BARRON, Circuit Judge.
This case requires us to decide whether an ordinance in the City of Portland, Maine that prohibits standing, sitting, staying, driving, or parking on median strips violates the constitutional guarantee of "the freedom of speech." U.S. Const. Amend. I.
In 2012, Portland's chief of police, Michael Sauschuck, identified an increase in panhandling on traffic medians in the City. Calling this increase a "public safety emergency," he recommended to the Public Safety, Health, and Human Services Committee of the Portland City Council that it adopt an ordinance barring virtually all
The proposed ordinance failed to pass. Nonetheless, concern about panhandling in the City's median strips did not abate. And, in July of 2013, the Council held a public hearing to reconsider the proposed ordinance. This time the City Council unanimously voted to adopt the median ordinance.
Portland City Code § 25-17(b) became effective on August 15, 2013. The ordinance provides that:
The ordinance defines a median strip as "a paved or planted area of [a] public right-of-way, dividing a street or highway into lanes according to the direction of travel." Portland City Code § 25-118. The ordinance does not specify any other features of a median strip — such as its size or its location relative to heavy or fast traffic. Nor does the ordinance restrict presence in the streets themselves.
The City has enforced the ordinance against just five people, in each case for panhandling. The City voluntarily stopped enforcing the ordinance when, on September 24, 2013, three individuals, now appellees — Michael W. Cutting, Wells Staley-Mays, and Alison E. Prior — brought the present action claiming the ordinance restricted their speech in various ways.
The suit contends that Portland's median strip ordinance violates the First Amendment, both as applied and on its face.
The District Court combined a hearing on plaintiffs' motion for a preliminary injunction with a trial on the merits. After receiving testimony and exhibits from the parties, the District Court issued a decision in plaintiffs' favor. See Cutting v. City of Portland, No. 13-cv-359-GZS, 2014 WL 580155 (D.Me. Feb. 12, 2014).
The District Court held that the median strips that the ordinance covered were traditional public fora, like sidewalks or parks. Id. at *7. The District Court then held that the City had adopted an "official interpretation" of the ordinance that excludes campaign signs from the ordinance's reach, thereby allowing signs bearing campaign messages to be posted in median strips but not signs communicating other messages. Id. at *6. On that basis, the District Court found that the ordinance discriminated on the basis of the content of the speech that occurs in the median strips, and so the District Court went on to ask whether the ordinance used the least
The City now appeals. We review the District Court's grant of a permanent injunction for abuse of discretion, its underlying conclusions of law de novo, and any factual findings for clear error. Asociación de Educación Privada de P.R., Inc. v. García-Padilla, 490 F.3d 1, 8 (1st Cir.2007).
We need to address two issues at the outset. The first concerns how to characterize, for First Amendment purposes, the type of places — median strips in Portland — that the ordinance targets. The second concerns whether the ordinance favors the content of certain messages or whether the ordinance instead restricts expression only because of where it occurs and thus without regard to its content.
The parties appear to agree that the City's median strips are what are known for First Amendment purposes as "traditional public fora." Those are places "held in trust for the use of the public ... for purposes of assembly, communicating thoughts between citizens, and discussing public questions." See Hague v. Comm. for Indus. Org., 307 U.S. 496, 515, 59 S.Ct. 954, 83 L.Ed. 1423 (1939). Given the role such places historically have played in fostering public discussion and debate, the government's authority to regulate speech within such places is especially limited. Id. at 515-16, 59 S.Ct. 954.
The classic traditional public fora are parks and sidewalks. Id. The City disputed below whether Portland's median strips qualify as traditional public fora. But the District Court resolved that dispute in favor of the appellees. Cutting, 2014 WL 580155, at *7. The District Court based its decision on the medians' "past uses," explaining that "the City's medians have routinely been the site of protected speech, including political protests, election campaigns by politicians, and solicitations by individuals for charity." Id.
The two circuits that have addressed whether median strips are traditional public fora held similarly, see Warren v. Fairfax Cnty., 196 F.3d 186, 196-97 (4th Cir.1999) (en banc); Satawa v. Macomb Cnty. Road Comm'n, 689 F.3d 506, 520-22 (6th Cir.2012), and the City makes no argument to us that its median strips are not traditional public fora. We thus decide this case on the understanding that, as the District Court found, the people of Portland have used median strips for expressive purposes in much the same way that they have used parks and sidewalks, as any argument to the contrary has been waived.
The City's concession bears on the second threshold issue that we must address: whether this ordinance favors certain types of messages on the basis of their content. A restriction on speech that targets the content of the message conveyed is known as a "content-based" law. And when such a content-based law restricts speech in a traditional public forum, it raises a very serious concern that the government is using its power to tilt public debate in a direction of its choosing, a particularly worrisome form of governmental regulation of free expression. As a result, such a law may be upheld only if that law uses the least speech restrictive means to serve what must be a compelling governmental interest. Globe Newspaper Co. v. Beacon Hill Architectural Comm'n, 100 F.3d 175, 182 (1st Cir.1996).
We thus need to decide whether this ordinance is content-based. For if the ordinance is not content-based, and the ordinance restricts speech without regard to the type of message communicated and only regulates the time, place, or manner of speech, then the ordinance is what is called "content-neutral." See id. at 183. Even though such a law might restrict a greater amount of expression in absolute terms than one that favors certain types of messages over others, it has the virtue of not singling out any idea or topic for favored or un-favored treatment. Thus, the government must show only that a restriction that is content-neutral is "narrowly tailored to serve a significant governmental interest, and that [it] leave[s] open ample alternative channels for communication of the information." Ward v. Rock Against Racism, 491 U.S. 781, 791, 109 S.Ct. 2746, 105 L.Ed.2d 661 (1989) (citations omitted).
Here, the District Court found from the evidence at trial that the City "favors one category of speech, campaign signs, over all others" because the City had adopted an "official interpretation" of the ordinance that exempts the posting of campaign signs from the ordinance's reach. Cutting, 2014 WL 580155, at *6 n. 5, *8-9.
To the extent that the District Court believed that the City's content-based "official interpretation" represented a construction of the ordinance's actual reach, the District Court erred in treating that construction as binding. For while we may read a law in light of the limits set forth in a government's "authoritative[] constru[ction]" of that law if doing so would "render [that law] constitutional," City of Lakewood v. Plain Dealer Publ'g Co., 486 U.S. 750, 770 n. 11, 108 S.Ct. 2138, 100 L.Ed.2d 771 (1988), consistent with the principle of constitutional avoidance, we may not do so to make that law more vulnerable to constitutional challenge, see McCullen v. Coakley, 571 F.3d 167, 178 (1st Cir.2009) (McCullen I) ("[A] state official's interpretation of a statute, even if generally authoritative, cannot render an
To the extent that the District Court meant only that the City had adopted an unwritten policy regarding how the City would implement or enforce the ordinance with respect to campaign signs, the District Court still erred in striking the ordinance down facially as content-based. The principle of constitutional avoidance would counsel against that approach, too. See id. And if the City does have an official policy of enforcing the ordinance that permits the posting of campaign signs but no others in median strips, the proper remedy must target that policy or the enforcement of the ordinance pursuant to the policy. See Hoye v. City of Oakland, 653 F.3d 835, 848-49 (9th Cir. 2011) ("[I]t would make little sense to invalidate a statute that is constitutional as written when only its implementation is defective. Doing so would only require legislative bodies to undertake the pointless exercise of re-enacting laws that were perfectly valid as enacted on the first go around.").
Thus, we set to one side any content-based "official interpretation" that may exist. And, having done so, we conclude that the ordinance restricts speech only on the basis of where such speech takes place. The ordinance does not take aim at — or give special favor to — any type of messages conveyed in such a place because of what the message says. See Reed v. Town of Gilbert, ___ U.S. ___, 135 S.Ct. 2218, 2227, 192 L.Ed.2d 236 (2015) (discussing what is meant by the term "content-based"). As a result, we must decide the constitutionality of the ordinance on the understanding that it is content-neutral.
Because the District Court deemed the law content-based, the District Court applied strict scrutiny. It thus did not decide whether the ordinance is narrowly tailored to serve a significant governmental interest. But no party asks us to remand the case for the District Court to rule on narrow tailoring in the first instance.
The City does note that we may do so, but the City makes no argument that the record is insufficient for us to decide the question, or that the City would be prejudiced were it not permitted to develop the record further. The parties were in agreement before trial that the ordinance is content-neutral and that the crucial question was whether the ordinance survives narrow tailoring, and the parties prepared their cases accordingly. We thus address the question of narrow tailoring despite the fact that the District Court has not passed on it.
The City asserts that the ordinance is narrowly tailored to serve the City's interest in protecting public safety. And we recognize that such an interest is a legitimate and significant one, as the Court most recently recognized in McCullen. 134 S.Ct. at 2535. But before examining whether the ordinance is narrowly tailored to that interest, we need to say a little more about the doctrine of narrow tailoring.
Outside the First Amendment context, the Court has stated that a litigant who brings a facial challenge to a statute must establish "that [there is] no set of circumstances ... under which [the regulation] would be valid." United States v. Salerno, 481 U.S. 739, 745, 107 S.Ct. 2095, 95 L.Ed.2d 697 (1987). But a content-neutral restriction on speech in a traditional public forum is facially unconstitutional if it does not survive the narrow tailoring inquiry, even though that ordinance might seem to have a number of legitimate applications. See McCullen, 134 S.Ct. 2518 (striking down content-neutral, sidewalk buffer-zone law facially on narrow tailoring grounds).
The reason is that the First Amendment interest in promoting free speech is so great that the government may not pass unnecessarily sweeping restrictions on speech and then force those burdened by them to challenge each problematic application. Thus, the seemingly tailored aspects of an untailored restriction on speech in a traditional public forum do not automatically save such a restriction from facial challenge. Id. at 2534 ("Where certain speech is associated with particular problems, silencing the speech is sometimes the path of least resistance. But by demanding a close fit between ends and means, the tailoring requirement prevents the government from too readily sacrificing speech for efficiency." (internal quotation marks and brackets omitted)); see Richard H. Fallon, Jr., As-Applied and Facial Challenges and Third-Party Standing, 113 Harv. L.Rev. 1321, 1354 (2000) (making a similar argument).
That said, the narrow tailoring doctrine does not require perfect tailoring. The doctrine requires only that a challenged speech restriction not burden "substantially" more speech than is necessary to further the government's interest. See McGuire v. Reilly, 260 F.3d 36, 48 (1st Cir.2001) (citing Ward, 491 U.S. at 799, 109 S.Ct. 2746).
We thus start our analysis by describing just how burdensome this ordinance is. We then consider whether the City's interest in protecting public safety justifies a restriction that is so burdensome. Finally, we consider whether there were less speech restrictive measures that the City bypassed in opting for this ban. In the end, we conclude that this ordinance cannot
There is no doubt that the ordinance imposes "serious burdens" on speech. See McCullen, 134 S.Ct. at 2535.
It is also hard to imagine a median strip ordinance that could encompass more spaces within its definition. The ordinance restricts speech in all median strips in the entire City of Portland. And the actual "strips" range widely in terms of their size and character. In this way, the ordinance applies without regard to whether the term "strip" really is a plausible descriptor when applied to the median in question.
To be sure, the ordinance applies to some very narrow strips of raised concrete between two lanes of traffic. These strips include the "eight-inch" strips that the City's police chief raised concerns about before the City Council. But we know, from the trial record, that not all of the median strips in Portland are eight inches wide. In fact, Chief Sauschuck acknowledged that most of the City's medians did not meet that description.
The record also shows the ordinance encompasses — by virtue of the definition used — some considerably larger medians. The ordinance applies, for example, to medians that are roughly eight feet wide, and even to the grassy expanse on Franklin Street, which runs for several blocks and is as wide as fifty feet in various places. The ordinance also applies to Boothby Square, a wide, raised grassy median that contains a park bench, and possibly other medians that are sufficiently wide to permit pedestrians to stay far away from traffic, as appellees argued below.
What is more, the ordinance's broad definition of the term "median strip" does not purport to consider other important factors, such as pedestrian and vehicle traffic patterns on the surrounding sidewalks and roadways. And so it is expansive in that way, too.
To be sure, plaintiffs may be able to engage in their speech in places other than medians — such as parks and sidewalks. But the fact that there are other places where plaintiffs may engage in their expressive activity "misses the point." McCullen, 134 S.Ct. at 2536. A flat ban on speech in a particular forum — like the median ordinance at issue here — can fail narrow tailoring even if it leaves open other channels for plaintiffs to engage in their expressive activity. And, in addition, plaintiffs "believe they can accomplish [their] objective" best if they are permitted to speak from traffic medians. Id. (considering
A protestor standing on a median with a double-sided sign may — as appellee Wells Staley-Mays asserts, based on his own experience — reach more people than he can standing on a sidewalk. And appellee Michael Cutting testified that there are "more interactions [with people] and acknowledgements on the median than from the sidewalk." According to Cutting, sidewalks also present obstacles to expression that medians do not: cars parked along sidewalks block drivers' views of him; storefronts and signs distract passersby from his message; and shop owners who line the sidewalks sometimes become agitated with his protest activities. In fact, appellee Alison Prior, who uses medians to panhandle, finds sidewalks so useless for her purposes that she now takes a bus to a different town in order to panhandle from medians. Similarly, parks may not enable protesters like Cutting and Staley-Mays, and panhandlers like Prior, to be seen by people constantly moving past them in both directions.
Notwithstanding the serious burdens on speech the ordinance imposes, we still need to decide whether the City's interest in public safety justifies such an all-encompassing ban. In the City's view, the dangers — to passersby as well as to those in the median strips — are sufficiently present, no matter the activity (expressive or not) taking place on the median strip, and no matter the nature of the median strip on which such activity occurs. "[T]here simply is no way to abate the City's significant safety concern," the City says, "except for an outright ban." For that reason, the City contends, the ordinance does not ban substantially more speech than necessary, even though it bans nearly every activity on every median in the City. But neither the City's interest in protecting people in the streets nor its interest in protecting people on medians holds up as a justification for the ordinance.
We start with the City's interest in eliminating the danger to drivers and other users of the streets, which was the City's focus when it enacted the bar to persons lingering in median strips. As described by Chief Sauschuck, who spearheaded the effort to pass the ordinance, that danger does not apply to all median strips in Portland. Instead, the record shows that the danger the City identified applies to only a limited number of median strips.
At trial, Chief Sauschuck agreed that "most of th[e] incidents" that citizens had called in "occurred at a handful of intersections" across the City. A map of the incidents, created by Chief Sauschuck and submitted as an exhibit at trial by plaintiffs, confirmed that they were indeed clustered around a few medians. And while Chief Sauschuck briefed the City Council on the dangers posed by individuals standing on "eight inch medians in the City of Portland," he conceded at trial that he did not know "exact measurements for any medians," that the eight-inch number he had used was "a pure guesstimate on [his] part," and that — at least based on the exhibits shown to him at trial — "probably most of the medians in town [are] wider than eight inches."
Absent evidence about whether the City's other median strips present the same or a similar danger, we have no basis for concluding that a substantial number of them do. The ordinance is thus geographically over-inclusive with respect to the City's concern that people lingering in all of the City's median strips — no matter
Moreover, the danger to drivers and other users of the streets that the City identified when it passed the ordinance was tied to concerns about disruptive and inattentive individuals on median strips. The record shows that the City was worried such median strip users were intentionally leaving the median strips to enter the roadway to accost passersby or stumbling — often under the influence of alcohol or drugs — into the roadway. Yet the trial record shows that plenty of people engage in expressive activities on median strips that the ordinance would ban but that do not pose the same threat to public safety that the City had identified.
For example, Chief Sauschuck conceded that he could not "recount any public safety problem with" either "individuals holding political campaign signs on the median strip" or "political activists holding issue related signs on median strips."
For these reasons, the risk to passersby posed by allowing people to linger in median strips does not justify banning as much speech, in as many places, as the City chose to ban. Given this record, that risk is simply not posed in many of the medians or by much of the expressive activity to which the ordinance applies.
Still, the City argues that the ordinance promotes public safety in another way, by ensuring that people are not on median strips and thus are not positioned to be hit by passing cars. With respect to this goal, the City says, quoting News & Sun-Sentinel Co. v. Cox, 702 F.Supp. 891, 900 (S.D.Fla.1988), there is no need for either "towering intellect []or an expensive `expert' study to conclude that mixing pedestrians and temporarily stopped motor vehicles in the same space at the same time is dangerous." Id. (internal quotation marks omitted). The City thus argues that, as a matter of common sense, there is a sufficient public safety need to impose this broad ban because, as a City official argued at trial, "motor vehicles are deadly weapons" and if a vehicle travels from a
To bolster this assertion, the City does offer two pieces of evidence. The City points to fourteen requests by the public that the City replace damaged signs in medians. And the City references three reports, over a four-year period, of cars veering off roads and into median strips, which the City found only after "really cull[ing] through the system looking for median related accident reports."
But the City's evidence is of limited value. It is not clear that the fourteen damaged signs reported by the public were in fact damaged by vehicles. The three reports of cars entering medians do prove the obvious proposition that cars sometimes veer off roads and hit medians. But one accident occurred at 1:10 a.m. Another occurred in treacherous winter conditions. And none actually involved pedestrians. The City even conceded at trial that, since 2008, there has been only one incident in which a driver hit a person on a median strip. In that incident, however, the driver hit a cyclist who was using the median strip to cross the road-an activity not prohibited by the ordinance.
Moreover, the City does not contest that Portland's median strips, as a group, are traditional public fora — that is, that they are places that "time out of mind" have been "held in trust" for the public's use for assembly, communicating ideas, and discussing public questions. Hague, 307 U.S. at 515, 59 S.Ct. 954. As such, Portland's medians would seem to be — as a class — presumptively fit for the very activities that the City now contends are obviously dangerous. In fact, at trial, Michael Bobinsky, Director of Portland's Public Services Department, conceded that "there are medians in the city that are safe," and that there were at least "a few" that were "basically ... island refuge[s]."
In sum, we must consider the City's claim that it is obvious that all medians are unsafe in light of the limited number of median strips with which the City expressed concern when it passed the ordinance, and the wide array of median strips that are subject to the ban. And, under that calculus, the ordinance is so sweeping that it does ban substantially more speech than necessary to serve the City's interest in preventing people on medians from being hit by drivers. Or, at the least, we cannot conclude otherwise given that the City has not shown that the "island refuges" that it concedes do exist are so few or so unattractive that only an insubstantial amount of expressive activity would occur there.
There remains the question whether there are any less restrictive means of accomplishing the City's concededly legitimate purpose of protecting public safety than the complete ban that the City chose to impose. But the City did not try — or adequately explain why it did not try — other, less speech restrictive means of addressing the safety concerns it identified. See McCullen, 134 S.Ct. at 2540 ("To meet the requirement of narrow tailoring, the government must demonstrate that alternative measures that burden substantially less speech would fail to achieve the government's interests....").
The City says that existing state and local laws that prohibit disruptive activity in roadways, such as prohibitions on obstruction of traffic, disorderly conduct, and abusive solicitation, "simply do not provide an adequate tool" because they are "reactive, rather than proactive, and require a police officer to directly observe the illegal behavior ... before taking action." For example, the City explains, "obstruction of
But the limitations in such laws do not suffice to show the need for the sweeping ban that the City chose. Such a ban is obviously more efficient, but efficiency is not always a sufficient justification for the most restrictive option. See id. at 2534-35. A more modest potential solution not addressed by the City, for example, would have been to strike the warning requirements that those laws contain, so that those laws would not be so reactive in their operation.
The City might also have considered an ordinance that focused more directly on the dangerous activities that were the source of initial concern, such as ordinances directed at public intoxication or belligerent behavior. Or it might have considered limiting activity on medians only at night, when the dark makes it more difficult for drivers to see, or during hazardous weather conditions, when slick roads increase the chances that a car will skid into a median. Nor did the City show that it contemplated and rejected as ineffective an ordinance limited to the few medians in Portland where the City had identified safety hazards in the past, an ordinance limited to the smallest or most dangerous medians, or even an ordinance with an exception for certain large park-like spaces — like Boothby Square and the wide median on Franklin Street — where the City had not observed safety hazards but which are especially attractive sites for expressive activity.
The City points out that it declined to extend the ordinance to "some intersection sidewalks," despite requests to do so from the public, "because the evidence did not indicate that the public safety danger was great in those locations." But the City's willingness to make that sensible, evidence-based limitation only underscores the point that tailoring is possible but was not fully considered, as Councilor Suslovic himself conceded at trial with respect to a number of these alternatives.
"In short, the [City] has not shown that it seriously undertook to address the problem with less intrusive tools readily available to it." Id. at 2539. Instead, it "sacrific[ed] speech for efficiency," and, in doing so, failed to observe the "close fit between ends and means" that narrow tailoring demands. Id. at 2534 (internal quotation marks omitted).
The City may have been motivated by a perfectly understandable desire to protect the public from the dangers posed by people lingering in median strips. But the City chose too sweeping a means of doing so, given the First Amendment interest in protecting the public's right to freedom of
