The Court issued four decisions late last week, including one of the more significant, and surprising, of the term:

  • In Allen v. Milligan (No. 21-1086), the Court declined an invitation to defang Section 2 of the Voting Rights Act, instead holding that Alabama’s latest congressional map likely violates the VRA by diluting the collective voting power of the State’s Black population;
  • In Jack Daniel’s Properties, Inc. v. VIP Products LLC (No. 22-148), the Court unanimously reversed a Ninth Circuit decision broadly applying the so-called Rogers test to trademark-infringement claims, holding that the test does not apply when a protected trademark is used to designate the allegedly infringing work’s source, as opposed to performing some other expressive function.
  • In Dubin v. United States (No. 22-10), the Court narrowed the circumstances under which a defendant can be convicted of aggravated identify theft under 18 U.S.C. § 1028A;

Then, just this morning, the Court issued three more:

  • Haaland v. Brackeen (No. 21-376), a 7-2 decision holding that the Indian Child Welfare Act, a federal statute governing state court adoption and foster care proceedings involving Indian children, does not exceed Congress’s Article I authority;
  • And Smith v. United States (No. 21-1576), where the Court unanimously held that when a criminal conviction is vacated because the defendant was tried in the wrong venue, the Double Jeopardy Clause does not bar the defendant from being retried in the proper venue.

We’ve got summaries of Allen, Jack Daniel’s, and Dubin for you today. And we’ll be back to talk about the rest in short order.

We’ll start with Allen v. Milligan (No. 21-1086), which may be the most surprising decision of the term (if not of the last several terms). It involved a challenge under Section 2 of the Voting Rights Act (VRA) to Alabama’s most recent congressional districts, which had the effect of creating just one majority-Black congressional district, while dividing the rest of the state’s Black population up among several majority-white districts. By a vote of 5-4, with Chief Justice Roberts and Justice Kavanaugh joining the Court’s three liberals, the Court concluded that the challengers were likely to succeed on their claim that Alabama’s map violated Section 2, effectively requiring Alabama to redraw its map to create a second majority-Black district.

Much turns on the history of the VRA, so we’ll start there. After the Civil War, the states ratified the Fifteenth Amendment, providing that the rights of citizens to vote shall not be abridged based on race. But that promise proved illusory, as Jim Crow laws throughout the South (like literacy tests and poll taxes) prevented the vast majority of Blacks from voting. Spurred by the Civil Rights movement, Congress enacted the Voting Rights Act in 1965. When originally enacted, Section 2 of the VRA largely tracked the Fifteenth Amendment’s text. As a result, in 1980’s City of Mobile v. Bolden, the Court held that Section 2 (like the Fifteenth Amendment) only prohibited voting practices with a discriminatory intent; it did not bar laws that merely had the discriminatory effect of disproportionately impeding members of a protected racial group from voting. The media and the civil rights community vociferously denounced Bolden soon after it was decided, seeing its discriminatory intent test as a major setback to ensure equal voting rights. But it had its defenders too—including the Reagan Justice Department, where Chief Justice Roberts was then a young hotshot lawyer—who worried that adopting a pure effects test would inevitably lead to a proportionality rule: Whenever a minority group won fewer seats in the legislature than its share of the population, the accusation would be that some state law had produced a discriminatory effect.

The issue of how to respond to Bolden soon made its way to Congress, which extensively debated whether and how to amend Section 2 of the VRA. Ultimately, Senator Bob Dole proposed a compromise: Section 2 of the VRA would include the effects test many championed, but it would also explicitly disclaim a proportionality requirement. That compromise swiftly won bipartisan support in Congress and was signed by President Reagan in June 1982. The resulting text thus provides that Section 2 is violated:

if based on the totality of circumstances, it is shown that the political process leading to nomination or election . . . are not equally open to participation by members of a class of citizens in that its members have less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice. . . . Provided, That nothing in this section establishes a right to have members of a protected class elected in numbers equal to their proportion in the population.

A few years later, the Court addressed the newly amended Section 2 of the VRA in Thornburg v. Gingles (1986). That case involved a challenge to North Carolina’s multimembers districting scheme, which allegedly diluted the vote of its Black citizens. The Gingles Court observed that the “essence” of a Section 2 VRA claim was that a “certain electoral law, practice, or structure interacts with social and historical conditions to cause an inequality in the opportunities enjoyed by black and white voters.” To succeed on a claim under the amended Section 2, a plaintiff needed to satisfy three preconditions: (1) They must show that a minority group was sufficiently large and compact to constitute a majority in a reasonably configured district; (2) The minority group must be “politically cohesive”; and (3) it must be shown that the white majority votes sufficiently as a block to defeat a minority group’s preferred candidates. Once all three preconditions are met, a plaintiff then must show that under the “totality of the circumstances”—a multi-factor test—the political process is not “equally open” to minority voters. That complicated framework has governed the Court’s Section 2 jurisprudence for the last 37 years, though in light of the Court’s recent voting-rights jurisprudence many feared (or hoped) that it would meet its end when the Court agreed to take up this case concerning Alabama’s redistricting plan.

For 115 years after Reconstruction, Alabama elected no Black representatives to Congress. That finally changed in 1992, after a Section 2 lawsuit forced the State to create its first majority-Black district. Alabama’s congressional map remained more or less the same ever since, notwithstanding changes in the composition and distribution of Alabama’s population. Following the State’s ratification of the same basic map again after the 2020 census, several plaintiffs sued, alleging that Alabama’s map violated the VRA by not creating a second majority-Black district centered in the “Black Belt,” a largely rural region of the state where much of its Black population is concentrated. A three-judge District Court presided over a week-long preliminary injunction hearing and then issued a 227-page opinion finding the case was not a “close call”: Alabama’s map clearly violated Section 2 under Gingles.The District Court thus enjoined Alabama from relying on that map for its upcoming 2022 congressional election. But Alabama sought a stay of the injunction in the Supreme Court, which was granted over the dissent of the Chief and Justices Breyer, Sotomayor, and Kagan. At the same time, the Court granted review to consider the case on the merits following full briefing and argument.

Given the stay, and the Court’s line of recent decisions rejecting Section 2 challenges (not to mention eliminating Section 4 and effectively Section 5 too), many observers expected a decision in Alabama’s favor. But instead, the Court affirmed the District Court, with Justice Kavanaugh joining the four Justices who’d dissented from the stay. The Chief Justice (whose own dissent from the stay was somewhat equivocal) wrote for the majority and began with a straightforward application of the Gingles factors. Addressing the first Gingles factor, he noted that the plaintiffs had produced their own proposed map creating a second majority-Black district, which conformed better to traditional districting criteria (like compactness and matching counties, cities, and towns) than Alabama’s actual map. Alabama responded that the challengers’ proposal divided up the “traditional community of interest” of the Gulf Coast region into two districts, but the Court agreed with the District Court that this Gulf Coast region wasn’t really an established community, and that even if it were, Alabama’s map severed a much more significant community of interest in the Black Belt region. On the remaining Gingles considerations, Roberts concluded the District Court was right to find the population of this new second district was politically cohesive; that given how racially polarized voting was in Alabama, its existing map was likely to prevent the Black population in other districts from electing their preferred candidates; and that the totality of circumstances supported the District Court’s factual finding that the political process was not equally open to Black Alabamans as it was to white voters. Under the deferential clear error standard review, the District Court’s factual findings on these points could not be disturbed.

The Chief then turned to what the case was really about: “Alabama’s attempt to remake” the law of Section 2. Alabama contended that so long as a state’s map approximated the one a mapmaker could draw based on traditional districting criteria, while ignoring the racial composition of the resulting districts, the map cannot have “denied or abridged” the right of anyone to vote on account of race. But that approach, the Chief concluded, was not consistent with the “hard-fought compromise that Congress struck” when it amended the VRA in 1982. After all, this race-neutral districting approach would effectively abolish the discriminatory-effects test Congress adopted in response to Bolden. And while Alabama asserted that rejecting its standard would effectively result in a proportionality requirement—something the 1982 VRA compromise also rejected—the Chief observed that lower courts’ application of the Gingles framework over the last decades hasn’t actually done that: Section 2 VRA challenges rarely succeed, and there are hardly any states where the fraction of districts in which minority-preferred candidates are likely to win is equal to the majority group’s percentage of the voting-age population. Roberts addressed many more arguments along these same lines—including challenges to the constitutionality of Section 2 as interpreted by Gingles—but they all led to the same place: The Court declined Alabama’s request to rethink its long-established approach to Section 2, either by scrapping Gingles entirely or by substantially modifying its elements.

Justice Kavanaugh separately concurred, agreeing with the Court in every respect (save for one short section of the opinion). As he saw it, there was no way to avoid the conclusion that Alabama was asking the Court to overrule Gingles. But that ran headlong into stare decisis, which Kavanaugh argued must be stricter in a statutory context (like this one) than on constitutional issues, given that Congress can easily reverse the Court’s interpretation of a statute through legislation. Given that Congress amended Section 2 following Bolden, but left Gingles alone, it logically followed that Congress had endorsed the Gingles framework. While Kavanaugh agreed with the Court in rejecting Alabama’s arguments, however, he warned lower courts that they must “rigorously” apply Gingles’s standards to avoid turning Section 2 into the proportionality mandate the statute prohibits.

Finally, we come to the two dissents. In the first, Justice Thomas, joined in part by Justices Alito, Gorsuch, and Barrett, offered quite a few reasons why he thought Alabama should prevail. To pick a few of the most important: First, the whole concept of Section 2 is that a minority group’s vote is “diluted,” which necessarily provokes the question, “diluted in comparison to what benchmark”? In his view, the obvious benchmark was a map drawn according to race-neutral criteria, just the standard Alabama proposed. The plaintiffs’ map, in his view, too closely resembled the sort of proportionality approach that Section 2 barred. And second, even if the challengers’ race-conscious method of drawing their proposed map was consistent with the VRA’s text, he believed it violated the Constitution’s requirement of race neutrality.

Justice Alito, joined by Justice Gorsuch, also dissented, focusing on Gingles’s first step: the requirement that a Section 2 plaintiff produce an illustrative majority-minority map that is “reasonably configured.” “Reasonably configured” means consistent with traditional districting criteria, like trying to avoid splitting up political subdivisions and communities of interest. From that, Alito concluded that a Section 2 plaintiff’s proposed map isn’t valid unless the plaintiff can show that map would be generated by traditional districting criteria without making the race of voters the predominant factor. In Alito’s view, the challenger’s map failed that requirement, because he believed their proposed map was drawn by giving race a predominant role.

Next up, in Jack Daniel’s Properties, Inc. v. VIP Products LLC (No. 22-148), the Court dialed back a Ninth Circuit expansion of the Second Circuit’s “Rogers test” for determining when a trademark claim must be dismissed at the outset because it relates to an “expressive work.”

Under the Second Circuit’s 1989 decision in Rogers v. Grimaldi—in which the dancer Ginger Rogers sued the distributor of an Italian film called “Ginger and Fred”—trademark claims concerning expressive works must demonstrate either that the challenged use “has no artistic relevance to the underlying work” or that it “explicitly misleads as to the source or content of the work” in order to advance to consideration of whether the use actually infringes on the existing trademark. Though the Supreme Court has never endorsed or adopted the Rogers test, it has been widely adopted in the circuits to filter out claims in which a trademark is used not to designate the source of a work, but rather to perform some other expressive function (like commenting on vapid materialism by singing “I’m a Barbie Girl, in a Barbie world…”). But the Ninth Circuit went a step further in this case, applying the Rogers test where the allegedly infringing use was itself a source designation.

Respondent VIP Products makes squeaky, chewable dog toys designed to look like (and parody) popular beverage brands. There’s Dos Perros (a take on Dos Equis), Smella Arpaw (Stella Artois), and Doggie Walker (Johnnie Walker). And then there’s Bad Spaniels, a take on Jack Daniel’s Tennessee Whiskey. As you can see below, the Bad Spaniels squeaky toy plainly evokes the familiar Jack Daniel’s bottle, but with a caniscatological spin:

In place of “Jack Daniel’s Old No. 7 Brand Tennessee Mash Whiskey,” we have “Bad Spaniels The Old No. 2 On Your Tennessee Carpet.” But it’s not just the words on the Jack Daniel’s bottle that are trademarked, but the entire look, from the white-on-black text to the swirling filigree, and even the distinctive square bottle.

Jack Daniel’s Properties, which owns these trademarks, didn’t appreciate VIP’s joke, and sent a letter stating that Bad Spaniels both infringed and diluted its trademarks and demanding that VIP stop selling the product. VIP responded with a lawsuit, seeking a declaratory judgment that Bad Spaniels neither infringed nor diluted Jack Daniel’s marks. After Jack Daniel’s counterclaimed under the Lanham Act, VIP moved for summary judgment on both the infringement and dilution issues. It argued that the infringement claim failed under the Rogers test, because the Bad Spaniels toy was an “expressive work” and Jack Daniel’s could not show that the challenged use had no artistic relevance to the underlying work or that it explicitly mislead as to the source or content of the work. And with respect to dilution, VIP argued that the “fair use” exception to the Lanham Act applied because its product was a parody of Jack Daniel’s. The District Court rejected both arguments for the same reason: VIP was using Jack Daniel’s trademarks to identify the source of its own product, not for some separate expressive purpose. But the Ninth Circuit disagreed, holding that, because Bad Spaniels is an expressive work (though “surely not the equivalent of the Mona Lisa, it “communicates a humorous message”) it should have been subjected to the threshold Rogers test. It further held that the dilution claim should have been dismissed because VIP’s use of the Jack Daniel’s marks was “noncommercial,” even though used to sell a product, because it parodies Jack Daniel’s.

The Supreme Court unanimously reversed, but in a “narrow” way. Writing for the Court, Justice Kagan stressed that “[w]e do not decide whether the Rogers test is ever appropriate, or how far the ‘noncommercial use’ exclusion goes.” Instead, it held that neither of these “threshold” exceptions to the usual trademark analysis applies where the alleged infringing use of a trademark is itself “source-identifying,” i.e. where “the challenged use of a mark is as a mark.” As Kagan explained, the “primary” function of a trademark is “to identify the origin or ownership of the article to which it is affixed.” While trademarks can do other things—“catch a consumer’s eye, appeal to his fancies, and convey every manner of message,” what makes it a trademark is that “it identifies a product’s source (this is a Nike) and distinguishes that source from others (not any other sneaker brand).” That core function serves both consumers and producers and promotes competition. And the benefits of trademark protection are strong enough that they generally outweigh whatever First Amendment interest there may be in infringing on the core function of a trademark.

In Rogers, however, the Second Circuit recognized that, when trademarks are used in “artistic works,” the use has an “expressive element” that implicates “First Amendment values” and, at the same time, poses only a “slight risk” of confusing consumers about “the source or content of the work.” It therefore imposed the threshold test described above for these cases. As Justice Kagan observed, the other lower courts that have adopted Rogers have similarly confined it to cases “in which a trademark is used not to designate a work’s source, but solely to perform some other expressive function.” In short, Kagan wrote, “whatever you make of Rogers—and again, we take no position on that issue—it has always been a cabined doctrine.” Only the Ninth Circuit in this case has applied it beyond its narrow origins, to protect an allegedly infringing use of a trademark as a mark. That application was error, Kagan concluded, for the “primary mission” of trademark law is to “indicate the source of goods and so to distinguish them from ones manufactured or sold by others.” The “cardinal sin” of trademark law is to undermine that function by confusing consumers about the source of a product. That kind of confusion is most likely when one person uses another’s trademark as a trademark—that is, as a “source identifier”—rather than for some other expressive function.

Here, there was no question that VIP used Jack Daniel’s trademarks as trademarks. Indeed, VIP specifically alleged that it “uses its Bad Spaniels trademark and trade dress as source identifiers of its dog toy.” Just because the use had an additional expressive element—to humorously send up Jack Daniel’s—doesn’t remove it from ordinary trademark scrutiny. If that were the case, the “Rogers might take over much of the world,” for “trademarks are often expressive, in any number of ways.” To be sure, the additional expressive element of a use may be relevant to whether it is infringing at all—which generally turns on the potential for consumer confusion. Here, for example, it is possible that VIP’s use of Jack Daniel’s marks was so humorous and parodic that consumers wouldn’t be confused into thinking it was Jack Daniel’s that was selling the squeaky toy. (We’ll return to that possibility in a moment.) The First Amendment may require some threshold filter like the Rogers test when a mark is sued solely for reasons outside of source identification. But “[w]hen a mark is used as a mark … the likelihood-of-confusion inquiry does enough work to account for the interest in free expression.”

Justice Kagan then turned to the dilution claim. The Ninth Circuit had held that Jack Daniel’s dilution claim failed because the Lanham Act excludes from dilution liability “any noncommercial use of a mark.” While it acknowledged that VIP used the contested marks to sell the Bad Spaniels toy, it insisted that the use was nevertheless “noncommercial” because it “parodies” and “convey[s] a humorous message” about Jack Daniels. Here again, the Court concluded that “[h]owever wide the scope of the ‘noncommercial use’ exclusion,” it does not apply when a mark is used as a mark. As Justice Kagan noted, the Lanham Act recognizes parody as a “fair use” protected from infringement claims, but the fair-use exclusion itself does not apply when the use is “as a designation of source for the person’s own goods or services.” It would make little sense if parody received broader protection when squeezed into the “noncommercial use” exception than it does in the “fair use” exception that it is explicitly associated with.

In sum, on both claims in the case—infringement and dilution—the Court concluded that ordinary trademark principles apply, without a threshold filter, when the challenged use of a trademark is as a source-identifying mark. The Court pointedly took no position on whether and how the Rogers test, or the noncommercial-use exclusion, should apply outside that core trademark territory. (With respect to the Rogers test, Justice Gorsuch did pen a short concurrence—joined by Justices Thomas and Barrett—cautioning lower courts to “handle Rogers … with care.”)

While the decision is a fairly sweeping repudiation of the Ninth Circuit’s reasoning in this case, it’s not a complete win for Jack Daniel’s. Recall that the District Court had earlier reached essentially the same result as the Court did here. The case then proceeded to a bench trial, where the District Court (relying largely on evidence of consumer surveys) sided with Jack Daniel’s on both its infringement and dilution claims. But while the Supreme Court agreed with the District Court’s threshold decision—neither Rogers nor the noncommercial-use exclusion barred Jack Daniel’s claims at the outset—it nevertheless remanded the case for further proceedings on the infringement and dilution claims. And some of the language in Justice Kagan’s decision suggests a degree of skepticism about whether consumers would really be confused about the source of the Bad Spaniels squeaky toy: “Self-deprecation is one thing; self-mockery far less ordinary. So although VIP’s effort to ridicule Jack Daniel’s does not justify use of the Rogers test, it may make a difference in the standard trademark analysis.”

Justice Sotomayor, joined by Alito, drafted a short opinion that might give VIP further hope on remand. She emphasized “that in the context of parodies and potentially other uses implicating First Amendment concerns, courts should treat the results of surveys with particular caution.” “Well-heeled brands with the resources to commission surveys” should not be “handed an effective veto over mockery.” Should the case continue to a second trial, therefore, it’s possible the District Court would regard Jack Daniel’s survey evidence a bit more skeptically.  

Finally, in Dubin v. United States (No. 22-10), the Court seriously cut back the Fifth Circuit’s expansive reading of the federal aggravated identify theft statute.

David Dubin helped his father run a psychological testing firm. Under his watch, the company submitted a bills for reimbursement to Medicaid claiming that certain tests were performed by a licensed psychologist, so they were billed at a higher rate. In fact, those tests were performed by a psychological associate, so a lower rate would have applied. Dubin was charged with both health care fraud and aggravated identify theft under 18 U.S.C. § 1028A. The addition of the identify theft count was significant as it carried with it a two-year mandatory minimum sentence.

The identify theft statute applies when a defendant, “during and in relation to any [predicate offense like health care fraud], knowingly transfers, possesses, or uses without lawful authority, a means of identification of another person.” The Government argued that Dubin violated the statute by using a patient’s Medicaid ID number in the course fraudulently billing at an inappropriately high rate. Because Dubin “use[d]” that means of identification, his health care fraud also met the elements of identity theft, so he was sentenced to an additional two years’ incarceration. Dubin appealed, arguing that the “use[]” of identification amounts to aggravated identify theft only when that “use[]” is the crux of the fraudulent scheme. The Fifth Circuit disagreed and affirmed, taking a more expansive view of the identity theft statute than the First, Sixth, Seventh, and Ninth Circuits. The Court granted cert to resolve that conflict.

In an opinion by Justice Sotomayor, joined by all but Justice Gorsuch, the Court rejected the Fifth Circuit’s expansive reading. The Court focused on two elements of the statute. First, the defendant must have “use[d]” the means of identification of another person. Second, that use must have been “in relation to” the predicate offense, like health care fraud. Both “use” and “in relation to” can have varied and broad meanings, so the Court has previously held that each phrase can only be understood in the context of the statute in which it occurs. Here, the title of the statute—“Aggravated identify theft”—illustrates that Congress intended that the term “in relation to” should mean that the use of the means of identification was the “crux” of the underlying scheme. Further, the title shows that the word “use” should mean that the identification was central to the fraudulent conduct. That is, the means of identification itself was deceptive, rather than merely a step along the way to some other method of fraud.

The Court next found support for its more restricted reading in the other words surrounding “uses”—that is, a defendant must have “transfer[red], possesse[d], or use[d]” the means of identification of another person. The Court applied a canon of construction called noscitur a sociis, Latin for “it is known by its associates.” Under that canon, when a word appears in a list, courts will construe that word to fit in thematically with the other words in the list. Here, “transfer” and “possess” clearly refer to theft. Even if “use” might be read more broadly on its own, in context, it is part of a “trio of verbs” intended to reflect the ordinary understanding of identity theft. Further, giving the term “use” the broad reading that the Government advocated would render the other words in the list superfluous. Read together, each word should be given a distinct meaning: “transfer” and “possess” refer to the appropriation of another’s identity, while “use” refers to the deceptive end to which that identity is put.

The Court emphasized that the Government’s proposed interpretation was overbroad for several reasons. First, the identity theft statute functions as an enhancement, adding a two-year mandatory minimum sentence onto certain predicate crimes. If the statute were as broad as the Government argued, then the enhancement and the underlying crimes would effectively collapse together, leading to significantly more severe sentences for a variety of offenses. The Court takes a restrained approach in interpreting criminal statutes, particularly when they are ambiguous. And the Court rejected the Government’s contention that a broader reading would be appropriate because federal prosecutors would wield the statute responsibly, declining to interpret the language of the statute based on the promise of reasonable prosecutorial discretion.

Applying its interpretation to Dubin’s conduct, the Court held that his use of Medicaid patient identification numbers was merely incidental to the overall scheme. That is, Dubin fraudulently misrepresented how and when services were provided, not the identity of the person receiving services. Accordingly, his aggravated identity theft conviction could not stand.

Justice Gorsuch concurred in the judgment only. He agreed that the Government’s interpretation of the statute was significantly overbroad. He criticized the majority opinion, though for articulating what the statute does not say without providing sufficient affirmative guidance on what it does say. Justice Gorsuch walked through a series of hypotheticals, pointing out that the Court’s new standard for aggravated identity theft may not always be simple to apply. Ultimately, he suggested that the statute might be unconstitutionally vague, and noted his hope that Congress would step in to clarify its intent. Otherwise, the statute may be back before the Court for further elucidation.