Two weeks ago, the Supreme Court decided Torres v. Lynch, a case about the meaning of a defined term of art in deportation provisions of the federal immigration laws, "aggravated felony," that also appears in the definitions of and sentencing provisions for several federal immigration crimes. Deciding the question in the deportation context, in a case which came up through the Second Circuit from the Board of Immigration Appeals, a closely divided Court agreed with the Board's result and reasoning but gave the Board no deference, instead producing a 21-page opinion which essentially acknowledged the linguistic ambiguity of the text in question and ultimately resolved that ambiguity by appeals to context, purpose, and far-flung analogues in the Assimilative Crimes Act and the federal three-strikes statute.
The parties and amici hotly contested whether Chevron applied to so-called dual-use statutes with civil and criminal applications, a point of emerging doctrinal debate even before Torres; the Court doesn't tell us. But its silence on deference in the face of confessedly ambiguous text has instigated a debate over whether the Court is signaling that Chevron does not apply to agency interpretations of criminal law, does not apply to so-called "crimmigration" questions (i.e. questions about which crimes lead to deportation) within the immigration agency's purview, or whether the Court is signaling nothing of the kind. Michael Kagan, a law professor at UNLV, argues that Torres signals that Chevron has a criminal immigration exception, and that that's a good thing. Patrick Glen, a senior litigation attorney in DOJ's Office of Immigration Litigation who was on the government brief in Torres, argues in response that Torres signals nothing and assures readers that his office will continue to seek Chevron deference to the immigration agency's interpretations of dual-use provisions of immigration law. As someone who, compared to Kagan and Glen, is relatively agnostic on whether there should be a criminal-law exception to Chevron (though I tend to sympathize with Glen), I thought I could neutrally referee this dispute.
Before turning to Kagan and Glen's disagreement on Torres itself, a key point of disagreement between Kagan and Glen is which prior cases, if any, bear on whether the Court is silently crafting an exception to Chevron. Kagan claims that the Court silently denied deference in many pre-Torres crimmigration cases; Glen says he's wrong.
Some background is needed here. The Board of Immigration Appeals is a strange body. Our immigration laws vest the power to deport in the Attorney General. The Attorney General, by regulation for nearly a century, has delegated much of this power by regulation to an entity of her creation, the Board of Immigration Appeals. The Board is a multi-member, court-like body which hears appeals from immigration judges. Like the federal courts of appeals, which it models itself after in many ways, the Board has adopted the modern practice of producing both precedential, published opinions, and non-binding, unpublished opinions. When the Board decides a question of immigration law by unpublished opinion, the courts of appeals are in unanimous agreement that under Mead the Board's decision only gets, at the most, Skidmore deference. Published opinions, by contrast, get Chevron deference. This is a long prelude to saying that, in some of the cases of seeming non-deference on which Kagan relies as evidence for the crimmigration exception, there was no published opinion of the Board to which the SG's office sought Chevron deference. (For an accurate - I have checked - inventory of the Court's recent crimmigration cases that details which even potentially presented a question of Chevron deference, see this page of the SG's brief in Mellouli v. Holder.) And, in some of the other cases Kagan cites, as the page just linked to and Glen's post detail, what was at issue was actually the meaning of a criminal provision of the U.S. Code -- cross-referenced by immigration law but clearly not, by virtue of those mere cross-references, proper subjects of Chevron deference, and therefore not subjects of requests for Chevron deference on the SG's part.
This leaves just two relevant cases,* though Glen disputes the relevance of even these: Nijhawan v. Holder, and Torres itself. Both Nijhawan and Torres involved the meaning of some subsection or subsections of the 21-subsection-long definition of "aggravated felony." This definition appears in a section of the Immigration and Nationality Act, and the term it defines is principally used in immigration laws. If an alien has committed an aggravated felony, he is both removable, and ineligible for discretionary cancellation of removal, making the definition of this term a matter of profound civil consequence. However, what an aggravated felony is also has considerable criminal-law salience, because, besides the term being an element of several immigration-related federal crimes, the prior commission of an aggravated felony increases the maximum sentence for illegally reentering the country by a whopping 900%, or 18 years. This sentencing provision has been cited over 16,000 times by the courts of appeals alone. Hence, if the Board were to get Chevron deference on its interpretation of "aggravated felony," the fates of throngs of federal prisoners would turn on the Board's interpretation of a term of immigration art. Both Nijhawan and Torres argued in their briefs that this would be an untenable state of affairs, Torres having the advantage of recent opinions by Justice Scalia and Judge Sutton agreeing with him about Chevron and criminal law. As it happened, while the Board's interpretation of "aggravated felony" prevailed in both Nijhawan and Torres, the Court said nary a word about deference in either opinion, despite (as discussed below) non-obvious questions of interpretation in Torres if not both.
*Glen would like to rely on the Court's fleeting mention of Chevron in Mellouli v. Lynch, though as he acknowledges, the Court didn't take Chevron very seriously in Mellouli. Mellouli is besides the point for another reason, at least as to an exception to Chevron for criminal law, namely that the crime-based ground of deportability at issue in Mellouli was just a ground of deportability, with no criminal-law implications - unlike the meaning of "aggravated felony," which has substantial criminal-law implications.
Glen's strategy for dealing with Nijhawan and Torres is to claim that the government didn't even seek deference in Nijhawan because the statute was so clear, and only sought deference in Torres as a back-up argument that the Court didn't have to reach because the statute was so clear. I think Glen is right about the government's litigating position in Torres, but little else. On Nijhawan, he is simply incorrect. As in Torres, the SG sought Chevron deference as a back-up argument, devoting some five pages of her brief to disputing Nijhawan's contention that the Board shouldn't receive deference to its interpretations of terms that appear in criminal laws. It is somewhat plausible to say, however, that Nijhawan was an easy enough case to make consideration of Chevron unnecessary. The Court unanimously held in Nijhawan that a subsection of the aggravated-felony definition including fraud offenses where the loss to the victim exceeded $10,000 did not call for the traditional categorical approach of matching convictions, element by element, to statutes attaching collateral consequences to certain crimes, for the simple reason that very few criminal statutes contain, as an element, loss in excess of $10,000. While the text of the statute did not expressly call for a departure from the categorical approach (which itself is an invention of the courts, not the Board), it could only be so fairly read. So the Court said, at any rate.
Nevertheless, the complete absence of the Board's reasoning from the Nijhawan opinion is suspect. The Court decided an issue, whether the categorical approach applied to subsection (M)(i) of the aggravated felony definition, that Congress very probably never thought about and certainly never addressed - that is, just the sort of issue that Chevron was created to handle - and it did so by drawing dubious inferences from a 50-state sua sponte survey of state fraud law loss thresholds that Congress probably never conducted. The Court's bottom line was that because only 21 states had any loss thresholds in their fraud laws at the time the aggravated-felony definition was enacted, and only 8 had thresholds right on the $10,000 button, Congress must have intended for immigration judges to inquire into the underlying facts of fraud offenses when determining whether potential deportees committed offenses involving $10,000 loss. The argument assumes that Congress knew this information - information which of course appears nowhere in the legislative history, and which the SG, in defending the Board's interpretation, didn't even bother to collect. Would Justice Breyer, Nijhawan's author, have charged his law clerk with conducting a survey of state fraud law that on any serious reflection proves nothing had there not been some qualms on the Court about simply deferring to the Board's eminently reasonable interpretation of the statute? Doubtful.
Torres, a 5-3 decision, is an even more suspect case of silence on deference. Torres involved the headscratcher of whether, for purposes of the aggravated-felony definition, an "offense described in" a federal statute named in that definition, "whether in violation of Federal or State law," includes state offenses that don't require the jurisdictional element of the federal statute in question. The Board thought the answer to that question was yes; the Second Circuit, finding ambiguity, deferred. The Court affirmed the Second Circuit without even mentioning Chevron, the Board's rationale, or the fact that the Second Circuit deferred. Instead, it concluded, in 21 pages of closely reasoned text, that the theoretical eligibility of state offenses to be "described in" aggravated felonies suggested that state offenses must be practically eligible as well, and that to be practically eligible, "an offense described in" a federal statute must not include that statute's jurisdictional element.
Glen says the Court reached this conclusion because the statute was unambiguous. Perhaps the Court thought so. One can't read Justice Kagan's opinion without getting the impression of great confidence in her result, borne out of the improbability that Congress would make serious state crimes non-deportable offenses merely because they lacked jurisdictional elements.
However, Torres is hardly a typical Step One opinion. For one thing, the Court never mentions Chevron. For good reason, the Court doesn't usually omit mention of Chevron just because it thinks an agency is unambiguously correct. This term in FERC v. Electric Power Supply Association, the author of Torres dropped a footnote stating that because the Court found the statute in FERC clear, it did not have to address Chevron. Footnotes like these are a necessity in a post-Brand X world, where absent a clear statement that the statute is clear and consideration of Chevron is therefore unnecessary, an agency can change its mind and abrogate the Court's decision; only determinations of unambiguity ensure that the Court's opinion will bind future agencies. Torres does not say reaching Chevron was unnecessary, nor does it use words like unambiguous or clear, except when acknowledging that Congress could have been more clear.
Instead of announcing its decision in terms of clarity, the Court flatly admits that the language at issue has multiple possible meanings and cannot decide the case, and that arguments from context, purpose, and an ostensibly "well-established background principle" about the relative unimportance of jurisdictional elements are needed to decide the case. Usually, when the Court says language has multiple possible meanings and can only be disambiguated by context, it deems that language ambiguous and either defers to the agency or finds some exception to Chevron and disambiguates the language itself.
Moreover, the Court's particular grounds for resolving the case are remarkably far-flung for a Step One opinion. Can "described in" really unambiguously except jurisdictional elements because of an analogous practice of disregarding jurisdictional elements under the Assimilative Crimes Act, a statute that regulates federal military enclaves and appears not to even contain similar language to the aggravated-felony definition? If the Court thought the statute unambiguous, would it need to cite, as evidence of its well-established background principle of selective jurisdictional deemphasis, Judge Diane Wood's views on the meaning of similar language in the federal three-strikes statute, or jurisdictional elements' irrelevance for purposes of mens rea? Ultimately, the Court places greater weight on what Diane Wood thinks about a statute that isn't in the case than on what the responsible agency thinks about the statute that is in the case. Even if the Court silently deemed the latter statute unambiguous, despite the many indications to the contrary, why is the agency's view never marshaled as confirmatory evidence of that unambiguity when so much else is?
The dissent, joined by Justices Sotomayor, Thomas and Breyer, is even odder in this regard. In a case where the government argued Chevron deference and the court below deferred, the dissenters feel free to disagree with the agency without explaining how they overcome Chevron. This too is not the norm; in FERC this term, for example, Justice Scalia explained in dissent that he gave the agency no deference because he thought FERC was clearly wrong.
Do the Torres dissenters think the Board was clearly wrong? Possibly. The dissent, quite unlike the majority, claims to rely on "plain language," "ordinary accepted meaning," and a "straightforward approach." How seriously it takes its claims about plain language, though, is unclear. The dissent devotes just a page to its plain-language argument, giving the weak example of a Craigslist ad that describes an apartment and claiming that the ad only "describes" the apartment if everything the ad says is in the apartment is there. Of course, though, one could and would say that one moved into the apartment "described in" the ad even if some feature of the apartment was altered between advertisement and move, pointing up that what matters here is not what "described in" means standing alone, but what it means to commit an offense described in a federal statute, whether in violation of that statute or state statutes. Ultimately, the dissent places much greater emphasis on what can be inferred from the structure of the aggravated-felony definition than on parrying the majority's claim that committing an offense described in a statute doesn't unambiguously mean committing that offense jot for jot.
So what is happening in Nijhawan and Torres? I think it's clear that something's happening. Total silence on Chevron, in cases where (1) the government asks for deference, (2) the parties hotly debate whether Chevron applies, and (3) the Court rules for the agency without explicitly or even, I would argue, implicitly finding unambiguity, is neither common nor easily explicable. This doesn't mean, however, that what Professor Kagan thinks is happening is happening. I see a few possible explanations.
Most simply, and minimally, at least some members of the Court think that whether Chevron applies to agency interpretations of dual-use statutes with criminal applications is a tough question, and didn't think it had to be decided in Torres or Nijhawan because the agency, if not clearly right, had the better of the arguments. A decision on Chevron's applicability in these contexts, the Court may reasonably feel, can wait for a case where an agency needs Chevron deference to prevail. While adopting the agency's view in Torres and Nijhawan without either expressly finding unambiguity or holding Chevron inapplicable doesn't really fit the doctrine, and creates a potential Brand X problem, that Brand X problem is highly unlikely to come to fruition, as the Board would have to overrule one of its precedents that the Supreme Court ratified and adopt, in either case, starkly pro-criminal-alien views. The doctrinal oddity of these decisions, and the small potential of a Brand X problem, then, is a small price to pay for avoiding a difficult question about Chevron's domain. I find this a very plausible account of what's happening in Torres and Nijhawan.
Equally plausibly, some Justices in the Torres majority and on the unanimous Nijhawan Court are or were firmly convinced that Chevron does not apply to agency interpretations of the aggravated-felony definition, and refused to join (or write) opinions saying that Chevron did apply. (We could also suppose that the Court has internally decided the question already, but if that were the case, it might say so.) Since reaching Chevron was practically, if not doctrinally, unnecessary, other members of the Torres majority and Nijhawan Court acquiesced and didn't write separate concurrences calling for deference. This is quite probable. The author of Nijhawan, Justice Breyer, is famous for a selective approach to the Step Zero question of whether Chevron applies. Justice Scalia, several years after Nijhawan, would write a statement respecting the denial of certiorari highly doubting Chevron's applicability to dual-use civil/criminal statutes. And in Torres, three members of the majority - the Chief Justice, Justice Alito, and Justice Kennedy - dissented in City of Arlington from a holding that Chevron applied programmatically to all ambiguities in agencies' statutes. The Torres dissenters, for their part, included Justice Breyer and Justice Thomas, who lately has begun to doubt whether Chevron is even constitutional. It's easy to imagine them encouraging Justice Sotomayor (if she needed encouraging) to not acknowledge Chevron's applicability in her dissent.
Finally, to the extent that some Justices evidently think that Chevron's applicability in the Torres/Nijhawan context is questionable or doubtful, we might ask why that is. Kagan thinks it has something to do with "the core liberty concerns" involved in immigration law, the Board's lack of political accountability (members, once appointed, can stay seemingly forever, and one Janet Reno appointee remains on the Board), and its lack of expertise in statutory interpretation. Glen harshly criticizes these suggestions. I agree with Glen that Kagan's explanations aren't convincing. The Court, as he points out, has repeatedly deferred to the Board on tough interpretive questions of immigration law, just so long as those questions don't involve the collateral immigration consequences of criminal convictions. Skepticism of Chevron's applicability to issues of deportation or detention, or of the Board's expertise and accountability writ large, can't explain Torres and Nijhawan.
What can explain some Justices' doubts about deference in this particular context is threefold. First, as Nijhawan and Torres argued, with help from amici, there is a decent argument that Chevron doesn't apply to interpretations of defined terms that appear in criminal laws. This isn't an argument to which I'm very sympathetic, as it, among other things, misconceives the relative weight of lenity and Chevron, but outside of the executive branch mine is a decidedly minority view.
Second, the Court may doubt, as the lower courts have at times, whether the Board has anything very useful to say on the classification of state and federal crimes. The substantial expertise the Board has in "pure" questions of immigration law, which can be as labryinthine as the Internal Revenue Code or habeas, may not seem implicated by disputes that call to mind nothing more than some of the Court's less difficult cases under the Armed Career Criminal Act, which also contains a list of generically described offenses that trigger collateral consequences. This is confused, as it either assumes, falsely, that there are correct legal answers about the meaning of the aggravated-felony definition that the Board is less skilled than the Court at finding, or that classifying criminal offenses is a matter for technical expertise - specifically, the technical criminal-law expertise of the Court, which the Court deploys, as in Nijhawan's survey of state fraud law and in similar exercises in Torres, to help carry out Congress's presumed purpose of geographically uniform deportability. One could instead view the many ambiguities in the aggravated-felony definition as delegations to the Board to ratchet up criminal deportability or ratchet it down - a policy choice which the Attorney General's delegates, however unaccountable, are better-suited to make than the Court. However, it would hardly be surprising if some members of the Court didn't see things that way.
Third, the Court's caution about deference in Torres may be a function of the waning fortunes of Chevron itself, which has perhaps only two true believers left on the Court since Scalia's death and Thomas's intellectual journey from writing Brand X to concluding Chevron likely violates Article III - Justices Kagan and Ginsburg. (While joining Scalia's opinion in City of Arlington, Justice Sotomayor's most notable writing on Chevron is a remarkably muscular reading of Mead in her Lawson v. FMR LLC dissent.) With so few votes left for a truly mandatory deference doctrine, it's unsurprising that the Court would decline to apply Chevron in an edge case.