A Legislative Response to 303 Creative

abstract. After the Supreme Court’s decision in 303 Creative LLC v. Elenis, public accommodations may discriminate in the provision of expressive services or products. But states are not without options to limit discriminatory impact while adhering to the Court’s ruling. This Essay argues that state legislatures can and should enact implied warranties of nondiscrimination. Such implied warranties would function as default rules, under which sellers implicitly promise not to discriminate unless they opt out. The 303 Creative Court struck down the Colorado Anti-Discrimination Act because the nondiscrimination law was mandatory in nature. Legislatures could respond by enacting laws making the nondiscrimination rule a disclaimable default with respect to the relevant First Amendment-protected goods and services, thereby vesting the ultimate decision to discriminate in expressive sellers while also giving presumptive effect to the majoritarian preference for nondiscrimination. Courts need not invalidate the application of the public-accommodations statute’s nondiscrimination norm to expressive sellers if those sellers have the power to opt out, because opting out relieves those sellers of any obligation to express ideas contrary to their own strongly held principles. The Essay proposes a simple altering rule—the process to displace the default rule—whereby sellers may disclaim the warranty by sending a brief statement to the state. The state could then compile such disclaimers into a central database. This database would not only allow the public to make informed consumer choices, but also mitigate the dignity-depleting impact of discrimination at the point of sale. The Essay considers and rejects a potential objection to our proposed default and altering rule: that the process for disclaiming the warranty unconstitutionally compels seller speech.

Introduction

In a momentous conclusion to its 2022-2023 Term, the Supreme Court struck down the application of the Colorado Anti-Discrimination Act to the expressive service of website design in 303 Creative LLC v. Elenis.1 Justice Gorsuch reasoned, over an impassioned dissent from Justice Sotomayor,2 that requiring a business to create websites celebrating the marriages of same-sex couples would force the seller to express a message with which it disagreed.3 The reasoning of 303 Creative is broad: after this opinion, public-facing businesses across the country may legally refuse to sell expressive services or products based upon not just a customer’s sexual orientation, but also their race, gender, and creed.4

The stakes are high, with the “equal dignity” of many market participants in jeopardy.5 As Justice Goldberg argued so powerfully in Heart of Atlanta Motel, Inc. v. United States, the “fundamental object” of public-accommodations law is to “to vindicate ‘the deprivation of personal dignity that surely accompanies denials of equal access to public establishments.’”6 Because the Court’s decision in 303 Creative frustrates this “fundamental object,” Justice Sotomayor warns, it not only harms LGBTQ+ consumers, but also “threatens to balkanize the market and to allow the exclusion of other groups from many services.”7

One might be tempted to believe that the State of Colorado can do nothing to respond: the Supreme Court has spoken, and the Constitution trumps state law. But the Colorado legislature is still free to deploy a well-used tool of contract law—an implied warranty—to promote equal access to public accommodations in the state. Colorado and other states can and should pass an implied warranty of nondiscrimination.

Implied warranties are a cornerstone of both contract and property law. The Uniform Commercial Code (UCC), for example, includes an implied-warranty provision stating that, absent contrary indication, goods are presumptively “fit for the ordinary purposes for which such goods are used.”8 Contracts for construction of new homes likewise typically include an implied warranty that the premises are habitable.9 The warranty is “implied” in that it need not be explicitly set forth in a contract to be enforceable; it exists in the background of all contracting to protect consumers, who often lack full information about vendors and may fail to negotiate explicitly for important protections and promises they value.

An implied warranty of nondiscrimination would be a default10 promise by a seller not to discriminate against any of the protected classes of consumers set forth in the statute—even with respect to the provision of expressive and bespoke services, like those at issue in 303 Creative. Like other legal defaults, it would allow sellers to opt out. Sellers would not be forced to offer expressive products or services with which they disagree, because they could disclaim the warranty. The implied warranty of nondiscrimination would run alongside public-accommodations statutes, and opting out would only be permitted with respect to expressive products and services. A wedding-cake seller, for instance, would be permitted to opt out with respect to bespoke cakes specially crafted for specific occasions, but would not be permitted to opt out with respect to off-the-shelf products routinely offered to the general market of consumers. This proposal is thus consistent with existing public-accommodations statutes, and it would do nothing to displace or modify their impact on the vast majority of public-facing businesses.

Expressive sellers could disclaim by notifying the state that they reserve the right to discriminate on the basis of certain protected characteristics of consumers regarding specific types of expressive services or products. The state, in turn, would compile these declarations into a searchable database that would inform the public about such business reservations.

Expressive businesses that fail to opt out of the warranty would be subject to the same remedial consequences for discrimination that apply to nonexpressive products and services under the public-accommodations statute.11 Instead of suing under the public-accommodations statute, however, customers subjected to discrimination could sue for the seller’s breach of its implied promise not to discriminate.12

A disclaimable warranty of nondiscrimination would have three benefits. First, by requiring businesses to disclaim in advance of denying services, such a policy would allow the state to investigate proactively whether particular products or services were sufficiently expressive to fall within the ambit of the Supreme Court’s 303 Creative rule. Justice Gorsuch acknowledged that “determining what qualifies as expressive activity protected by the First Amendment can sometimes raise difficult questions.”13 Preemptive notification to the state would enable legislatures and courts to flesh out the contours of protected expressive activity before legal challenges would commence.

Second, the state registry would mitigate the dignity-depleting impact of discrimination by warning potential victims before they attempt to contract with a business that discriminates. Justice Sotomayor vividly described this danger in her dissent:

Discrimination is not simply dollars and cents, hamburgers and movies; it is the humiliation, frustration, and embarrassment that a person must surely feel when he is told that he is unacceptable as a member of the public because of his [social identity]. It is equally the inability to explain to a child that regardless of education, civility, courtesy, and morality he will be denied the right to enjoy equal treatment.14

Part of the Colorado statute that the Supreme Court struck down is a “Communication Clause.”15 This clause prohibits a public accommodation from indicating that a person will be denied “the full and equal enjoyment” of services based on a protected classification.16 Our proposal flips the statute’s Communication Clause on its head, instead requiring expressive firms to disclose such a warning if they want to opt out of the statute and reserve the right to deny provision of certain expressive products and services. Moreover, it requires the disclosure of this reservation in a centralized, searchable, online location, away from the public square or the site of the commercial activity, where the message could be as harmful as the actions it describes. Potential customers would be able to search the database, sparing themselves the ignominy of in-person service denial.17

Finally, the state registry would facilitate informed association, an independent First Amendment value.18 Some customers would prefer to patronize businesses that stand behind a warranty of nondiscrimination, while others might gravitate towards sellers that disclaim the warranty. The aggregate outcome—a mix of boycotts and buycotts—would likely usher in greater accessibility to public accommodations for marginalized groups. We also provide examples below to show why discriminatory preferences are likely to be over-accommodated in a competitive market—so that buyers are likely to have ample options for both disclaiming and nondisclaiming sellers.19

Because sellers who want to opt out of potential liability would be required to declare their stance to the state publicly, some may wonder whether the implied warranty itself creates a constitutional “compelled speech” problem.20 But the public declaration we propose falls squarely within the bounds of commercial-speech regulation. No specific state-endorsed ideology is imposed. Indeed, opting out of the warranty would involve speech the state does not prefer.21 Businesses merely unveil and thereby clarify a detail of their contractual commitment. The Supreme Court, in its landmark decision in Zauderer v. Office of Disciplinary Counsel, held that requiring “purely factual and uncontroversial” disclosures is permissible if the disclosures are “reasonably related to the State’s interest in preventing deception of consumers.”22 This disclosure disclaiming any responsibility for certain types of discrimination simply informs consumers—factually—of the legal attributes of their transactions. Just as sellers have been required to disclaim expressly various UCC warranties for years, sellers transacting under our proposed regime would be analogously required to disclaim expressly a nondiscrimination warranty.23

Compelled contractual speech is an inevitable aspect of contracting. Because of the Supreme Court’s 303 Creative decision, if sellers of expressive services want to make clear to consumers that they are not reserving the right to sell on a discriminatory basis, they must now include explicit nondiscrimination promises, while those who reserve the right to discriminate may remain silent. Under an implied warranty of nondiscrimination, the reverse would be true.

The rest of this Essay is divided into three Parts. Part I addresses what should be the default presumption with regard to whether a retailer of expressive products or services promises to serve the public without discriminating on the basis of protected characteristics. Part II discusses various “altering” rules—the requirements for opting out of the default—and argues that expressive retailers should be able to opt out of an implied warranty by sending a straightforward notice to the Secretary of State reserving the right to discriminate in specific types of contracts with regard to certain otherwise protected groups. Part III then defends the constitutionality of our proposal by addressing the concern that the implied warranty compels speech by some commercial entities.

I. choosing the default

In his 1992 book, Forbidden Grounds, Richard A. Epstein argues that our civil-rights laws inappropriately displaced a bedrock principle of commercial activity: freedom of contract.24 Epstein’s central claim is that freedom of contract never had a chance because prior to 1964, state law required employers to discriminate against Black people; after 1964, Title VII prohibited discrimination.25 Epstein thought that if state governments had protected businesses that were willing to hire and sell to minorities, unprejudiced entrepreneurs would have efficiently served the needs of the Black community.26 His argument was not that all bigoted employers would be driven from the market, but that nonbigoted employers would enter the market and provide minorities with sufficient opportunities so “it would be as if bigotry didn’t exist.”27

A central problem with Epstein’s argument is the non sequitur between his criticism of the civil-rights statutes and his remedy. Epstein’s criticism is that our civil-rights laws are mandatory rules limiting freedom of contract. His remedy—repealing all civil-rights statutes—is a non sequitur because doing so goes beyond eliminating the statutes’ mandatory nature.28 Even if one accepted Epstein’s libertarian argument that it is wrong for civil-rights law to displace freedom of contract, one would need a separate argument to explain why a default that permits discrimination is preferable to one that forbids discrimination. The former requires businesses to “opt in” to nondiscrimination if they want to signal credibly to the public their willingness to serve all consumers equally, while the latter requires businesses to “opt out” if they want to signal to consumers that they might discriminate against some consumers.

A parallel disconnect is at play in 303 Creative. Justice Gorsuch, writing for the majority, was principally concerned that, with regard to expressive products and services, the mandatory nature of Colorado’s public-accommodations law “compel[s] an individual to create speech she does not believe.”29 But the Court’s holding that struck down the nondiscrimination law when applied to such contexts does more than simply nullify the mandatory nature of the law and hence does not exemplify judicial restraint. The decision also changes the presumptive or default coverage of the civil-rights law. Accepting the Court’s conclusion that sellers of expressive goods and services cannot be forced to express views with which they disagree does not mean that the legislature could not create a default rule that presumptively binds such sellers unless they opt out. Nothing in the Court’s opinion stops expressive sellers from promising nondiscrimination—because those sellers would be choosing whether or not to be so bound. But the same would hold true if the state legislature flipped the default with a statute presuming that expressive sellers implicitly warrant nondiscrimination unless they explicitly opt out by disclaiming the default.

Some may be concerned that an implied warranty of nondiscrimination is not comparable to other contractual warranties because it could create potential liability in the absence of privity. This could happen when a seller who has failed to disclaim the warranty nonetheless discriminates, refuses to deal, and then is held liable—even though (because of the refusal) there is no contractual relationship with the victim plaintiff. We have three responses to this concern, grounded in (1) statutes prohibiting unfair, deceptive, or abusive practices; (2) solicitation fraud; and (3) third-party-beneficiary theories. Our first two responses offer bases for liability in the absence of privity, resorting to statutory and tort theories outside of contract; the third response returns to contract and shows how liability for discrimination could arise absent privity even within a more standard contract regime.

Civil duties arising under “Unfair, Deceptive or Abusive Practices” (UDAP) statutes, which prohibit unfair or deceptive marketing, are not limited to settings where enforceable contracts have been formed.30 UDAP duties have already been interpreted to cover discrimination and resulting failure to contract as “unfair.”31 For example, at the federal level, the Consumer Financial Protection Bureau announced on March 16, 2022, that “it considers discrimination to be a UDAAP and will begin examining for discrimination itself and for whether companies are adequately ‘testing for’ discrimination in their advertising, pricing, and other activities.”32

A theory of solicitation fraud also supports potential liability in the absence of privity.33 The seller, in soliciting offers from the public, implicitly represents that it will not discriminate among such offers on the basis of an offering consumer’s protected status. Consumers who rely upon this implicit representation and undertake to deal with the seller are harmed when the seller will not abide by the implied representation of nondiscrimination, and that harm gives rise to a cognizable right of action. Solicitation fraud is a species of promissory fraud, and in this way an implied warranty of nondiscrimination would fall within a standard interaction of tort and contract law.

Finally, third-party-beneficiary theory can support liability in the absence of privity. Under traditional contract doctrine, the implied warranty can be viewed as an implied representation that the seller does not discriminate, with members of the protected class as the intended third-party beneficiaries. The implied representation is made to all of the consumers with whom the seller is in privity and thus becomes a part of their contracts.34 When the implied warranty is breached through a refusal to deal with a member of the protected class, third-party-beneficiary law creates a right of action for the harmed individual.35

A principal advantage of flipping the default rule from “discrimination allowed” to “discrimination prohibited” is that it likely better aligns with contracting preferences. It is probable that a nondiscrimination warranty is a majoritarian default; this reduces the transaction cost of retailers attempting contractually to opt into a commitment not to discriminate.36

To assure that a nondiscrimination warranty is a majoritarian default, it might be prudent for it first to be enacted in “bluer” states with broader public commitments to equal treatment (indeed, the sort of public-accommodations statute involved in 303 Creative would provide strong evidence of such majoritarian preferences). The statute creating an implied warranty of nondiscrimination might also include a provision empowering the state’s office of civil rights to sunset the provision if it finds that a majority of sellers are disclaiming the warranty.37

But even in the most progressive states, we should expect that some group of sellers will disclaim the warranty. Some will disclaim as a matter of conscience, but others might respond to market incentives and disclaim the warranty to serve niche demand for discrimination. To see how competition might even lead to an overrepresentation of sellers who reserve the right to discriminate, let us imagine a stylized market consisting of ten sellers of some expressive service.38 Suppose that the ten sellers’ expressive products are equally appealing to the average consumer, such that each seller’s market share is 10% of the market.39 Suppose further that 5% of customers support the kind of discrimination at issue so strongly that they will go out of their way to buy from companies that reserve the right to discriminate on that basis. And imagine that four times as many consumers—20%—actively dislike or disapprove of discriminatory sellers (enough so that they will avoid purchasing from companies that have reserved the right to discriminate). The remaining 75% of the consumers do not care one way or the other. Now consider what happens to the first company that disclaims the nondiscrimination warranty. Even if that company loses all of its business from the pro-equality consumers (2%—their share of the market’s consumers who dislike discrimination), that difference is more than made up by the pro-discrimination consumers who are induced to buy from the disclaiming producer (5% of the total market, all of whom are drawn to the disclaiming business). The first mover increases from a market share of 10% (before disclaiming the warranty) to a 12.5% market share (after disclaiming the warranty). This increase in demand results despite the fact that the company loses all of the pro-equality customers’ business. The disclaiming company is still getting its random tenth of the consumers who do not care (one-tenth of 75% = 7.5%), plus all the pro-discrimination consumers (5%).

The result of our thought experiment contradicts the intuitions of some scholars who worry that boycotts would destroy businesses with the temerity to disclose their intention to discriminate.40 How can it be that a firm has an incentive to disclaim the warranty when consumers who prefer the warranty outnumber those who do not four to one? The answer is that most of the pro-equality consumers were not going to buy from the first-mover firm anyway. Because there were ten identical firms in the market, the first mover only had a 10% chance of getting any consumer to buy. From the first mover’s perspective, the pro-equality consumers fall from a 10% chance to a 0% chance of buying. But the pro-discrimination consumers rise from a 10% chance to a 100% chance of buying. Because of this disproportionate change in shifting probabilities, the buycott effect is likely to be much stronger than the boycott effect for first adopters in markets with many firms.41

Of course, in the real world, the pro-discrimination consumers will not go all the way to a 100% probability of buying. But the underlying idea that first adopters will not be deterred, even in the face of considerable antidiscrimination consumer sentiment, still holds true. Indeed, in our stylized example, a second firm will have an incentive to disclaim the nondiscrimination warranty as well. The two disclaiming firms will now split the pro-discrimination consumers, so each can expect to control 10% of the market—7.5% (one-tenth of consumers who do not care) plus 2.5% (one-half of the pro-discrimination consumers). Had the second adopter not disclaimed, it would have controlled only 9.5% of the market—7.5% (one-tenth of consumers who do not care) plus 2% (one-tenth of the pro-equality consumers). The example shows that a pro-discrimination demand segment as small as 5% might be able to induce 20% of producers to disclaim the nondiscrimination warranty.42 One should expect a similar response to a nondiscrimination warranty, where discrimination-friendly businesses might even affirmatively produce an analogue to the Negro Motorist Green Book43—this one designed for consumers who wish to reward discrimination. Thus, there are both theoretical and empirical reasons to believe that an implied nondiscrimination warranty will produce an equilibrium that provides substantial civil-rights protection while it simultaneously allows sufficient diversity in the marketplace to assure courts that businesses have a viable means of disclaiming the warranty. Indeed, our simple model provides reasons to think that in a world with substantial numbers of consumers who are indifferent to whether businesses pledge nondiscrimination, we should expect minority preferences for discrimination to be over-represented by a competitive market.

This thought experiment might make some readers wonder whether our proposal risks increasing the presence of discriminatory businesses in some markets. If publicly disclaiming the warranty comes with a competitive advantage in some markets, even those with majoritarian preferences for nondiscrimination, then is it possible that a public-disclaimer requirement will cause more discrimination than the status quo? We think the answer to this turns on how we describe the status quo. If the status quo, after 303 Creative and prior to passage of an implied warranty of nondiscrimination, is that any expressive business can refuse service on the basis of some conscientious objection without prior warning, then discrimination may be occurring more often than we realize. It is quite likely that behavior carrying no negative legal consequences would receive little monitoring and occur without any centralized information gathering to quantify the frequency of discrimination. Our proposal, at least, provides a basis for data gathering.

Moreover, in our stylized model, the incentives for businesses to reserve the right to discriminate begin and end with their own expectations about consumer preferences. In dynamic markets, the percentage of consumers who are positive, negative, or neutral about discrimination on the basis of a particular characteristic may be in flux, and as other societal forces influence consumer hearts and minds, the business impact from declaring or disclaiming discrimination will change. In this way, the market participates in, but does not solely drive, the preferences for discrimination that attract or repel customers in our hypothetical market.

II. choosing the altering rules

Altering rules are the set of necessary and sufficient conditions for displacing a default rule.44 Conditions that artificially impede businesses from disclaiming a nondiscrimination warranty might unreasonably burden constitutional interests by making it too difficult for business owners to preserve their free-speech rights.45 Accordingly, it is prudent to craft an implied nondiscrimination warranty so that it can be easily disclaimed. We propose that any expressive business can accomplish this by sending their Secretary of State a letter reserving the right to discriminate and indicating (1) the expressive products or services that they sell, and (2) the (otherwise) protected characteristics on which they wish to reserve the right to discriminate. Thus, the plaintiff in 303 Creative might have preserved her right to refuse service to a same-sex couple simply by sending a one-time, twenty-four-word message to the state: “303 Creative reserves the right to discriminate on the basis of sex and sexual orientation with regard to its service of creating wedding websites.”

But beyond the concern that altering rules can be unconstitutionally burdensome, there are other dimensions on which policy analysts might reasonably differ. In the remainder of this Part, we consider two such dimensions—whether the law should require ex ante or ex post disclaimers, and whether the disclaimers should be more public.

One might wonder whether ex post disclaimers (i.e., those that occur only after a business discriminates on a protected basis) should be sufficient.46 After all, the Boy Scouts were allowed to wait until they were sued to announce publicly that they reserved the right to discriminate on the basis of sexual orientation.47 We think there are decisive arguments favoring ex ante disclaimers. Just as the UCC does not allow post-sale disclaimers of merchantability warranties,48 businesses should be required to disclose to the public in advance the terms on which they are doing business, including whether they reserve the right to discriminate.

A requirement that disclaimers occur ex ante helps foster the three core informational benefits set forth in this Essay’s Introduction. First, ex ante disclaimers inform the state and thereby give it the ability to assess and potentially contest whether the claimed products and services are sufficiently expressive to fall within the protection of the 303 Creative decision going forward. This is important because the Court’s opinion in 303 Creative does not provide guidance on how to evaluate the expressive nature of other products or services that sellers may withhold on a discriminatory basis (such as dressmakers, limo drivers, or florists, to name just a few businesses often associated with weddings). Filings with the state disclaiming the warranty would give each state the opportunity to establish the contours of expressive products and services, perhaps in ways that would provide advisory guidance to other businesses considering disclaimers for their activities. The state could provide—on the very website where businesses make their disclaimers—advisory guidelines and examples of business activities that either do or do not meet the standards for expressiveness protected by 303 Creative. If the state determines that a business is not expressive and publicizes both the nature of the business and the reasoning behind the rejection of the disclaimer, it could preemptively deter sellers from discriminating if those sellers perceive similarities between their business and one that was found not to be expressive.

Second, ex ante disclaimers can be standardized to allow the state to create a database that is searchable by consumers. Disclaiming in advance can allow potential victims of discrimination to reduce the dignitary harms of being refused service in person. And finally, the searchable database made possible by ex ante disclaimers facilitates more informed associational consumer choices—including both buycotts and boycotts.49

A more difficult question concerns how conspicuous disclaimers ought to be. For example, in addition to sending a disclaimer to the Secretary of State’s office, a disclaiming business might be required to post a sign or contract provision outside the store or at the point of sale, informing all consumers that the business has reserved the right to discriminate on certain bases with regard to certain products. Public disclaimers could force businesses to “put their money where their mouth is” in ways that more rigorously test the sincerity of their beliefs and then hold businesses accountable for their discrimination, including potential boycotts by customers.50 Public disclaimers might even dissuade businesses from reserving the right to discriminate, even though they would have been willing to disclaim in less public ways. Visible disclaimers might also stop some consumers from purchasing, even though they would have been willing to purchase if they were not forced to acknowledge to themselves that they were choosing to purchase from a potentially discriminatory business.51

Ronan Avraham and Daniel Statman have offered some thoughtful analysis of refusals to sell or serve, with an aim to help distinguish between refusal to serve a client that “stems from objection to the content of the service” versus “rejection of the client,” with the former not constituting discrimination.52 They have also suggested that, to minimize the dignity-deteriorating impact of an effective disclaimer, the wording should be “formulated not only negatively but also positively,” such as the following: “We happily serve all clients regardless of their religion, race, gender or sexual orientation. However, for religious reasons, we apologize for not being able to provide [for instance] photographing services to same-sex weddings.”53

They do worry that disclaimers “exactly because of their overtness, might be ‘contagious,’ increase social polarization, and make service providers less compromising and less tolerant.”54 They also worry that in “more conservative towns or neighborhoods, the cost to businesses might be much lower (such businesses might even benefit financially from such disclaimers), and same-sex couples might be worse off as a result.”55

Douglas Laycock, conversely, has worried that public disclaimers might subject business owners to harassment or even vandalism. Laycock writes:

The real reason we can’t require public notice is that no conscientiously objecting merchant could afford to give it. It was not clear in 2008, but it is clear now that any such merchant would risk boycotts, defamatory reviews, and, simultaneously, repeated confrontational demands for service from gay couples. The merchant would also risk vandalism and worse. . . . [T]he conscientious objectors’ only hope in much of the country is to lie low as much as possible and to invoke any exemption rights as quietly and diplomatically as possible.56

Laycock’s suggestion that objectors “lie low,” however, may give inadequate weight to the associational interests and sincere scruples of many consumers, who want to know whether they are supporting businesses that would refuse service or products to members of protected classes. If a product or service is expressive, giving a business owner associational and expressive interests in withholding their work from certain consumers, surely those who purchase those products or services (because their identity is not objectionable to the proprietor) also have cognizable interests in knowing what the proprietor is expressing by accepting their business (e.g., “You’re OK since you’re not gay”). They might wish not to be a part of such a message.

Andrew Koppelman acknowledges the potential danger Laycock has identified, but argues that leaving consumers in the dark is no solution: “[T]he state can create its legal preconditions: rules that if obeyed, will create safe space for everyone. . . . A disclosure regime can do that. A regime in which gay people face unforeseeable discrimination cannot.”57

We should acknowledge that a public list of expressive businesses opting out of the nondiscrimination warranty might harm not only the businesses but also the otherwise protected groups that have been specified for denial of the relevant services. This raises a difficult and important question about the psychological harm of discrimination. Which is worse: to view a public list of businesses that have declared their unwillingness to sell to or serve you, or to walk through the world and into a business knowing that any potential seller could be the one that will deny you a service or product, because the law does not require them to disclaim nondiscrimination ex ante?

The specter of discord and harassment may be sufficiently threatening that we should give careful consideration to the way disclaiming businesses make known their reservation of rights to discriminate. It is possible that if the implied warranty opt-out system requires too much explanation from businesses—even at the level supposed by Laycock and Avraham & Statman—the warranty could force citizens to speak about their personal (religious) beliefs in a way that goes beyond the normal bounds of commercial-speech regulation. This is why our proposal calls upon businesses to specify only two things: the expressive nature of the product or service they wish to sell selectively, and the characteristics of consumers to whom they reserve the right not to sell. We do not call for disclaimers to set forth the rationale for their reservation of the right to discriminate. The Court’s holding in 303 Creative is not limited to religious expression, so we see no reason to require such explanations from businesses.

The state might assist sellers who wish to opt out by providing safe-harbor language, with blanks for the seller to fill in, such as the following: “Our business involves the following product and/or service: _______ and we reserve the right to deny our services on the basis of _______.” The state could make things even easier by providing a check-box form listing (1) qualities of products or services that would make them expressive and (2) personal characteristics protected by the public-accommodations statute that would be the basis for denial of service. Sellers would remain free to go beyond this language and provide customers with additional information in more accessible and salient ways if they chose.

Altering-rule requirements can powerfully influence the mixture of buycotts and boycotts. For example, consider the “Orthodox Union” emblem (a letter “U” encased in a larger circle or letter “O”) certifying that a product is kosher.58 The symbol is so opaque to non-Jewish people that consumers who might be disinclined to buy kosher products are likely to miss its meaning.59 By creating a kind of “acoustic separation” between consumers who disprefer and consumers who prefer kosher products, the symbol is likely to promote more buycotts than boycotts.60 A similar strategy might be adopted by businesses who signal to like-minded consumers that they oppose celebrating same-sex marriage.61 These more opaque signals of disclaiming are likely to become known by consumers with more intense preferences on the issue—both those who strongly support and those who strongly oppose nondiscrimination—and less likely to be understood by the many consumers who have less passionate preferences on the issue.62 The downside of opaque signaling will be experienced primarily by consumers who care, but are unwilling or unable to research the meaning of an opaque symbol disclaiming the warranty. Because this undermines the value of transparency and raises consumer information costs relative to more overt signage or online disclosures, it is not the optimal approach in our view.

In the end, we would not require point-of-sale signage (whether overt or opaque) because, to our minds, the costs to protected group members of having to see salient messages of discrimination—somewhat akin to “Irish need not apply”63—outweigh the potential benefits of transparency when businesses publicly disclaim the implied warranty at the point of sale. But we acknowledge that reasonable people could differ on this question.

III. is an implied warranty of nondiscrimination constitutional?

The preceding two Parts have attempted to defend, as a welfare-enhancing policy, our proposed default of nondiscrimination as well as the nonburdensome method by which the warranty may be disclaimed. In this Part, we respond to the concern that the altering rule constitutes unconstitutional compelled speech.

Some readers of this Essay could be concerned that our proposal runs afoul of the First Amendment’s compelled-speech doctrine.64 After all, our proposal compels businesses of expressive products and services to speak if they wish to reserve the right to discriminate on the basis of a characteristic that is otherwise protected by a public-accommodations statute. This Part examines whether forcing sellers to send a few words to the state unconstitutionally compels speech.

The speech compelled by an implied warranty of nondiscrimination does not offend the Constitution for two reasons, one structural and one doctrinal. First, structurally, the government cannot avoid setting a default respecting whether or not a business reserves the right to discriminate. And any default rule compels speech of contractors who wish to opt out of the default and recreate the kind of potential liability for discrimination that existed before the 303 Creative decision. Our proposal compels speech from businesses that want to disclaim a warranty of nondiscrimination. But the 303 Creative decision also created a default which analogously compels speech because, unless sellers of expressive services contractually commit not to discriminate, consumers can reasonably assume that businesses are free to discriminate without legal consequence.65 That is, the 303 Creative default rule compels speech by forcing pro-equality sellers of expressive products and services who want to credibly communicate their views to the public to come forward and expressly warrant that they do not discriminate.

And while any default compels speech by imposing altering rules for those who wish to opt out of the default, the nondiscrimination default compels less speech, both because its altering rules are not cumbersome and because it is likely to be a majoritarian default. The implied warranty of nondiscrimination default has a simple altering rule with which a seller can comply, as described above, in a single twenty-four-word statement to the enacting state’s Secretary of State. In contrast, the 303 Creative default requires a more cumbersome and repetitive contracting process establishing potential victims as intended third-party beneficiaries.66 Moreover, a substantial majority of the public disapproves of businesses discriminating against employees or customers, so there is every reason to believe that an implied warranty of nondiscrimination is a majoritarian default.67 When compelled speech cannot be avoided, majoritarian defaults structurally reduce the number of contractors who would want to opt out.68

The second justification for the speech required by our proposal grows out of the Supreme Court’s own jurisprudence. Commercial speech only began to receive First Amendment protection in 1976.69 And the Court has repeatedly made clear that commercial speech is subject to “modes of regulation that might be impermissible in the realm of noncommercial expression.”70 Quite simply, “there can be no constitutional objection to the suppression of commercial messages that do not accurately inform the public about lawful activity.”71 Because the state may lawfully suppress inaccurate commercial speech, even “[c]ontent discrimination is . . . routinely practiced within commercial speech” despite its “impermissib[ility] within public discourse.”72

In addition to the prohibition on false and misleading commercial speech, the Court has emphasized the potential reasonableness of business-disclosure mandates:

[A business’s] constitutionally protected interest in not providing any particular factual information in [its] advertising is minimal. Thus, in virtually all our commercial speech decisions to date, we have emphasized that because disclosure requirements trench much more narrowly on an advertiser’s interests than do flat prohibitions on speech, “warning[s] or disclaimer[s] might be appropriately required . . . to dissipate the possibility of consumer confusion or deception.”73

As Robert Post has recently noted, lawmakers have not been reticent in mandating business disclosures:

If we just open our eyes, we can see that American society is full of examples of compelled pure speech, ranging from required product disclosures, to disclosures in real estate transactions, to the required testimony of witnesses in a trial, to a raft of statutory obligations to report various events and circumstances, to the myriad of miscellaneous disclosure requirements imposed on commercial transactions.74

The relatively relaxed scrutiny of commercial-speech regulation makes an implied warranty of nondiscrimination even more likely to pass constitutional muster than, for instance, a default rule requiring that noncommercial actors disclose if they reserve the right to discriminate.75

The clearest articulation of the Supreme Court’s approach to scrutinizing compelled commercial speech can be seen in its landmark Zauderer v. Office of Disciplinary Counsel decision.76 In Zauderer, the Court upheld the constitutionality of compelled speech in the context of a state requirement that attorney advertisements contain particular factual disclosures:

In requiring attorneys who advertise their willingness to represent clients on a contingent-fee basis to state that the client may have to bear certain expenses even if he loses, Ohio has not attempted to prevent attorneys from conveying information to the public; it has only required them to provide somewhat more information than they might otherwise be inclined to present.

. . . .

. . . Ohio has not attempted to “prescribe what shall be orthodox in politics, nationalism, religion, or other matters of opinion or force citizens to confess by word or act their faith therein.” The State has attempted only to prescribe what shall be orthodox in commercial advertising, and its prescription has taken the form of a requirement that appellant include in his advertising purely factual and uncontroversial information about the terms under which his services will be available.77

The Court concluded that “an advertiser’s rights are adequately protected as long as disclosure requirements are reasonably related to the State’s interest in preventing deception of consumers.”78 To be reasonably related to this interest, the mandated disclosure must also “relate to the good or service offered by the regulated party.”79 If a mandated commercial disclosure fulfills these prerequisites, then it is subject only to rational-basis review.80

An implied warranty of nondiscrimination falls squarely within the Zauderer safe harbor. Businesses are only compelled to disclose “purely factual and uncontroversial information about the terms under which [their] services will be available.”81 As Robert Post has explained, the requirement that a disclosure be “uncontroversial” is best understood as “a description of the epistemological status of the information that a speaker may be required to communicate . . . . [I]f the truth of information is seriously controverted, the state cannot appeal to the relaxed Zauderer test to sanction its compelled disclosure.”82 When an expressive business opts to disclaim the implied warranty of nondiscrimination, the compelled disclosure is “factual” information regarding the terms of service—it clarifies whether or not those terms include a nondiscrimination warranty.83 In Zauderer’s terms, the disclosure is “uncontroversial” in that it is not subject to epistemological dispute—it truthfully presents whether or not the terms include a nondiscrimination warranty.

In addition, the disclosure is “reasonably related to the State’s interest in preventing deception of consumers.”84 The deception would occur if, while the commercial entity seemed to be holding itself out as doing business with the public generally, it was only willing to transact with a selective subset of the public.85 The government has a substantial interest in ensuring that consumers are adequately informed about whether the seller is offering its goods or services to (or soliciting offers from) them or not.86 Accordingly, an implied nondiscrimination warranty does not unconstitutionally compel speech.

Conclusion

Constitutional thinking about defaults is underdeveloped. We are accustomed to treating constitutional rules as mandates that either prohibit or require certain results. This is all the more true with regard to civil rights, where mandatory rules prohibiting discrimination take center stage. The underappreciation of defaults yields non sequiturs like Richard Epstein’s, in which criticism of the mandatory nature of civil-rights legislation led mistakenly to his proposal for a full undoing of such legislation without considering the more responsive middle ground of a disclaimable prohibition.87

More work needs to be done by courts and commentators to flesh out the circumstances under which the government can constitutionally flip defaults.88 At a minimum, when private actors have ultimate control over an issue and when there are constitutional interests on both sides, states should be empowered to establish majoritarian defaults with nonburdensome altering rules. An implied warranty of nondiscrimination meets this standard. Businesses that sell expressive goods or services have a right to make binding promises not to discriminate as well as a right—per 303 Creative—not to make such a commitment. There are constitutional interests on both sides of the issue: the Court has made clear that expressive sellers have a First Amendment interest in not being forced to espouse views with which they disagree, but it is equally clear that customers have a First Amendment interest in informed association.89 Indeed, patrons may feel defrauded if they learn they were transacting with a business that discriminates on the basis of a protected characteristic.

Legislatures might also do well to pay more attention to default choice. Given the mandatory nature of the Colorado Anti-Discrimination Act and its collision with business owners’ freedom of expression as understood by the 303 Creative majority, invalidation of the Colorado law seemed inevitable as applied to such expressive businesses. But this need not be the final chapter for nondiscrimination norms in Colorado and other states. If legislatures enact an implied warranty of nondiscrimination—and provide a reasonable process for expressive businesses to disclaim the warranty—states can address the Court’s central constitutional misgiving regarding the mandatory nature of the Colorado Anti-Discrimination Act90 and still preserve consumers’ interest in access to a market free of discrimination where sellers fail to opt out.

Some might argue that the 303 Creative decision was so unexpected that legislatures could not have reasonably foreseen the possibility that nondiscrimination mandates would be struck down with regard to expressive sellers. And the decision might be considered a unique one, cabined to the rare circumstances of what Justice Gorsuch characterized as regulating “pure speech.”91 But the Supreme Court may not be finished paring back the reach of our civil-rights laws. In his concurrence in Ricci v. DeStefano, Justice Scalia warned that “the war between disparate impact and equal protection will be waged sooner or later, and it behooves us to begin thinking about how—and on what terms—to make peace between them.”92

There is a nontrivial chance that the Court’s recently empowered conservative majority will strike down disparate-impact liability as invidious government discrimination.93 Foreseeing this possibility, one could imagine state or federal legislation with a savings clause creating a disclaimable warranty not to discriminate in ways that create unjustified disparate impacts. Reasonable legislatures might eschew such a proactive savings-clause approach because it could be seen as inviting judicial invalidation of the original nondiscrimination mandates. The more prudent legislative approach would be to respond following invalidation.

In any case, now, in the wake of 303 Creative, Colorado and other states need not remain silent. By enacting an implied warranty of nondiscrimination, states can adroitly furnish themselves and their populaces with valuable information about what is and is not being promised. In such a regime, most sellers would stick with the nondiscrimination default; others would take advantage of the state’s opt-out provision and declare their intention potentially to discriminate in ways otherwise prohibited by a nondiscrimination statute. In this regime, a state would be able to cabin appropriately the discriminatory impact of the 303 Creative decision while simultaneously maintaining a vibrant space for free expression.

Ian Ayres is Oscar M. Ruebhausen Professor of Law, Yale Law School. Jennifer Gerarda Brown is Professor of Law, Quinnipiac University School of Law. The authors thank for helpful comments Jed Rubenfeld and participants at the University of California College of the Law, San Francisco, Center for Negotiation and Dispute Resolution’s 20th Anniversary “DEI and Dispute Resolution: Reimagining the Field” Conference. Ji Young Kim and Richard Peay provided helpful research assistance.