rates in asylum claims as a reason to reform asylum policy).

  127. See generally Barnett, supra note 65, at 1023 (arguing that “challenges to adjudicators’ appearance of partiality are well positioned to be part of the new wave of structural challenges to the administrative state”).

  128. Matter of L-E-A- (L-E-A- II), 27 I. & N. Dec. 581, 589 (Att’y Gen. 2019).

  129. Matter of Acosta, 19 I. & N. Dec. 211, 233 (B.I.A. 1985).

  130. Matter of L-E-A- (L-E-A- I), 27 I. & N. Dec. 40, 43 (B.I.A. 2017), rev’d, 27 I. & N. Dec. 581 (Att’y Gen. 2019).

  131. Id.

  132. L-E-A- II, 27 I. & N. Dec. at 589.

  133. Id. (“I also recognize that certain courts of appeals have considered the requisite elements of a ‘particular social group’ and . . . have nonetheless suggested that shared family ties alone are sufficient to satisfy the INA’s definition of ‘refugee’ . . . .”).

  134. Thomas v. Gonzales, 409 F.3d 1177, 1188–89 (9th Cir. 2005), vacated, 547 U.S. 183 (2006).

  135. L-E-A- II, 27 I. & N. Dec. at 590.

  136. Id. at 591 (providing that “determination and ruling by the Attorney General with respect to all questions of law shall be controlling” under the INA (citing 8 U.S.C. § 1103(a)(1))).

  137. Id. at 591–92 (“Congress thus delegated to the Attorney General the discretion to reasonably interpret the meaning of ‘membership in a particular social group,’ and such reasonable interpretations are entitled to deference.” (citing Chevron U.S.A., Inc. v. Nat. Res. Def. Council, Inc., 467 U.S. 837, 844 (1984))).

  138. Id. at 592 (citing Nat’l Cable & Telecomms. Ass’n v. Brand X Internet Servs., 545 U.S. 967, 980 (2005)).

  139. Cf. Perez-Sanchez v. Att’y Gen., 935 F.3d 1148, 1158 n.7 (11th Cir. 2019) (declining to review the validity of Matter of L-E-A- on the facts of the case presented).

  140. Pena Oseguera v. Barr, 936 F.3d 249, 251 (5th Cir. 2019).

  141. See Procedures for Asylum and Withholding of Removal; Credible Fear and Reasonable Fear Review, 85 Fed. Reg. 80274, 80280–81 (Dec. 11, 2020) (to be codified at 8 C.F.R. pts. 1003, 1208, 1235).

  142. Pangea Legal Servs. v. U.S. Dep’t of Homeland Sec., No. 20-CV-09253, 2021 WL 75756, at *2–6 (N.D. Cal. Jan. 8, 2021).

  143. S.E.R.L. v. Att’y Gen. U.S., 894 F.3d 535 (3d Cir. 2018).

  144. Id. at 540.

  145. Id. at 541.

  146. Id.

  147. Id. at 541–42.

  148. Id. at 544 (quoting Valdiviezo-Galdamez v. Att’y Gen., 663 F.3d 582, 595 (3d Cir. 2011) (quoting Matter of Acosta, 19 I. & N. Dec. 211, 233–34 (B.I.A. 1985))).

  149. Id. at 545–46.

  150. Id. at 546 & n.12 (citing Valdiviezo-Galdamez, 663 F.3d at 603–09) (summarizing national litigation about the BIA’s interpretation during this period).

  151. Id. at 546–47.

  152. Matter of M-E-V-G-, 26 I. & N. Dec. 227, 237 (B.I.A. 2014); Matter of W-G-R-, 26 I. & N. Dec. 208, 212 (B.I.A. 2014).

  153. Matter of M-E-V-G-, I. & N. Dec. at 237.

  154. Id. at 239.

  155. Id. at 240–41.

  156. S.E.R.L., 894 F.3d at 549.

  157. Id. at 555.

  158. Id. at 549–50.

  159. Id. at 550.

  160. Id. at 549–55.

  161. Id. at 550.

  162. See Reyes v. Lynch, 842 F.3d 1125, 1135 (9th Cir. 2016).

  163. See, e.g., id. at 1135–36 (“The BIA’s statement of the purpose and function of the ‘particularity’ requirement does not, on its face, impose a numerical limit on a proposed social group . . . . Nor is it contrary to the principle that diversity within a proposed particular social group may not serve as the sine qua non of the particularity analysis. . . . The ‘social distinction’ requirement is not, as Garay contends, a ‘new’ requirement.” (citation omitted)).

  164. See Melnik v. Sessions, 891 F.3d 278, 286 n.22 (7th Cir. 2018).

  165. See Barnett & Walker, supra note 41, at 17–18.

  166. See, e.g., Cinto–Velasquez v. Lynch, 817 F.3d 602, 606 (8th Cir. 2016) (“[W]e give Chevron deference to the BIA’s reasonable interpretation of this ambiguous statutory phrase.”); De Guevara v. Barr, 919 F.3d 538, 540 (8th Cir. 2019) (quoting the same).

  167. See, e.g., Muiruri v. Lynch, 803 F.3d 984, 986 (8th Cir. 2015) (“[T]he BIA’s interpretation of immigration laws and regulations receives substantial deference.”) (citing Habchy v. Gonzales, 471 F.3d 858, 862 (8th Cir. 2006)); Bernal–Rendon v. Gonzales, 419 F.3d 877, 880 (8th Cir 2005) (holding the court would “accord substantial deference to the BIA’s interpretation of immigration law and agency regulations”) (first citing Tang v. INS, 223 F.3d 713, 718–19 (8th Cir. 2000); and then citing Ikenokwalu–White v. INS, 316 F.3d 798, 804 (8th Cir. 2003)).

  168. See Anker, supra note 13, § 5:43.

  169. Perez-Zenteno v. U.S. Att’y Gen., 913 F.3d 1301, 1308–10 (11th Cir. 2019).

  170. See, e.g., Del Carmen Amaya-De Sicaran v. Barr, 979 F.3d 210, 213–14, 216 (4th Cir. 2020); Diaz-Reynoso v. Barr, 968 F.3d 1070, 1079–80 (9th Cir. 2020); Gonzales-Veliz v. Barr, 938 F.3d 219, 235 (5th Cir. 2019).

  171. Gonzales-Veliz, 938 F.3d at 232–34; see also De Pena-Paniagua v. Barr, 957 F.3d 88, 93 (1st Cir. 2020) (affirming that Matter of A-B- does not categorically change asylum law).

  172. Diaz-Reynoso, 968 F.3d at 1079–80 (“We recognize that the Attorney General began the opinion in Matter of A-B- by offering some general impressions about asylum and withholding claims . . . . But the holding of Matter of A-B- plainly does not endorse any sort of categorical exception . . . .”).

  173. S.E.R.L. v. Att’y Gen. U.S., 894 F.3d 535, 554–55 (3d Cir. 2018) (citation omitted).

  174. See Negusie v. Holder, 555 U.S. 511, 516 (2009) (“[T]he BIA is entitled to deference in interpreting ambiguous provisions of the INA.”).

  175. Kagan, supra note 2, at 532.

  176. See Gutierrez-Brizuela v. Lynch, 834 F.3d 1142, 1154 (10th Cir. 2016) (Gorsuch, J., concurring) (“Not only is Chevron’s purpose seemingly at odds with the separation of legislative and executive functions, its effect appears to be as well.”).

  177. Sweeney, supra note 20, at 179–87.

  178. INS v. Cardoza-Fonseca, 480 U.S. 421, 437–38 (1987); United Nations Convention Relating to the Status of Refugees, July 28, 1951, 189 U.N.T.S. 150.

  179. Sweeney, supra note 20, at 184, 186, 188; see also Murray v. Schooner Charming Betsy (Charming Betsy), 6 U.S. 64, 118 (1804).

  180. Cardoza-Fonseca, 480 U.S. at 448.

  181. The Shifting Burdens of Immigration Law, 8 Immigr. L. Advisor (U.S. Dep’t Just. Exec. Off. for Immigr. Rev.), Oct. 2014, at 2–3.

  182. 8. U.S.C. § 1229a(c)(7)(C)(ii).

  183. See id. § 1229a(c)(4)(B); Woodby v. INS, 385 U.S. 276, 277, 282, 284–85 (1966) (“[I]t is incumbent upon the Government in such proceedings to establish the facts supporting deportability by clear, unequivocal, and convincing evidence.”).

  184. See 8 U.S.C. § 1252(a)(2)(B).

  185. Pereira v. Sessions, 138 S. Ct. 2105, 2110 (2018).

  186. Id. at 2113–14, 2121.

  187. Id. at 2113.

  188. Id. at 2121, 2129 (Alito, J., dissenting).

  189. Id. at 2121 (Kennedy, J., concurring).

  190. Id. at 2113 (majority opinion); see also Kagan, supra note 2, at 524–28.

  191. See, e.g., Paiz–Morales v. Lynch, 795 F.3d 238, 245 (1st Cir. 2015) (rejecting a proposed particular social group “on the record of this case” only, based on a substantial evidence standard); Velasquez v. Sessions, 866 F.3d 188, 194, 196–98 (4th Cir. 2017) (affirming under substantial evidence for lack of nexus to a particular social group); Cruz-Guzman v. Barr, 920 F.3d 1033, 1038 (6th Cir. 2019) (concluding that “[t]he record does not compel” finding that the proposed ground caused the fear of persecution); Bernal–Rendon v. Gonzales, 419 F.3d 877, 880–81 (8th Cir. 2005) (“While petitioners correctly contend that a nuclear family can constitute a social group, petitioners fail to prove that a specific threat exists to their family as a social group.”) (citations omitted); cf. Oliva v. Lynch, 807 F.3d 53, 61 (4th Cir. 2015) (remanding for the BIA to consider all evidence about whether the proposed group met requirements).

  192. See Thomas W. Merrill, Article III, Agency Adjudication, and the Origins of the Appellate Review Model of Administrative Law, 111 Colum. L. Rev. 939, 948–50, 954–55, 999–1000, 999 n.293 (2011) (discussing historical origins of deference based on technical expertise). But see Gary Lawson & Stephen Kam, Making Law Out of Nothing at All: The Origins of the Chevron Doctrine, 65 Admin. L. Rev. 1, 11 (2013) (critiquing epistemological expertise as a ground for deference).

  193. Int’l Fabricare Inst. v. EPA, 972 F.2d 384, 389 (D.C. Cir. 1992).

  194. See Antonin Scalia, Judicial Deference to Administrative Interpretations of Law, 1989 Duke L.J. 511, 514–16 (1989); Lawson & Kam, supra note 193, at 71–72 (2013); United States v. Mead Corp., 533 U.S. 218, 250 (2001) (Scalia, J., dissenting).

  195. Kisor v. Wilkie, 139 S. Ct. 2400, 2413 (2019) (Kagan, J., concurring) (quoting Martin v. Occupational Safety & Health Rev. Comm’n, 499 U.S. 144, 151 (1991)).

  196. See Wadhia & Walker, supra note 17, at 1201, 1218–20, 1222–23; Sweeney, supra note 20, at 174–75.

  197. See Kagan, supra note 2, at 513–17 (critiquing standard arguments for deference as applied to the BIA).

  198. 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, 352 (1990).

  199. See Fatma Marouf et al., Justice on the Fly: The Danger of Errant Deportations, 75 Ohio St. L.J. 337, 339 & n.2 (2014).

  200. See David Zaring, Reasonable Agencies, 96 Va. L. Rev. 135, 177, 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); Wadhia & Walker, supra note 17, at 1221 (stating that circuit courts are less deferential in immigration adjudication cases than in other administrative matters).

  201. See generally Kagan, supra note 2, at 530, 532, 538–39, 542.

  202. See, e.g., S.E.R.L. v. Att’y Gen. U.S., 894 F.3d 535, 549 (3d Cir. 2018); Negusie v. Holder, 555 U.S. 511, 517 (2009).

  203. Scialabba v. Cuellar de Osorio, 573 U.S. 41, 55–57 (2014) (citations omitted) (quoting INS v. Aguirre-Aguirre, 526 U.S. 415, 425 (1999)).

  204. See Matthew J. Lindsay, Disaggregating "Immigration Law," 68 Fla. L. Rev. 179, 182–83, 197 (2016); cf. Trump v. Hawaii, 138 S. Ct. 2392, 2419 (2018) (noting that a “narrow standard of review” is appropriate “in admission and immigration cases that overlap with ‘the area of national security’”) (quoting Kerry v. Din, 135 S. Ct. 2128, 2140 (2015)).

  205. See, e.g., Moncrieffe v. Holder, 569 U.S. 184, 187–90, 206 (2013); Mellouli v. Lynch, 135 S. Ct. 1980, 1983–84, 1990–91 (2015); Esquivel-Quintana v. Sessions, 137 S. Ct. 1562, 1567, 1572–73 (2017).

  206. U.S. Dep’t Just. Exec. Off. for Immigr. Rev., Executive Office for Immigration Review: An Agency Guide 2–4 (2017), https://www.justice.gov/eoir/page/file/eoir_an_agency_guide/download [https://perma.cc/WKG6-GBF2].

  207. See 8 U.S.C. § 1229b(a).

  208. See id. § 1229b(b)(1).

  209. See, e.g., In re Andazola-Rivas, 23 I. & N. Dec. 319, 323–24 (B.I.A. 2002) (considering whether educational opportunities for children can cancel removal); In re Monreal-Aguinaga, 23 I. & N. Dec. 56, 63 (B.I.A. 2001) (stating that caring for elderly parents or “a qualifying child with very serious health issues, or compelling special needs in school” may qualify for removal cancellation).

  210. See 8 U.S.C. § 1254a(b)(1).

  211. See id. § 1254a(a)(1), (c)(1)(A).

  212. Trump v. Hawaii, 138 S. Ct. 2392, 2415, 2421 (2018).

  213. See Note, American Courts and the U.N. High Commissioner for Refugees: A Need for Harmony in the Face of a Refugee Crisis, 131 Harv. L. Rev. 1399, 1400–02 (2018).

  214. See discussion supra Part V (discussing the Charming Betsy doctrine’s application in Chevron analysis).

  215. See 8 C.F.R. § 1003.1(d)(1), (d)(1)(ii).

  216. 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).

  217. See Hawaii, 138 S. Ct. at 2419–20 (noting the President’s need “to respond to changing world conditions”) (quoting Mathews v. Diaz, 426 U.S. 67, 81 (1976)).

  218. See Robert M. Cannon, Comment, A Reevaluation of the Relationship of the Administrative Procedure Act to Asylum Hearings: The Ramifications of the American Baptist Churches’ Settlement, 5 Admin. L.J. 713, 722–23 (1991) (“The determination of whether an individual fits within the definition of a refugee is meant to be free of considerations of ideology, foreign policy, and geographic origin.”).

  219. See Shane M. Sorenson, Note, Immigration and Naturalization Service v. Cardoza-Fonseca: Two Steps in the Right Direction, 3 Admin. L.J. 95, 99 (1989).

  220. Id. at 104 (“[T]he Refugee Act removed the words ‘in his opinion’ from section 243(h). This eliminated the Attorney General’s discretionary power to withhold deportation. The Act mandates relief if the statutory requirements of that section are satisfied.”).

  221. See Eliot Walker, Asylees in Wonderland: A New Procedural Perspective on America’s Asylum System, 2 Nw. J.L. & Soc. Pol’y 1, 12–14, 16–18 (2007).

  222. Am. Baptist Churches v. Thornburgh, 760 F. Supp. 796, 799 (N.D. Cal. 1991).

  223. See Joan Fitzpatrick & Robert Pauw, Foreign Policy, Asylum and Discretion, 28 Willamette L. Rev. 751, 758 (1992).

  224. See id. at 751–52 (discussing INS v. Doherty, 502 U.S. 314 (1992)).

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

  226. S.E.R.L. v. Att’y Gen. U.S., 894 F.3d 535, 545 (3d Cir. 2018).

  227. Kagan, supra note 36, at 2354.

  228. Id. at 2331–32.

  229. 8 C.F.R. § 1003.1(d)(1)(ii) (“Board members shall exercise their independent judgment and discretion in considering and determining the cases coming before the Board.”); see also Stephen H. Legomsky, Restructuring Immigration Adjudication, 59 Duke L.J. 1635, 1665 (2010) (stating that BIA members are career appointees).

  230. Wadhia & Walker, supra note 17, at 1230,1232.

  231. See Kent Barnett et al., Administrative Law’s Political Dynamics, 71 Vand. L. Rev. 1463, 1480–81 (2018).

  232. See Jaclyn Kelley-Widmer & Hillary Rich, A Step Too Far: Matter of A-B-, “Particular Social Group,” and Chevron, 29 Cornell J.L. & Pub. Pol’y 345, 357–63 (2019) (discussing the precedential history of interpreting the term “particular social group”). See generally Wadhia & Walker, supra note 17, at 1210 & n.77, 1211 & n.84, 1212 & n.92.

  233. Wadhia & Walker, supra note 17, at 1201–02.

  234. See Daniel E. Chand, Protecting Agency Judges in an Age of Politicization: Evaluating Judicial Independence and Decisional Confidence in Administrative Adjudications, 49 Am. Rev. Pub. Admin. 395, 398 (2019) (stating that the controversy over independence of non-APA judges “is most notable in immigration court, overseen by immigration judges (IJs), who make up the largest population of non-APA judges”).

  235. Id. at 395.

  236. Cf. Kent Barnett, Regulating Impartiality in Agency Adjudication, 69 Duke L.J. 1695, 1698–1700 (2020) (arguing that the statutory requirement that administrative law judges have their impartiality shielded by prohibiting “at-will” dismissal is constitutionally problematic, but could be accomplished through regulation or executive order).

  237. See Chand, supra note 235, at 399.

  238. See Adrian Vermeule, Conventions of Agency Independence, 113 Colum. L. Rev. 1163, 1211 (2013).

  239. Catherine Y. Kim, The President’s Immigration Courts, 68 Emory L.J. 1, 35–36 (2018) (noting that Justice Kagan’s oft-cited arguments for presidential control over administration treated adjudication as distinct).

  240. See, e.g., Colmenar v. INS, 210 F.3d 967, 971 (9th Cir. 2000) (“[T]he IJ behaved not as a neutral fact-finder interested in hearing the petitioner’s evidence, but as a partisan adjudicator . . . .”).

  241. Kim, supra note 240, at 36–37; see also Londoner v. City of Denver, 210 U.S. 373, 386 (1908); Bi-Metallic Inv. Co. v. State Bd. of Equalization, 239 U.S. 441, 445 (1915).

  242. Kim, supra note 240, at 36–37.

  243. See, e.g., Shoaira v. Ashcroft, 377 F.3d 837, 842–43 (8th Cir. 2004) (“[N]ear the beginning of the hearing, the IJ said, ‘I am one of those judges that is not the least bit interested with the process. I don’t care about the procedure. I don’t care about the process here. It’s all for show.’ Taken in isolation and coming from a judge, this language is deplorable.” (citation omitted)).

  244. Sanchez–Cruz v. INS, 255 F.3d 775, 779 (9th Cir. 2001) (internal quotations omitted).

  245. Cf. Merrill, supra note 193, at 999, 1002 (arguing that the appellate review model for agency rulemaking may invite too much judicial intervention, on the theory that the adjudication looks much like the judicial trial system).

  246. 8 C.F.R. § 1003.1(g).

  247. See Michael Kagan et al., Invisible Adjudication in the U.S. Courts of Appeals, 106 Geo. L.J. 683, 701 (2018).

  248. Amy Semet, An Empirical Examination of Agency Statutory Interpretation, 103 Minn. L. Rev. 2255, 2260 (2019) (“[W]hen agencies rely on precedent to the exclusion of other tools, agencies may abdicate their responsibility to be democratically accountable by failing to fully consider [policy concerns].”).

  249. See James Durling & E. Garrett West, May Chevron Be Waived?, 71 Stan. L. Rev. Online 183, 193 (2019); Glob. TelLink v. FCC, 866 F.3d 397, 407–08 (D.C. Cir. 2017); Aaron L. Nielson, D.C. Circuit Review – Reviewed: A New Step for Chevron?, Yale J. on Regul.: Notice & Comment (June 16, 2017), https://www.yalejreg.com/nc/d-c-circuit-review-reviewed-a-new-step-for-chevron/ [https://perma.cc/2QF5-GDFU].

  250. See Durling & West, supra note 250, at 191–92.

  251. See Crowell v. Benson, 285 U.S. 22, 50 (1932).

  252. Christopher J. Walker, Constitutional Tensions in Agency Adjudication, 104 Iowa L. Rev. 2679, 2680 (2019).

  253. Id. at 2688.

  254. Id. at 2691–93.

  255. Id. (discussing Stern v. Marshall, 564 U.S. 462, 489 n.6 (2011)); see also Kim, supra note 240, at 41–42 (noting the distinction between adjudicative and legislative facts); Heckler v. Campbell, 461 U.S. 458, 467–68 (1983).

  256. SEC v. Chenery Corp. (Chenery II), 332 U.S. 194, 201–02 (1947).

  257. See Aaron L. Nielson, D.C. Circuit Review – Reviewed: "I Vote for Chenery I, Not Chenery II," Yale J. on Regul.: Notice & Comment (Nov. 24, 2017), https://www.yalejreg.com/nc/d-c-circuit-review-reviewed-i-vote-for-chenery-i-not-chenery-ii/ [https://perma.cc/TV8H-PFVE].

  258. Wadhia & Walker, supra note 17, at 1202, 1224–26.

  259. See id. at 1232; Batterton v. Marshall, 648 F.2d 694, 703 (D.C. Cir. 1980) (stating that notice-and-comment procedures “reintroduce public participation and fairness to affected parties after governmental authority has been delegated to unrepresentative agencies”).