ICLE Amicus Letter Supporting Review in Liapes v Facebook
RE: Amicus Letter Supporting Review in Liapes v. Facebook, Inc. (No. S282529), From a Decision by the Court of Appeal, First Appellate District, Division 3 (No. A164880)
The International Center for Law & Economics (“ICLE”) is a nonprofit, non-partisan global research and policy center aimed at building the intellectual foundations for sensible, economically grounded policy. ICLE promotes the use of law and economics methodologies and economic learning to inform policy debates and has longstanding expertise evaluating antitrust law and policy. We thank the Court for considering this amicus letter supporting Petitioner Facebook’s petition for review in which we wish to briefly highlight some of the crucial considerations that we believe should be taken into account when looking at the intermediary liability principles that underlie the interpretation of the Unruh Act.
The Court of Appeal’s decision in Liapes v. Facebook has profound implications for online advertising and raises significant legal and practical concerns that could echo beyond the advertising industry. Targeted advertising is a crucial aspect of marketing, enabling advertisers to direct benign, pro-consumer messages to potential customers based on various considerations, including age and gender. The plaintiff’s argument, and the Court of Appeal’s acceptance of it, present a boundless theory of liability, suggesting that any targeted advertising based on protected characteristics is unlawful. This theory of liability, unfortunately, fails to take account of the nature of Facebook as an online intermediary, and the optimal limitations on liability that this requires when weighing the bad acts of third parties against Facebook’s attempt to provide neutral advertising tools to the benefit of millions of users.
The Unruh Act Is Not a Strict Liability Statute
While the Unruh Act prohibits intentional discrimination, California Civil Code §§ 51 and 51.5, California courts have consistently emphasized that the statute does not impose strict liability for all differential treatment. Rather, the Unruh Act allows for distinctions that serve legitimate nondiscriminatory purposes.
Courts have held that the Unruh Act does not bar practices “justified by ‘legitimate business interests.'” Koebke v. Bernardo Heights Country Club, 36 Cal. 4th 824, 851 (2005). The statute prohibits only discrimination that is “arbitrary, invidious or unreasonable.” Javorsky v. Western Athletic Clubs, Inc., 242 Cal. App. 4th 1386, 1395 (2015). Reasonable, nonarbitrary distinctions are therefore permissible. Differential treatment may qualify as reasonable and nonarbitrary if there is a public policy justification for the distinction. For example, discounts for senior citizens have been deemed nonarbitrary because they advance policies like assisting those with limited incomes. Sargoy v. Resolution Trust Corp., 8 Cal. App. 4th 1039, 1044 (1992). And it is “reasonable” discrimination on the basis of age to prevent minors from entering bars and adult bookstores. Koire v. Metro Car Wash, 40 Cal. 3d 24, 31 (1985).
Thus, the Unruh Act does not impose strict liability merely for practices that have a disparate impact. Harris v. Capital Growth Investors XIV, 52 Cal. 3d 1142, 1149 (1991). While the Unruh Act provides robust protections, it was not intended to forbid all differential treatment. Distinctions based on legitimate justifications remain permissible under the statute’s exceptions.
Firms like Meta operate services facilitating billions of interactions between users and advertisers. In this vast, complex environment, interpreting any ad targeting based on protected class membership as a per se Unruh Act violation would amount to imposing de facto strict liability on the online advertising industry. Setting aside the fact that the Unruh Act is not a strict liability statute, drawing the liability line at this point would have drastic practical consequences.
First, a de facto strict liability standard fails to account for the immense scale and complexity of services like Facebook. Given the number of third-party advertisers and users, as well as the speed and quantity of ad auctions, some incidental correlations between ad delivery and protected characteristics are likely inevitable even absent purposeful exclusions. The Court of Appeal’s opinion exposes both advertisers and platforms like Facebook to litigation based on such correlations, on the theory that the correlations may be “probative” of the intentional discrimination the Unruh Act forbids.
Second, advertisers may have many reasonable, nonarbitrary motivations for targeting their ads to certain demographic groups. For example, targeting older people for certain kinds of medicines, or members of religious groups with information about services in their religion. The Court of Appeal’s opinion will lead to extensive, costly litigation about potential justification for such ad targeting, and in the meantime consumers will be deprived of useful ads.
Finally, if any segmentation of ad targets based on protected characteristics triggers Unruh Act violations, online advertising loses an essential tool for connecting people with relevant messages. This impedes commerce without any showing of invidious discrimination.
Although the Unruh Act provides important protections, overbroad interpretations amount to strict liability incompatible with the realities of a massive, complex ad system. Nuance is required to balance anti-discrimination aims with the actual welfare of users of services. In order to properly parse the line between reasonable and unreasonable discrimination when dealing with a neutral advertising service like Facebook and the alleged bad acts of third parties, it is necessary to incorporate the legal principles of intermediary liability into an analysis under the Unruh Act.
Principles of Intermediary Liability
In public policy and legal analysis, a central objective is to align individual incentives with social welfare, thereby deterring harmful behavior and encouraging optimal levels of precaution. See Guido Calabresi, The Cost of Accidents: A Legal and Economic Analysis 26 (1970). In the online context, this principle necessitates a careful examination of intermediary liability, especially for actors indirectly involved in online interactions.
Intermediary liability applies to third parties not directly causing harm but who can influence primary actors’ behavior to reduce harm cost-effectively. This is particularly relevant when direct deterrence is insufficient, and the intermediary can prevent harm more effectively or at a lower cost than direct enforcement. See Reiner Kraakman, Gatekeepers: The Anatomy of a Third-Party Enforcement Strategy, 2 J.L. Econ. & Org. 53, 56-57 (1986). However, not every intermediary in a potentially harmful transaction should be a target for such liability.
The focus is on locating the “least-cost avoider” – the party that can reduce the likelihood of harm at the lowest overall cost. See Harold Demsetz, When Does the Rule of Liability Matter?, 1 J. of Leg. Stud. 13, 28 (1972); see also Kraakman, supra, at 61 (“[t]he general problem remains one of selecting the mix of direct and collateral enforcement measures that minimizes the total costs of misconduct and enforcement”). This approach aims to balance the costs of enforcement against the social gains achieved as well as the losses that flow from the chilling effects of liability.
Imposing liability involves weighing the administrative costs and the potential lost benefits society might enjoy in the absence of liability. See Ronald Coase, The Problem of Social Cost, 3 J.L. & Econ. 1, 27 (1960) (“[W]hat has to be decided is whether the gain from preventing the harm is greater than the loss which would be suffered elsewhere as a result of stopping the action which produces the harm.”). The least-cost avoider is determined by considering whether the reduction in costs from locating liability on that party is outweighed by the losses caused by restricting other activities that flow from that liability. Calabresi, supra at 141.
The internet comprises various intermediaries like interactive computer services, internet service providers, content delivery networks, and advertising networks, which facilitate interactions between users, content platforms, and various service providers. See generally David S. Evans, Platform Economics: Essays on Multi-Sided Businesses (2011). Sometimes, intermediaries are the least-cost avoiders, especially when information costs are low enough for them to monitor and control end users effectively, or when it is difficult or impossible to identify bad actors using those platforms. But this is not always the case.
While liability can induce actors to take efficient precautions, intermediaries often cannot implement narrow precautions due to limited information or control. Facebook’s platform illustrates this challenge: Facebook has limited to no access to information about the motivations or design of every one of the millions of ad campaigns from millions of individual advertisers on its platform at any given time. Thus, avoiding liability risk might entail broad actions like reducing all services, including those supporting beneficial activities. If the collateral costs in lost activity are significant, the benefits of imposing intermediary liability may not justify its implementation.
Here, overbroad liability could end up severely reducing the effectiveness of advertising in general. This could result in 1) less relevant advertisements for users of online services; 2) reduced value to advertising for businesses, harming in particular small businesses which have limited advertising budgets, and 3) less revenue for online services which rely on advertising revenue, pressuring them to increase revenue through other means like higher ad prices and subscriptions.
The individuals and businesses placing advertisements, not the intermediary ad platform, are the primary actors choosing whether and how to use tools for targeting. As we noted above, under the Unruh Act there are permissible uses of targeted advertising, even when focusing on protected classes. The focus in discouraging discrimination should be on primary actors.
It is not hard to locate parties misusing Facebook’s advertising tools in a way that potentially violates the Unruh Act when evidence of discrimination is presented. On the other hand, intermediaries like Facebook will often lack particularized ex ante knowledge of specific discriminatory transactions or direct control over advertisers’ targeting choices. The only avenue for Facebook to comply with broad liability under the Unruh Act is to altogether remove the ability of businesses to use any characteristic that might theoretically trigger Unruh Act liability, which would result in the harms described above. In situations like this, where the intermediary has little ability to effectively police certain misuses of otherwise benign, neutral tools that enhance social welfare, the case for imposing collateral liability is weakened.
Moreover, some statistically disproportionate ad delivery outcomes may be inevitable given the vast scale of platforms like Facebook. Disparate effects should not automatically equate to impermissible discrimination absent purposeful exclusion. The creation of neutral tools for use by advertisers who then use them to break the law does not imply intentionality by Facebook (or any other advertising platform) to break the law. No one would suggest that a hammer company intends for its product to be misused by customers who use it to bludgeon another human being. Nuance is required.
Broad Unruh Act liability risks unintended harms. Imposing a de facto strict liability regime that treats all ad targeting of protected classes as impermissible under the Unruh Act would drive services like Facebook to restrict lawful advertising tools for all users in order to mitigate liability risks. This impairs a large amount of indisputably legal commerce to deter allegedly illegal advertising by a subset of third parties. Moreover, the effects of such a decision would echo not only throughout the advertising ecosystem, but throughout the internet ecosystem in general where intermediaries might provide similar neutral tools that could run afoul of such a broad theory of liability.
Conclusion
The intermediary liability principles outlined above strongly counsel against the overbroad Unruh Act interpretation embraced by the Court of Appeal in the present matter.
The primary actors are the advertisers choosing whether and how to target ads, not Facebook. The Court of Appeal’s broad view wrongly focused on Facebook’s provision of neutral tools rather than advertisers’ specific uses of those tools.
Given the context-dependent nature of an Unruh Act analysis, the Court of Appeal failed to appropriately balance between deterring allegedly illegal acts by advertisers with the potential loss of value from targeted advertising altogether. The proper duties of intermediaries like Facebook should be limited to feasible actions like removing impermissibly exclusionary ads when notified. They should not include disabling essential advertising tools for all users. The Court of Appeal’s overbroad approach would ultimately harm consumer access to targeted advertising.
With the foregoing in mind, we respectfully urge this Court to grant the pending petition for review. Careful examination of the Court of Appeal’s ruling will reveal it strays beyond the Act’s purpose and ignores collateral harms from overdeterrence. Guidance is needed on balancing antidiscrimination aims with the liberty interests of platforms and their users. This case presents an ideal vehicle for this Court to provide that guidance.