It is settled doctrine that agencies receive Chevron deference1 when resolving statutory ambiguities. But what happens when a statute contains an apparent conflict between two otherwise clear provisions? Should the agency still receive Chevron deference when it chooses to follow one statutory directive and not the other? Last Term, in Scialabba v. Cuellar de Osorio,2 a plurality of the Supreme Court held that it is permissible for an agency to choose between two contradictory statutory directives in the Child Status Protection Act3 (CSPA). The Court’s opinions give insight into fundamental questions of judicial deference to agency interpretations: when a statute completely contradicts itself, as CSPA arguably does, any rationale for agency deference rooted in congressional intent begins to erode.
The Immigration and Nationality Act4 allows U.S. citizens and lawful permanent residents (LPRs) to petition for family members — including spouses, siblings, and children — to obtain visas to enter and reside in the United States.5 These family members are referred to as “principal beneficiaries,”6 and their visa petitions fall into one of five “family preference” categories7: F1, the unmarried adult children of U.S. citizens; F2A, the spouses and unmarried minor children of LPRs; F2B, the unmarried adult children of LPRs; F3, the married children of U.S. citizens; and F4, the siblings of U.S. citizens.8 The number of visas within each family preference category is capped by law,9 and visas are doled out on a first-come, first-served basis.10 Principal beneficiaries’ close family members — namely, spouses and children under the age of twenty-one — are afforded the “same status, and the same order of consideration.”11 This provision is of critical importance: between bureaucratic delays and the time spent waiting for visas to become available, the immigration process can take decades. As a result, children who were under twenty-one at the time the visa petition was approved by the United States Citizenship and Immigration Services (USCIS) may be substantially older by the time visas become available. Before 2002, families with children who had aged out were forced either to stay in their home countries or to leave those children behind.12 In an effort to remedy this issue, Congress passed the CSPA in 2002.13 Section 3 of the CSPA states:
If the age of an alien is determined under paragraph (1) to be 21 years of age or older for the purposes of subsections (a)(2)(A) and (d) of this section, the alien’s petition shall automatically be converted to the appropriate category and the alien shall retain the original priority date issued upon receipt of the original petition.14
This provision is far from limpid. The first half appears to propose broad-based relief by granting relief to all who have been determined to be twenty-one years or older under the statute; the second half, though, arguably appears to provide a remedy that works only for a small subset of the eligible individuals identified in the first half. The crux of the tension is the requirement of automatic conversion: Does the automaticity contemplate a situation in which a new petition could be created with a different principal beneficiary? Or must automatic mean immediate, and without any bureaucratic wrangling?
This was precisely the question confronted by Rosalina Cuellar de Osorio and her son, Melvin. In 1998 Ms. Cuellar de Osorio, a native of El Salvador, became the principal beneficiary of a visa petition filed by her mother, a U.S. citizen.15 Ms. Cuellar de Osorio hoped to enter the United States and become an LPR; she also hoped to bring Melvin with her as a derivative beneficiary of the visa petition.16 By the time they reached the front of the visa line in November 2005, Melvin was no longer under the age of twenty-one.17 The United States consulate told them that, because Melvin had “aged out” of minor derivative beneficiary status, he had to begin the visa application process anew.18
Ms. Cuellar de Osorio, along with other similarly situated parents, filed suit in the United States District Court for the Central District of California,19 arguing that section 3 of the CSPA, codified at 8 U.S.C. § 1153(h)(3), granted a remedy to all minors who have “aged out” of the sponsoring petition by the time the visa becomes available.20 Under her interpretation, section 3 required that USCIS automatically convert the child’s petition to be derivative of the newly LPR parents’ petition, and — critically — that the child’s petition would maintain the same priority date.21 In other words, the plaintiffs argued that immigration officials could simply approve the primary beneficiary first, then convert the derivative beneficiary’s application to piggyback on the newly approved LPR.22
While Ms. Cuellar de Osorio’s suit was pending, the Board of Immigration Appeals (BIA) considered the meaning of section 3 in In re Wang.23 The BIA concluded that the phrase “automatically be converted” meant that conversion could only occur when a derivative beneficiary’s petition could move from one family preference category to another without creating a new sponsor.24 In other words, the child could not “automatically” become a derivative beneficiary of a parent’s newly granted LPR status — instead, the child must have had an independent relationship with the original petitioning U.S. citizen or LPR.
After the BIA interpreted section 3, Judge Selna granted the government’s motion to dismiss Ms. Cuellar de Osorio’s suit, holding that the BIA’s interpretation in In re Wang was supported by the text and legislative history of the statute, and, accordingly, was reasonable.25
The parents in Ms. Cuellar de Osorio’s suit consolidated their claims with those of another set of similarly situated individuals26 and appealed to the United States Court of Appeals for the Ninth Circuit. Judge Tallman, writing for a unanimous panel, affirmed the judgment of the district court.27 After rehearing en banc, the Ninth Circuit reversed.28 Writing for the majority, Judge Murguia reasoned that § 1153(h)(3) must be considered in its statutory context: subsection (h)(3) is triggered when subsection (h)(1) determines that the alien is twenty-one or older, and subsection (h)(1) explicitly applies to all familial visa categories.29 As a result, aged-out children could automatically convert to another appropriate familial category and retain their original priority dates.30 And since the statutory text contained no suggestion that Congress had intended to exclude children in certain family preference categories, it was not the type of “‘rare and exceptional circumstance[ ]’ that renders the plain meaning of a statute impracticable.”31
The Supreme Court reversed and remanded. Writing for a plurality of the Court, Justice Kagan32 held that the statute’s provisions — namely, the predicate condition in § 1153(h)(1) of determining the alien to be twenty-one, and the remedy in § 1153(h)(3) of “automatic conversion”33 — were in irreconcilable tension, and that the Court must therefore accord deference to the BIA.34 The three Justices in the plurality believed that the CSPA contains an internal contradiction: “We might call the provision Janus-faced,” wrote Justice Kagan, “The two faces of the statute do not easily cohere with each other: Read either most naturally, and the other appears to mean not what it says.”35 Cases like this, she reasoned, are “the kind of case[s] Chevron was built for.”36 And because the BIA chose a reasonable construction of the contradictory commands in the text, the Court deferred to the BIA.37
Justice Kagan also rejected the respondents’ arguments in favor of overturning the BIA’s judgment in In re Wang. First, the respondents argued that aged-out beneficiaries could automatically convert to a new immigration category by “substituting new sponsors for old ones, and by ‘managing the timing’ of every conversion to ensure such a new petitioner exists on the relevant date.”38 But Justice Kagan rejected the contention that the “‘automatic conversion’ procedure permits a change in the petitioner’s identity.”39 Furthermore, Justice Kagan held that the mechanics of “managing the timing”40 precluded the type of automaticity contemplated by the CSPA by requiring “special intervention, purposeful delay, and deviation from standard administrative practice.”41
The respondents also attempted to reconcile the section’s remedial clause by reading § 1153(h)(3) — “the alien’s petition shall automatically be converted to the appropriate category and the alien shall retain the original priority date issued upon receipt of the original petition”42 — as conferring the benefit of retention of the original priority date as “a benefit wholly independent of automatic conversion.”43 But, Justice Kagan wrote, such an interpretation would “make ‘retention’ conditional on something the statute nowhere mentions — a highly improbable thing for Congress to have done.”44
Finally, Justice Kagan rejected the respondents’ contention that the BIA acted unreasonably in choosing the more restrictive reading of § 1153(h)(3): “At the least, the Board’s interpretation has administrative simplicity to recommend it,”45 and in any event, “immigration law’s basic first-come-first-served rule” is served by the BIA’s interpretation that an aged-out beneficiary should not receive credit for the parent’s time spent waiting in the visa queue.46
Chief Justice Roberts, joined by Justice Scalia, authored an opinion concurring in the judgment. Although Chief Justice Roberts agreed that the BIA’s interpretation of § 1153(h)(3) was reasonable, he objected to the plurality’s suggestion that deference was warranted because of a direct conflict between clauses in the statute.47 Indeed, he denied that there exists an internal contradiction in the statute.48 The justification for Chevron deference to agency interpretation of statutes is grounded in the presumption that “Congress intended to assign responsibility to resolve the ambiguity to the agency.”49 But “[d]irect conflict,” he wrote, “is not ambiguity, and the resolution of such a conflict is not statutory construction but legislative choice. Chevron is not a license for an agency to repair a statute that does not make sense.”50 Instead, the concurrence suggested that “[w]hatever other interpretations of that provision might be possible, it was reasonable . . . for the [BIA] to interpret section 1153(h)(3) to provide relief only to a child who was a principal or derivative beneficiary of an F2A petition.”51
Justice Alito penned a brief dissent declaring that the BIA could not ignore what he saw as a “clear statutory command” to convert applications and retain priority dates for aged-out beneficiaries.52
Justice Sotomayor also filed a dissenting opinion, which was joined by Justice Breyer and joined in part by Justice Thomas.53 Justice Sotomayor advanced two reasons that the Court failed to “interpret the statute ‘as a . . . coherent regulatory scheme’”54 by reaching its conclusion that the BIA was free to interpret § 1153(h)(3) as giving relief only to one category of aged-out children.55 First, she argued that the plurality erred by permitting an aged-out child to retain her original priority date only if her petition may be automatically converted; the limitations on automatic conversion, in other words, do not preclude all aged-out derivative beneficiaries from retaining their priority dates.56 Next, the “ordinary meaning” of the words “automatically” and “converted” compel the changing of “an old petition into a new petition in an appropriate category upon the occurrence of some predicate event, without a further decision or contingency.”57
As scholars have observed, the theoretical justification for Chevron deference is unclear.58 Two possible theoretical bases are explored in the Scialabba plurality opinion and in Chief Justice Roberts’s concurrence, respectively. One is the plurality’s narrative of agency expertise and accountability. The other is the concurrence’s narrative of congressional intent59: Chevron is really a way to ensure that courts defer appropriately to Congress, rather than substitute their own predilections for the considered judgment of Congress to defer to the agency. In most cases, the difference between these approaches has limited practical consequence because the end result is the same. But when, as here, a statute contradicts itself, these different understandings of Chevron’s rationale point in opposite directions, and the second basis for Chevron begins to erode. In other words, the doctrine cannot consistently function as a proxy for congressional intent when deference is afforded to internally contradictory statutes.
Justice Kagan’s opinion appears to be undergirded by her belief that the “search for legislative intent [is] chimerical”60: because Congress does not typically give courts direct instruction for the level of scrutiny optimally applied to agency decisionmaking, the court creates for itself “a constructive substitute for an actual statement of legislative desire.”61 Chevron is a creature of the judiciary, created to signal to Congress that the judicial branch will defer to the statutory interpretation of executive agencies unless Congress indicates otherwise.62 And courts have adopted that institutional choice for the simple reason that it reflects judicial intuitions about sound administrative policy to defer to more expert agencies.63 That deference is, as Justice Kagan noted, particularly critical when cases involving immigration come before the Court, because “decisions about [the] complex statutory scheme often implicate foreign relations.”64
The competing theoretical justification for Chevron emphasizes, instead, congressional intent. Chevron, under this view, simply gives voice to the will of the legislative branch, which understands that any statutory ambiguity “would be resolved, first and foremost, by the agency, and desire[s] the agency (rather than the courts) to possess whatever degree of discretion the ambiguity allows.”65 Justice Scalia, among the most vocal supporters of this intent-based reasoning on today’s Supreme Court, has argued that “[t]he extent to which courts should defer to agency interpretations of law is ultimately a function of Congress’ intent on the subject.”66 Later, he would write that the Court accords agencies deference because Congress understood that “[statutory] ambiguity would be resolved, first and foremost, by the agency, and desired the agency (rather than the courts) to possess whatever degree of discretion the ambiguity allows.”67 But, “[a]s both Justices and commentators have noted, . . . this rationale is a fiction.”68 And nowhere is that fiction more exposed than in the unusual circumstances presented by Scialabba: the existence of a statute so mired in contradiction and confusion casts new light on this long-simmering rift.
The logic of the Chief Justice’s concurrence reveals the weaknesses in the congressional-intent justification for Chevron. Chief Justice Roberts’s concurrence maintained that while the statute was ambiguous as to which petitions may be converted, there was no conflict between the statutory provisions — meaning that the statute is not internally contradictory — and he therefore affirmed the agency under Chevron Step Two.69 Indeed, under the Chief Justice’s approach, a world in which Congress grants agency discretion through a self-contradicting statute is not even conceivable.70 Loath to describe the statute as having “internal tension”71 but also loath to grant broad-based relief,72 the Chief Justice was hamstrung. He was forced to attribute intent where none can be found, and resultantly, unconvincingly claimed that Congress did not speak to which petitions may be automatically converted. The concurrence asks the reader to conclude that Congress clearly and intentionally left the decision to deny relief to certain petitioners to the agency — a conclusion that is difficult to justify given established methods of statutory interpretation,73 the legislative history of the CSPA,74 and the amicus brief of current and former members of Congress that argued that “CSPA leaves no room for interpretation as to its scope. . . . [The provision] unambiguously appl[ies] to any derivative-beneficiary child whose age is calculated to be over 21.”75
Since the CSPA is evidence that a statute with an internal contradiction is indeed a possibility, it becomes difficult to understand the congressional-intent justification for Chevron. At the most basic level, when Congress commands an agency to execute both one policy and its opposite, it makes little sense to speak of agency deference in terms of following legislative intent. The canonical objections to judicial efforts to divine legislative intent — that the intent of a body of individuals is “undiscoverable in any real sense”76 and that legislative language may be a product of compromise so it is impossible to determine what Congress truly intended77 — become even more salient when it appears as though Congress made a drafting error. When the true intent of Congress is so obscured by conflicting statutory language that it becomes impossible to speak of the will of the body, a judge may still wish to defer to an agency’s interpretation of the statute because the agency has relatively more expertise or the agency is relatively more accountable. But to defer to the agency’s choice between one policy and its opposite on a theory of legislative intent requires a judge to perform interpretative jujitsu to declare that the statute contains a clear command from Congress.
The natural consequence of Chief Justice Roberts’s approach is to read the statute as though Congress had hoped to vest the agency with a choice: either execute one policy, or execute its exact opposite. But if Congress had meant to legislate in that way, surely it could have said so explicitly, rather than furtively and confusingly abdicating its own authority over a major foreign policy decision. In other words, if it is true that Congress does not generally “hide elephants in mouseholes,”78 it may also be true that Congress does not generally draft elephant-sized legislation when it means to cover mousehole-sized territory. As then-Professors Kagan and Barron put it, “[t]he Court’s cavalier attitude toward the relevant statutory language also suggests a certain disingenuousness in describing the Chevron doctrine as a product of legislative decision.”79 That cavalier attitude — if it ever exists — can only be heightened in the case of a statute that is so internally contradictory that it is impossible to argue that Congress would have intended both policies to be executed.
Courts might consider practical solutions to these theoretical issues. There are at least two doctrinal paths that could create a more harmonious resolution to the problem of self-contradictory statutes. First, similar to the reasoning advanced by Justice Kagan’s plurality opinion, a court might determine that the agency’s expertise and accountability outweigh its own and might afford Chevron deference even when the agency’s answer is based on a choice between two conflicting interpretations of the statute. Such a solution would require a subtle expansion of Chevron: it preserves Step One, but at Step Two broadens deference to agencies beyond “reasonable accommodation of conflicting policies”80 to a frank admission that the agency may choose whatever policy it prefers as between two irreconcilable ones. This approach has the benefit of simplicity and congruence with current doctrine, but it comes at a price: namely, that agency expertise vel non offers weaker support for the legitimacy of agency exercise of authority where Congress has not explicitly endowed the agency with authority to act.
Alternatively, courts might properly understand Chevron as a judicially manufactured background presumption that “must give way to evidence that Congress harbored a different intent.”81 Using this theoretical justification, a court may treat a nonsense statute in a manner analogous to a statute with a scrivener’s error. Just as a “scrivener’s error obscures what was Congress’ real intent,”82 thereby permitting a court to correct technical mistakes when necessary, so too might it be said that in cases of internally contradictory statutes, courts may give “unusual (though not unheard-of) meaning”83 to words in the statutory text. In so doing, a court could resolve the contradiction in a manner that complies with the perceived will of Congress, which might be ascertained through an analysis of legislative history and briefing. In other words, in the extraordinarily narrow subset of statutes with clear internal contradictions, a court may excise agency interpretations that appear to defy congressional will from the realm of Chevron deference. This solution is most apt in circumstances in which the statute at issue is of such importance that it may be fairly assumed that Congress would not want to defer to whichever method of resolving the contradiction that the agency happened to adopt.84