Greetings, Court Fans!

The Nine have gone their separate ways, but we’re still soldiering on (albeit from vacation and paternity leave, respectively). We’ve got six more cases for you, so we’ll split them up into manageable trios. Read on for summaries of Kisor v. Wilkie (No. 18-15), in which the Auer doctrine narrowly escaped the jurisprudential dustbin; Food Marketing Institute v. Argus (No. 18-481), in which the Court’s textualists applied a strict construction to one of the Freedom of Information Act’s statutory exceptions; and United States v. Haymond (No. 17-1762), a potential sleeper extending the Apprendi line of cases to the context of probation and supervised release.

A continuing theme this Term has been whether to apply stare decisis to preserve the Court’s precedents or to overrule them as discarded relics of another era. On two notable occasions this term, the Court has followed that latter route, overruling longstanding decisions governing access to federal court for takings claims and to state court for lawsuits against other states. Justice Kagan (along with the other liberals) dissented from those rulings, but in Kisor v. Wilkie (No. 18-15) she stepped forward to assemble a majority to preserve another longstanding precedent, the Auer doctrine of judicial deference to a federal administrative agency’s reasonable interpretations of its own ambiguous regulations. Though Auersurvived, it hangs on by a thread. Justice Kagan’s opinion narrows the doctrine considerably and was only joined in part by the Chief (who provided the decisive fifth vote, along with Ginsburg, Breyer, and Sotomayor). And, although they technically concurred in the judgment, Justices Gorsuch, Thomas, Alito, and Kavanaugh still have Auer in their sights.

The mundane facts of the case belie its potential importance. James Kisor, a Vietnam War veteran, was denied disability benefits by the Department of Veterans Affairs (VA) in 1982, but in 2006 Kisor successfully reopened the decision and obtained benefits due to his PTSD—but the VA, as affirmed by a single administrative judge at the Board of Veterans’ Appeals, refused to make the benefits retroactive. The Board’s decision turned on the meaning of “relevant” in a regulation making benefits retroactive only if there were “relevant” department records not considered in the initial denial. Kisor’s new documents were not considered previously, but also were not directly on point to the reason for the initial denial, and the Board took the latter approach to relevancy. The Federal Circuit affirmed, applying Auer deference to uphold the Board after finding the regulation ambiguous and both sides’ views to be reasonable.

The Supreme Court affirmed the concept of judicial deference to an agency’s interpretation of its own regulations, but spent much effort in its opinion explaining and limiting the doctrine, indeed criticizing previous opinions that reflexively deferred to agency interpretations. Justice Kagan’s opinion laid down strict rules for when a court may defer to an agency: (1) the regulation must be genuinely ambiguous, which requires a court to engage in its own rigorous consideration of the regulation’s text, structure, history, and purpose just as if the agency had never opined on the meaning, which will often resolve ambiguities without resort to deference or, at least, define the range of permissible interpretation; (2) the court must evaluate if the “character and context” of the agency’s interpretation warrants deference, i.e., whether (a) the interpretation is authoritative or an official position rather than an ad hoc or informal agency position, (b) the topic of the interpretation is one where the agency’s subject-matter expertise matters, and (c) the interpretation reflects fair and considered judgment, so it’s not a post hoc rationalization to support a litigated position or a new interpretation creating unfair surprise or in conflict with an earlier agency interpretation. The Court emphasized the strong judicial role in interpreting agency rules, to ensure that agencies are given leeway to interpret their rules only when appropriate.

Specifically addressing Kisor’s request to overrule the Auerdoctrine, Justice Kagan discussed the importance of stare decisis in adhering to settled precedent, particularly in a case like this where (1) Auer, while decided in 1997, has its roots in cases going back more than 75 years; (2) abandoning Auer would cast doubt on many settled constructions of rules; (3) unlike with constitutional precedent, where the Court has the last word, Congress could (if it so desired) amend an agency’s governing statutes or the Administrative Procedure Act to alter the interpretive methodology for judicial review. As for Kisor’s own dispute with the VA, the Court applied its modified Auer doctrine and held that the Federal Circuit did not do its job when deferring to the VA: The Federal Circuit failed to engage in its own thorough interpretation of the regulatory text when “casually” declaring an ambiguity, and it failed to consider the character of the Board of Veterans’ Appeals’ decision as coming only from a single administrative judge that may not reflect the agency’s considered judgment or have precedential value even at the agency. The Supreme Court therefore remanded to the Federal Circuit to apply the Auer doctrine as newly explicated in its majority opinion.

That was the extent of the five-Justice majority opinion, but Justice Kagan wrote more for the plurality of four liberal Justices, without the Chief. The additional parts of the opinion discussed the need for expert agencies who wrote the rules to explain the rules in new or unanticipated contexts, and viewed deference as consistent with congressional intent for agencies to play the primary role in resolving regulatory ambiguities under their delegated authority and to ensure more uniform application of rules. Another additional part of the plurality opinion defended the Auer doctrine as consistent with the Administrative Procedure Act’s provisions governing judicial review and rulemaking. As for the latter, Justice Kagan emphasized that an interpretive rule not adopted with formal notice and comment rulemaking doesn’t have the force of law, even if deferred to by a court, and cannot itself form the basis for an enforcement action. And for notice-and-comment rules, courts retain the final say using the rubric approved by the majority for when to defer to the agency. Justice Kagan also found no evidence or logic in Kisor’s argument that agencies will promulgate vague, inscrutable regulations just so they can later interpret them and receive judicial deference.

Chief Justice Roberts wrote a separate concurrence explaining his joining the majority opinion, largely on the basis of stare decisis and his agreement with the new limitations placed on Auer deference. He also sought to emphasize that there is little “distance” between the majority opinion and the four Justices who would overrule Auerbut still let courts be guided by the persuasive value of well-reasoned agency interpretations of their rules. In fact, he thought in most cases the result would be the same under either approach. On a side note, he also stated that nothing in the majority opinion touches on the deference courts should give to an agency’s interpretation of a federal statute under the separate Chevrondoctrine.

Though he technically concurred in the judgment, Justice Gorsuch’s separate opinion (joined by Thomas, Kavanaugh and, in large part, Alito) reads much more like a dissent. Not content to merely narrow Auer, Justice Gorsuch would have overruled it outright and left it to courts to give litigants their “best independent judgment of the law’s meaning.” He viewed Auer deference as an accidental historical anomaly deviating from the court system’s usual role. He also viewed that deference as inconsistent with the judicial review and rulemaking provisions of the Administrative Procedure Act and with the Constitution’s vesting of judicial power solely with the Judiciary and not the Executive Branch. (Interestingly, Chief Justice Robert’s views on those issues are unstated, as he didn’t join the relevant portions of Justice Kagan’s opinion, instead joining only Justice Kagan’s stare decisis rationale for keeping Auer in some form.)

Justice Gorsuch went on, in a portion of his opinion joined only by Thomas and Kavanaugh, to rebut Justice Kagan’s policy reasons for deferring to an agency’s current view of a previously promulgated regulation, and to rebut the majority’s invocation of stare decisis. Gorsuch questioned whether stare decisis even applies to precedent addressing a method of regulatory interpretation and saw no reason to apply the principle anyway as even the majority saw the need to greatly modify Auer from how courts have been applying it. Plus, with the growth of the regulatory state, it is all the more important now to correct course. And, as Justice Gorsuch wrote, there need not be concern about how overruling Auer would cast doubt on settled constructions of regulations because (1) the majority already has changed how Auer should be applied, without worrying about that effect, and (2) even after Auer is gone, earlier decisions construing specific laws could still be given stare decisiseffect as settled precedent. Justice Gorsuch’s bottom line is to leave courts to apply another earlier decision, Skidmore, and follow an agency’s views only to the extent they are persuasive.

Finally, Justice Kavanaugh wrote a short opinion (joined by Justice Alito) agreeing with Justice Gorsuch about overruling Auer but also emphasizing what the Chief Justice wrote in his separate opinion: The courts, if they apply the majority’s more rigorous application of Auer, should lead in most cases to the same result as applying an independent analysis without Auer; and this decision should not be construed to address the viability of Chevron deference by courts to an agency’s interpretation of a statute as opposed to its own regulations.

Justice Gorsuch led the Court in Food Marketing Institute v. Argus Leader Media (No. 18-481), providing a discourse on statutory construction in the course of interpreting the Freedom of Information Act (FOIA).

A South Dakota newspaper, the Argus Leader, sought data collected by the U.S. Department of Agriculture (USDA) identifying all retail stores participating in the food-stamp program (Supplemental Nutrition Assistance Program or SNAP) and each store’s annual SNAP redemption data for a five-year period. The USDA provided the list of stores but withheld the store-specific data based on FOIA’s Exemption 4, which shields from disclosure not only trade secrets, but also “commercial or financial information obtained from a person and privileged or confidential.” 5 USC § 552(b)(4). After a two-day bench trial, the district court had found that stores would suffer competitive harm from disclosure yet the harm would not be “substantial.” The Eighth Circuit therefore upheld the USDA’s decision because it required a showing of “substantial harm” to the “competitive position” of the stores that provided the data to the government before the commercial information can be treated as “confidential” under the statutory exemption.

Several circuits employ that same standard derived from the 1974 D.C. Circuit’s National Parks decision. Yet, the Supreme Court reversed in a 6-3 ruling, discarding this entire line of cases. Justice Gorsuch, putting on his reading glasses, could not find in the statute the words “substantial competitive harm” from disclosure, as the statute requires commercial information only to be “confidential.” The Chief Justice and Justices Thomas, Alito, Kagan, and Kavanaugh joined him. Before interpreting the statute, he first upheld the Food Marketing Institute’s standing on behalf of the grocery retailers, who would suffer concrete injury, traceable to the lower court’s disclosure order, that would be redressed by a successful appeal. As for reading the statute, Justice Gorsuch, “as usual,” asked what the term “confidential” meant when Congress enacted FOIA in 1966. He turned to 1960s dictionaries to find a meaning of “private” or “secret”—and he construed the data as meeting that definition because both (1) the stores customarily kept the data private, and (2) the recipient of the data (the USDA) provided assurances of secrecy. The Court held that item (1) was necessary to make the data confidential but left for another day whether item (2) was also necessary, although it was satisfied here.

Justice Gorsuch referred to the 1974 D.C. Circuit decision, which adopted the “substantial competitive harm” standard, as “a relic from a bygone era of statutory construction” that impermissibly engaged in “a casual disregard of the rules of statutory interpretation.” He held out particular disdain for the D.C. court’s resort to legislative history before consulting statutory text and structure, especially the reliance on testimony at congressional hearings that didn’t even match the later official committee reports that hewed closely to the statute’s plain meaning. He also rejected the argument that “confidential commercial information” is a term of art; after all, Congress didn’t use that exact term, instead referring in the statute to “commercial … information [that is] … confidential.” And he rejected the argument that Congress ratified the 1974 decision by since enacting similar phrases in other statutes, because those later enactments don’t tell us what Congress meant when FOIA was enacted in 1966 and the ratification doctrine would only have relevance if Congress had, after 1974, reenacted the same FOIA statute using the same language, which it hadn’t done. In the final analysis, the standard used by the lower courts was motivated by a policy decision to favor FOIA disclosure, but FOIA’s exemptions are as much a part of the statute as the disclosure requirements, and each part of the statute must be construed based on a fair reading of the text. That reading means that the store-level SNAP data are “confidential” within the meaning of FOIA’s Exemption 4.

Justice Breyer dissented, joined by Justices Ginsburg and Sotomayor. He actually agreed with the majority that the FOIA exemption doesn’t include the requirement of showing “substantial competitive harm” from disclosure. But, he also would not let the industry and a government agency defeat FOIA’s goals by themselves turning non-confidential information into confidential information simply by treating it as private or secret. He reads the term “confidential” as necessarily referring to information where some harm would arise from disclosure. It need not be “competitive” harm but could be some other sort of business or economic harm. It need not be “substantial” harm but need only be “genuine” harm. He would remand for a new determination of whether release of the data would cause genuine harm to a business owner’s economic or business interests.

Finally (for today), United States v. Haymond (No. 17-1672)continues the trend of late-term decisions that are unusually fractured. Four members of the Court (Justice Gorsuch, joined by Justices Ginsburg, Sotomayor, and Kagan), concluded that 18 U.S.C. § 3573(k)’s mandatory minimum five-year term of imprisonment for anyone who possesses child pornography while on supervised release violates a defendant’s right to a jury trial under the Fifth and Sixth Amendments. Four others (Justice Alito, joined by the Chief and Justices Thomas and Kavanaugh) warned that Justice Gorsuch’s opinion was potentially revolutionary, calling into question whether the entire system of federal supervised release violated the Constitution. In the middle was Justice Breyer. Though he “agree[d] with much of the dissent,” he also agreed with the Justice Gorsuch-led plurality that Section 3583(k) was unconstitutional given the somewhat unique way it operates. For now at least, the revolution is on hold.

Petitioner Andre Haymond was convicted by a jury of possessing child pornography in violation of federal law. He was sentenced to a relatively short term of 38 months’ imprisonment, followed by 10 years of supervised release. While on release, an unannounced search of his computer turned up child pornography, so the government sought to revoke Haymond’s supervised release and obtain a new, additional prison term. In a hearing, held without a jury, the district judge found it more likely than not that Haymond had knowingly downloaded and possessed some of the images. Ordinarily, under 18 U.S.C. § 3583(e)(3), a defendant who violates conditions of his supervised release may be sentenced to a new prison term up to the maximum period of supervised release authorized by statute for the defendant’s original offense, with judges having substantial discretion to select the appropriate sentence. (That would have resulted in a sentence of no more than two years.) But Section 3583(k) changes this default rule, providing that a defendant who commits one of several enumerated offenses (possession of child pornography among them) while on supervised release must be sentenced to an additional term of at least five years, regardless of the length of prison term authorized for the defendant’s original conviction. Based on that provision, the district judge concluded he had no choice but to impose an additional prison term of five years’ imprisonment.

On appeal, the Tenth Circuit held that Section 3583(k) violated the Fifth and Sixth Amendments. In its view, a jury had convicted Haymond beyond reasonable doubt of a crime carrying a prison term of zero to ten years. But now, Haymond faced a potential prison term of five years to life. Because the mandatory minimum for this new sentence rested on facts found by a judge, rather than a jury, the Tenth Circuit concluded Haymond’s right to a jury trial had been violated and remanded for resentencing based on the general rule of Section 3583(e)(3). (The district court then sentenced him to time served, as Haymond had been in prison for more than two years while his appeal was litigated.) The Supreme Court granted certiorari to consider the Tenth Circuit’s approach to Section 3583(k).

Justice Gorsuch’s plurality opinion began with a long paean to the importance of the right to a jury trial in all criminal prosecutions. For most of U.S. history, the division of responsibility between the judge and jury in criminal convictions and sentencing was straightforward: The jury would find the facts of the offense, and the judge would then prescribe the sentence based on those facts. True, sometimes judges would find aggravating or mitigating facts that affected the sentence, but these facts did not increase the penalty beyond what the law provided for the acts charged and found by the jury.

This division came under strain with more recent sentencing regimes. Beginning with Apprendi v. New Jersey (2000), the Court has held that it violates a defendant’s right to a jury trial for a defendant to be sentenced to an a term of imprisonment longer than the statutory maximum of the offense found by the jury based on facts found by a judge. In Alleyne v. United States (2013), the Court extended Apprendi from maximum sentences to mandatory minimums. Thus any facts that subject a defendant to an increase in the mandatory minimum penalty must also be found by the jury.

Much like the Tenth Circuit had done, Justice Gorsuch and the plurality found that these cases resolved Haymond’s. The offense found by the jury subjected Haymond to a prison term of zero to ten years. But then a judge, without a jury and on a preponderance of the evidence standard, found that Haymond had engaged in additional conduct that subjected him to a term of five years to life. Just like in Alleyne, these facts that resulted in a mandatory minimum sentence had to be found by the jury.

The only real question was whether it changed anything that Haymond’s imprisonment was for violating conditions of his supervised release. The simple answer was no. While the government (and the dissent) offered various arguments for why supervised release should make a difference, the basic fact of the matter was that Haymond was subject to a mandatory minimum sentence based on facts never found by the jury, and that result violates the Fifth and Sixth Amendments regardless of the terminology one uses to describe the basis of the sentence.

Before getting to Justice Breyer, it is perhaps best to turn to Justice Alito’s dissent. Much of it was concerned with just what follows from the plurality’s opinion. He noted that many of Justice Gorsuch’s pronouncements suggested that a defendant has a right to a jury trial in any supervised-release revocation proceeding. This would make the current federal supervised release regime largely impossible. And even if the plurality would not go that far, at the very least the logic of its opinion appeared to require a jury whenever the defendant’s sentence for violating the conditions of supervised release exceeds the maximum sentence he could have received for his original offense. In an apparent attempt to work the refs (i.e., the Court of Appeals Judges who will have to decide how Haymond applies to this different context in future case), Justice Alito argued this too was wrong, because the facts found by the jury in the defendant’s original conviction entail not only the ordinary term of imprisonment but also the term of imprisonment that could follow from violation of conditions of supervised release (generally a mandatory component of federal criminal sentences). Based on similar logic, Justice Alito would have held that Section 3583(k) does not run afoul of Apprendi or Alleyne, because the maximum period of confinement for the defendant’s original conviction (found by the jury) is the maximum term of imprisonment for that offense plus the maximum term of imprisonment that would follow from revocation of supervised release. In this respect, supervised release is much like the system of parole it replaced. And it is well established that there is no right to a jury trial in parole revocation hearings. Justice Alito would follow this same rule here.

Now we can finally turn to Justice Breyer’s concurrence, which, as the fifth vote, is the controlling opinion. As noted, he began by stating his agreement with much of the dissent: Supervised release is little more than a variation on parole proceedings, and the jury trial rights implicated by one should not be meaningfully different from the other. To avoid “potentially destabilizing consequences,” Justice Breyer would therefore not apply the Apprendi line of cases to the supervised release context in general. Despite that, he agreed with the plurality that Section 3583(k) was unconstitutional, because it departed from standard role of supervised release. Generally, supervised release serves to punish the defendant for violating the terms and conditions of the supervised release imposed by his original sentence; it does not exist to imprison the defendant based on new criminal conduct. While the former aptly describes supervised release revocation in the ordinary case (i.e., under Section 3583(e)(3)), Section 3583(k) seems much more like the latter: It applies only when the defendant commits certain federal criminal offenses and it greatly restricts the judge’s discretion over what term of imprisonment is necessary in light of the violation of supervised release conditions. Taken together, these provisions make Section 3583(k) much more like the punishment of a new offense than punishment for violating supervised release, so the ordinary jury-trial rights that apply to new criminal prosecutions should apply here as well.

That’s enough for now. We’ll be back tomorrow to finish out the term, with summaries of Tennessee Wine and Spirits Retailers Assn. v. Thomas (No. 18-96), a con-law novelty combining the Twenty-First Amendment and the dormant Commerce Clause; Rehaif v. United States (No. 17-9560), a crimmigration case addressing the mens rea for violations of 18 U.S.C. § 924(a)(2); and Dutra Group v. Batterton (No. 18-266), an admiralty case on whether punitive damages are available for “unseaworthiness” claims.