New Jersey law allows candidates for elected office to place a six-word slogan next to their names on the ballot.1 However, candidates who want to use the name of a person or New Jersey corporation in their slogans must first obtain that person’s or corporation’s consent.2 Recently, in Mazo v. New Jersey Secretary of State,3 the Third Circuit held that this consent requirement was a content-neutral regulation of speech that did not violate candidates’ First Amendment rights.4 In holding that the New Jersey consent requirement was content neutral, the Third Circuit adopted an implausibly broad reading of the Supreme Court’s recent decision in City of Austin v. Reagan National Advertising of Austin, LLC.5 The Third Circuit set forth a new “extrinsic features” test for content neutrality,6 which exempted from strict scrutiny a law that effectively prohibits certain categories of political speech. Mazo exemplifies the potential danger of City of Austin’s departure from the clear content-neutrality rule of Reed v. Town of Gilbert.7
New Jersey permits candidates running in primary elections for state or federal office to choose a slogan of up to six words that will appear next to their names on the ballot.8 However, there is a catch. If candidates want their slogans to include or “refer to” the name of any person or any New Jersey corporation, they must first obtain that person’s or corporation’s written consent.9 Candidates must file that written consent with the Secretary of State as part of their petitions to appear on the ballot.10
This “consent requirement” proved troublesome for Eugene Mazo and Lisa McCormick, who were candidates in New Jersey’s July 7, 2020, Democratic primary election for the House of Representatives.11 Mazo initially sought to use the following three slogans, one in each of three different counties within New Jersey’s Tenth Congressional District: “Essex County Democratic Committee, Inc.”; “Hudson County Democratic Organization”; and “Regular Democratic Organization of Union County.”12 But because those slogans all referred to the names of New Jersey corporations whose consent to be named Mazo had not obtained, state officials rejected the proposals.13 Mazo eventually chose different slogans.14 McCormick initially submitted the slogan “Not Me. Us.,” but that slogan was denied on the basis that it “referred to” a New Jersey corporation.15 She then tried to use the slogan “Bernie Sanders Betrayed the NJ Revolution,” but that too was rejected since it referenced a person — Senator Bernie Sanders — whose consent McCormick had not obtained.16 McCormick ultimately used the slogan “Democrats United for Progress” after receiving the consent of that corporation.17
Five days before the primary election, Mazo and McCormick sued the New Jersey Secretary of State and various election officials in the United States District Court for the District of New Jersey.18 They sought declaratory relief and an injunction against the enforcement of the consent requirement, arguing that it violated their First Amendment rights.19 In granting the defendants’ motions to dismiss, the district court first concluded that the plaintiffs’ claims were not moot and were ripe, although the primary election had already passed and the next primary was over a year away.20 The district court then upheld the constitutionality of the consent requirement after applying the Anderson–Burdick21 balancing framework.22
The Third Circuit affirmed.23 Writing for the panel, Judge Krause24 first affirmed the district court’s conclusion that the case was neither unripe nor moot.25 The court then concluded that the consent requirement should be assessed as an election regulation under the Anderson–Burdick balancing framework, as opposed to a regulation of pure speech that would trigger a First Amendment doctrinal test.26 The Mazo court adopted a “two-track approach” to the Anderson–Burdick framework.27 Under that two-track approach, a law that imposes a “severe” burden on constitutional rights is subject to strict scrutiny.28 But if a burden imposes only “reasonable, nondiscriminatory restrictions” on constitutional rights, then “the State’s important regulatory interests are generally sufficient to justify the restrictions.”29 The court concluded that the consent requirement imposed only a minimal burden on First Amendment rights.30 The law applied equally to all candidates and slogans and left open “ample and adequate alternatives for expression and association.”31 As to whether the law was “nondiscriminatory,” the court noted that “[w]hether a law is viewpoint- or content-based may also bear on the severity of the burden imposed.”32
The court then held that the consent requirement was a content-neutral regulation of speech.33 The court based its holding on the Supreme Court’s decision in City of Austin.34 In that case, the Court upheld the sign code of Austin, Texas, which permitted businesses to use digital signs to advertise goods and services located on the same premises as the sign, but prohibited digital advertisements for things located on different premises.35 The Mazo court explained that under City of Austin, a law can remain content neutral even if it requires the enforcement official to examine speech, but only if the speech is examined “in service of drawing neutral lines.”36 The Mazo court read City of Austin as expressly endorsing two categories of content-neutral line drawing: (1) when examining the speech is necessary to determine compliance with a neutral time, place, or manner regulation, such as the on-/off-premises distinction in City of Austin itself; and (2) when examining speech is necessary to determine its “function or purpose,” such as whether speech constitutes regulable “solicitation.”37 In addition to those two enumerated categories, the Mazo court then read City of Austin as supporting a “third category” in which a state may examine speech: to determine whether the speech contains “extrinsic features unrelated to the message conveyed.”38 According to the court, the consent requirement fell into this third category, since an enforcement official needed to examine a ballot slogan’s content only to determine objectively whether the candidate satisfied the consent requirement.39 Finally, the court concluded its Anderson–Burdick analysis. Since the law’s burden on speech was minimal, it was justified by New Jersey’s four asserted interests in the consent requirement: “preserving the integrity of the nomination process, preventing voter deception, preventing voter confusion, and protecting the associational rights of third parties who might be named in a slogan.”40
Mazo is particularly concerning for its holding that New Jersey’s consent requirement is content neutral. To reach that conclusion, the Third Circuit read City of Austin as allowing for a content-neutrality determination if an official examines speech for the purpose of identifying “extrinsic features unrelated to the message conveyed.”41 This expanded test for content neutrality rests on an implausibly broad reading of City of Austin. That case does not contemplate that a law can be considered content neutral if it authorizes a government official to ban entire categories of political messages from a particular sphere of public discourse, as New Jersey’s consent requirement does. Although Mazo should have adopted a narrower reading of City of Austin, its expansion of that case is perhaps a foreseeable side effect of City of Austin’s departure from Reed v. Gilbert’s clear content-neutrality rule. To prevent additional expansions of City of Austin, the Supreme Court should clarify its limits in a future case.
City of Austin rejected as “too extreme” the lower court’s holding that a regulation is content-based whenever it requires an enforcing official to read the sign at issue.42 The Court explained that Austin’s ordinance was content neutral because it required an examination of speech “only in service of drawing neutral, location-based lines.”43 However, the case did not define what counts as a “neutral” line. In lieu of a comprehensive definition, the Court gave two precedent-based examples of laws that require evaluating speech while still remaining content neutral: (1) those that require an official to distinguish “between on-premises and off-premises signs,” and (2) those that require an official to “identify whether speech entails solicitation.”44 But the case provided no further guidance regarding which other types of restrictions, if any, could count as neutral.
The Third Circuit tried to fill this gap by defining “neutral” restrictions to include those that regulate speech “based on extrinsic features unrelated to the message conveyed.”45 But that is too broad a reading of City of Austin. The central problem with the court’s new test is that “extrinsic features” will often determine which messages can be spoken and which cannot. The extrinsic feature at issue here — names of people and corporations who had not consented — is a prime example. Suppose New Jersey banned any candidate from referencing any deceased person in any election speech, for the same reasons as it passed the consent requirement.46 This would prevent candidates from saying things like “Ronald Reagan would want you to vote for me.” The mention or non-mention of a deceased person would be an “extrinsic feature” of the election speech. This restriction would be “unrelated to the particular message conveyed” to the same extent as the consent requirement is unrelated to the particular message contained in a ballot slogan, since this hypothetical ban would “appl[y] to all [speeches], regardless of message.”47 But such a rule would not pass City of Austin, or even “the laugh test.”48
Another hypothetical will further illustrate the implausibility of the Mazo court’s expanded test for content neutrality. Suppose that New Jersey had established a database of “hurtful words or phrases” that it deemed offensive to the state’s citizens, and then banned electoral candidates from using any of those designated words in their six-word ballot slogans. Although this hypothetical regulation would be a transparent attempt to regulate the content of speech, it would seem to qualify as a content-neutral law under the Third Circuit’s test, since the law identifies an “extrinsic feature[] unrelated to the message conveyed.”49 The official would need to examine the speech only in service of evaluating whether an extrinsic feature, capable of objective determination, was present — does the slogan use any of the banned words? — regardless of whatever message the slogan expressed. As stark as this example might seem, it is directly analogous to the New Jersey consent requirement at issue in Mazo. New Jersey established a category of prohibited words that could not be used in a slogan without consent: the name of any living person and the name of any New Jersey corporation.50 A candidate may make her desired speech only after obtaining the consent of the individual or corporation at issue; if that consent is not forthcoming, then the effective “banned names list” operates to block the candidate’s choice of message.51
The Mazo court would have done better to narrowly read City of Austin as endorsing only the two categories of neutral line-drawing that the case expressly mentioned: (1) “location-based lines” such as the “off-premises distinction” at issue in the case itself, and (2) determining whether speech counts as solicitation.52 The Mazo court could have distinguished the case before it by showing that City of Austin sanctioned only the “zoning” of particular messages, not the outright prohibition of them. Indeed, New Jersey’s consent requirement directly regulates the content of speech in a way that the law at issue in City of Austin did not. Even if City of Austin did treat certain categories of speech differently than others,53 it at least did not permit a total ban on digitizing certain messages. Rather, the Court approved Austin’s attempt at “zoning” speech into different locations. The billboard operators in Austin could digitize any message they wanted — so long as that message was in the right place. The slogan “Bob’s Place: Best Barbecue in Town” could be digitized on Bob’s premises, but not off Bob’s premises. Either way, the speaker could digitize the message he sought to convey: Bob’s Place has the best barbecue in town.54
By contrast, New Jersey’s electoral candidates are not free to communicate any message they want in their slogans. They will inevitably be prohibited from naming or “refer[ring] to” certain people and corporations, since it will not always be possible to obtain the consent of those people or corporations.55 The result is that certain messages will be purged from this form of political discourse altogether. For example, New Jersey prohibited McCormick from criticizing a major political figure in her slogan because she did not obtain his consent to be named.56 Likewise, a politician seeking to use the slogan “Not Me. Us.” may believe that those words are the most effective way for her to convey a message of unity with her constituents.57 But here, McCormick was prevented from using exactly that phrase, since it “referred to” an existing corporation.58 The Third Circuit blessed as content-neutral New Jersey’s attempt to regulate which words and messages are allowed to feature in political speech — an attempt that should have been struck down as an “obvious content-based inquiry.”59
Mazo’s content-neutrality holding is perhaps a foreseeable side effect of City of Austin’s retreat from the Court’s previous bright-line rule for content neutrality. In 2015, Reed v. Gilbert held that a law is facially content based if it “applies to particular speech because of the topic discussed or the idea or message expressed.”60 The sign ordinance at issue in that case triggered strict scrutiny, since it “require[d] Town officials to determine” certain facts about the purpose of the speech after reading it — an “obvious content-based inquiry.”61 Lower courts, including the Fifth Circuit in City of Austin, distilled from Reed that if a law requires a government official to read the speech’s message in order to implement a restriction, then that alone renders the restriction content based.62
But City of Austin retreated from that clear rule. Like the law invalidated in Reed, Austin’s sign ordinance clearly required enforcement officials to read the sign’s message in order to determine whether to apply the on-/off-premises restriction.63 However, the Court rejected the Fifth Circuit’s holding that a law is content-based whenever it requires an enforcement official to examine its message.64 Aside from the two examples regarding sign ordinances and solicitation, the Court gave little guidance for how lower courts should decide what counts as “neutral” in future cases. Indeed, the dissent worried that the Court did not sufficiently explain how lower courts should determine what counts as a “sufficiently substantive or specific content-based classification.”65 Introducing these exceptions to the previous rule made it possible for lower courts to adopt broad interpretations of the content-neutrality principle. Mazo appears to vindicate the concern that departing from Reed may have jeopardized First Amendment freedoms.
The Mazo court should have narrowly read City of Austin as establishing two exceptions to Reed: examination of speech is permissible only to (1) “distinguish based on location”66 or (2) determine whether the speech qualifies as solicitation.67 That would have preserved as much of Reed’s rule-like clarity as possible.68 Such clarity is particularly important in this area of the law, where censorship is often subtle and sometimes operates even when the government has innocent motives.69 But the extrinsic features test threatens to swallow the content-based versus content-neutral distinction altogether. To prevent additional expansions of City of Austin, the Supreme Court should clarify its limits in a future case.