Tobacco regulations: bans on manufacturers' speech

Part II: Bans on advertisers’ own speech

The court—now Judge Clay is back to speaking for the unanimous panel again—turned to restrictions on speech about “modified risk” tobacco products. Under the Act, people may only market a product as “modified risk” if the FDA determines that the product, as actually used by consumers, will significantly reduce harm/risks to individual tobacco users and benefit the health of the population as a whole. A “modified risk” product is a tobacco product whose label or advertising represents, implicitly or explicitly, (1) that it presents a lower risk or is less harmful than other tobacco products on the market, (2) that it or its smoke contains a reduced level/reduces exposure to a substance; (3) or that it or its smoke is free of a substance. A “modified risk” product is also a tobacco product that uses “light,” “mild,” or “low” or similar descriptors, or one as to which the manufacturer “has taken any action directed to consumers through the media or otherwise . . . respecting the product that would be reasonably expected to result in consumers believing” (1)-(3) above. Plaintiffs argued that this was overbroad because it would suppress their participation in the public health debate over harm reduction, even in scientific symposia, press releases, or news programming, and that they were being subjected to impermissible viewpoint discrimination. The district court agreed with them.

The court of appeals found that the premarket approval requirement was a prophylactic regulation of commercial speech, and thus not subject to traditional prior restraint analysis. While plaintiffs argued that their commercial speech was inextricably intertwined with political speech, but the court disagreed. The Act only applied to products where the labeling or advertising made particular health claims or used key words, or where the manufacturer took “any action directed to consumers through the media or otherwise . . . respecting the product.” There was no reason to think the Act touched on plaintiffs’ noncommercial speech in the public health context, including scientific symposia, regulatory press releases, or news programming, or to think that the Act would apply when manufacturers limited their speech to discussions of generic product categories such as smoke-free tobacco. Just because an ad is linked to a public debate doesn’t make it protected the way noncommercial speech is.

Even so, plaintiffs argued, they were subject to unconstitutional viewpoint discrimination because government agencies, ideological anti-tobacco organizations, and commercial manufacturers of tobacco-cessation products were “all free to publicly denigrate the relative health risks of Plaintiffs’ products” but plaintiffs couldn’t respond without receiving FDA preapproval. (I talk about why this argument shouldn’t work in Fighting Freestyle.) The court was unconvinced, because the government can ban commercial speech that’s more likely to deceive than to inform. Plaintiffs can respond to claims that modified risk products are unhealthy “if and when the veracity of the health claims they make in response have been demonstrated.”

Plaintiffs argued that this portion of the Act covered truthful, nonmisleading speech because it barred them from making claims about individual harm reduction even if they unambiguously disavowed any population-wide health benefit. It also barred them from marketing tobacco products as additive-free “to appeal to naturalists and smokers who prefer organic products, even when such preferences are not linked to perceptions of health benefit.” Moreover, though the Act allows products to be marketed as “smokeless” or “smoke free,” it barred manufacturers from “explain[ing] the relative health significance” of those designations. Furthermore, these restrictions allegedly didn’t materially advance the government’s interests, nor were they narrowly tailored.

In a potentially significant move, the court accepted plaintiffs’ argument that there’s no difference between a law that bars promotional speech and one that renders a product’s sale illegal based on speech used to promote it. This would seem to imply that the entire FDA preapproval edifice, which has precisely this structure (a drug is misbranded if it’s advertised for an unapproved purpose) must pass Central Hudson review—perhaps in each individual instance, though the court hedged its bets by parenthetically referring to a previous case as having “distinguish[ed]” normal FDA regulation.

The government’s main defense was that the industry’s speech about modified risk products was, and had historically been, misleading.  But Central Hudson still applied because “the provision undoubtedly ensnares some speech which may be completely truthful and nonmisleading.” Thus, the court turned to whether the government’s interest was substantial. This interest was the risk of fraudulent claims. (If all prophylactic antideception measures have to pass Central Hudson despite being enacted to prevent deception, what happens when the legislature finds that, for example, “organic” can only be used to describe food produced in a particular way?  In other words, this formulation of Central Hudson requires a very clear idea of how much deference a court should be willing to give the legislature on its factual premises.) A paternalistic assumption that the public will use truthful, nonmisleading information unwisely can’t justify its suppression, according to 44 Liquormart, even if the result is bad choices. In light of the industry’s history of deception, though, the government had shown a significant problem of misleading ads and marketing tactics, which had, for example, raised the market share of “low tar” cigarettes—which are no better for health—from 2% in 1967 to 81.9% in 1998. Query: So does this mean that paternalism is justified if, instead of an “assumption,” the government has actual evidence that people use information unwisely? Probably not under current doctrine, but then we shouldn’t really be talking as if the problem were paternalistic assumptions.

The premarketing approval requirement had the appropriate fit to further the substantial interest in preventing deception. It covered only commercial speech. The contention that plaintiffs couldn’t market “additive-free” products to “naturalists and smokers who prefer organic products, even when such preferences are not linked to perceptions of health benefit,” was unpersuasive because plaintiffs presented no evidence that such people actually exist. “On the contrary, we may safely presume that naturalists and those who subscribe to organic products do not engage in unmotivated or arbitrary behavior—common sense dictates the conclusion that they prefer such products precisely because they believe that natural and organic products confer health advantages over conventional products.” Note: While the court cited evidence on this with respect to smokers, it’s also possible to believe that organic production is better for the environment even though consumption of organic products may not confer health benefits.

Moreover, the requirement to show harm reduction at both the individual and population levels survived Central Hudson. “A claim that a product is less risky if it reduces harm to an individual, when that harm is externalized to others, is inherently misleading.” If marketing a product as “modified risk” raises the aggregate number of people, especially kids, who use tobacco by making them think the product is relatively safe, then that impairs the government’s compelling interest in reducing juvenile tobacco use.

Finally, plaintiffs’ suggested less restrictive alternatives—mandatory disclaimers, post-marketing review of ads, enhanced enforcement of fraud laws, and government public advertising campaigns—didn’t convince the court that there was a problem with fit. “Plaintiffs would have us believe that the government has only recently come to recognize the harms associated with false tobacco safety claims, and that the [Act] is the government’s first up at bat. But on this issue, the government is at play in the major leagues, and the alternatives suggested by Plaintiffs have already been tried and found wanting.” Congress expressly concluded that disclaimers wouldn’t work, because of evidence that consumers have misinterpreted ads claiming harm reduction even with disclaimers. Nor did the costs of correcting these misimpressions have to be externalized from the tobacco industry to the government and taxpayers. “And, although the idea of post-market review of deceptive claims may be appealing to the tobacco industry, the government has made a reasonable determination that, in the context of a deadly and highly addictive product, it would be a virtual impossibility to unring the bell of misinformation after it has been rung.”

Restrictions on the use of color and imagery in most tobacco ads, by contrast, failed Central Hudson scrutiny.  The Act mandated black text on a white background for any labeling or advertising, except for circumstances unlikely to reach juveniles: places where vending machines and self-service displays were permitted if the advertising wasn’t itself visible from outside and adult publications with 15% or less and 2 million or fewer of under-18 readers.  The district court struck this down as overbroad because it covered many innocuous images and colors (“images that teach adult consumers how to use novel tobacco products, images that merely identify products and producers, and colors that communicate information about the nature of a product, at least where such colors and images have no special appeal to youth”).

The government argued that the tobacco industry’s history of targeting juveniles through colorful and graphic advertising justified the breadth of the restriction.  As it framed the problem, “[t]he industry’s campaign to attract minors is not waged with tools of rational persuasion that invoke the ‘merits’ of taking up tobacco use. Instead, the industry relies on peripheral cues and irrational associations to distract would-be users from the fact that tobacco products are lethal and addictive.”  This strategy was more effective with juveniles than with adults.  The government contended that the ban was narrowly tailored because it exempted adult publication ads and didn’t apply to packaging, and allowed manufacturers to communicate via text, and because narrower bans would be subject to the industry’s history of evasion.

Plaintiffs argued that the ban prevented them from “attracting attention and differentiating their products in the retail environment” and communicating with adults, including by advertising their products’ packaging or logos and colors.  The ban covered direct mail to adults, some magazines primarily directed to adults, most retailers that sell tobacco, and taverns that allow juveniles “to enter for meals with their parents or during restricted times.”  Moreover, the burden of having competent and reliable survey evidence of readership, as required by law to advertise in an adult publication, was too heavy since a readership survey would generally be more expensive than the ad itself.

The court began with the proposition that truthful speech (or, in this case, nonfalsifiable speech, since we’re not talking about truthful communication for almost all of the uses of imagery involved here) is entitled to First Amendment protection.   IMS v. Sorrell: “the State may not seek to remove a popular but disfavored product from the marketplace by prohibiting truthful, nonmisleading advertisements that contain impressive endorsements or catchy jingles. That the State finds expression too persuasive does not permit it to quiet the speech or to burden its messengers.”  Thus, one implication of this decision is First Amendment protection for trademarks themselves as such, which contradicts prior Supreme Court precedent, Friedman v. Rogers, 440 U.S. 1 (1979), but I think is likely to find favor with the present Court.

The government’s argument was that colorful advertising distracted potential users from the harms of tobacco use and created positive lifestyle associations that overrode risk information.  This was not a claim of deception.  Note that the FTC and the FDA have long taken the position that distracting people from disclosures/disclaimers with color and sound can be deceptive; perhaps the court here would agree that specific uses of images and sound can meet that standard, but if that's true why isn't the evidence that consumers don't notice the current disclosures evidence that current uses of color and images are distrcting them?  

Plaintiffs rejoined that accepting the government's premise would allow bans of attractive ads for all age-restricted products, from movies to cars.  The court wasn’t entirely convinced—there are safe ways to consume alcohol, movies, lottery tickets, and cars, but there is no non-harmful way to consume tobacco. 

But it was convinced enough.  Ads always try to create positive associations.  “Perfume and cologne do not make people more beautiful, chewing gum does not make them more athletic, coffee does not make them more intelligent or urbane. By the same token, though the government would have us believe otherwise, using tobacco does not necessarily preclude a person from mountain biking, playing games, or engaging in romantic relationships.”  A broad prophylactic rule against images was unjustified.  From Zauderer: “We are not persuaded that identifying deceptive or manipulative uses of visual media in advertising is so intrinsically burdensome that the State is entitled to forgo that task in favor of the more convenient but far more restrictive alternative of a blanket ban on the use of illustrations . . . .”

As nondeceptive uses of color and imagery, the court identified informational ads, ads intended to “reinforce” consumer preferences by simply showing the package, and uses of color that grab attention in a crowded marketplace, “letting consumers know that their preferred brand or product is available at a particular retailer.” In addition, “there are surely certain color graphic tobacco ads that have nominal to zero appeal to the youth market.”  (Really?  Which ones?)  Suppressing them all imposed an undue burden.

The government’s argument that color was used to convey mood—red for passion and power, green for harmony and health—and circumvent other ad restrictions.  But that wasn’t enough to justify a blanket ban.  “Packaging shape, product shape and color, display location, and any number of other factors may also convey meaning through association.”  (Note the implication, about to come to fruition in the next section, that conveying meaning makes something speech.)  Congress could have exempted innocuous images and colors that merely teach consumers how to use novel tobacco products, identification-promoting images, and information-providing colors, at least where they had no special appeal to juveniles.  This would involve more difficulty in implementation—the government noted the controversy surrounding the “Camel Farm” ads, which also produced an interesting right of publicity case—but that’s required by the First Amendment.
Finally, the plaintiffs lost their challenge to bans on express or implied claims that a product is safe or less harmful due to FDA regulation or compliance with FDA standards.  They argued that a ban on any consumer-directed speech that conveys that the FDA is achieving its objectives was overbroad.  The district court found that the ban went beyond commercial speech and failed strict scrutiny, but the court of appeals disagreed.  Manufacturers were banned from making such claims in consumer-directed communications, but journalists, doctors, scientists and politicians weren’t banned from making noncommercial statements, so this was an ordinary commercial speech regulation.  The ban served the legitimate interest of keeping manufacturers from confusing consumers about the FDA’s role with respect to tobacco, which was significantly different from its ordinary regulatory role.  For the industry to claim or imply that tobacco products are made safe or less harmful by virtue of FDA regulation would be “inherently misleading and patently false.”  So this ban survived Central Hudson.