Some Thoughts on the Relevance of Customer Behavior to Discrimination Law: Who Counts as a “Customer”?

In Discrimination by Customers,1 Katharine Bartlett and Mitu Gulati draw attention to the apparent asymmetry in the law’s treatment of “firms” and “customers.” Firms are prohibited from engaging in discrimination against customers. Customers, however, are generally free to act on discriminatory preferences in their interactions with firms even though, as Bartlett and Gulati show, discrimination by customers can hardly be regarded as benign. They claim that “[i]t is a puzzle why we generally take for granted the right of customers to discriminate when they exercise their buying power.”2

In arguing for greater legal attention to the problem of discrimination by customers, Bartlett and Gulati consider and persuasively reject two possible arguments for limiting discrimination liability to firms (the “efficacy” and “privacy and individual autonomy” rationales) and then offer a proposed legal intervention to address discriminatory customer conduct.3 Their discussion of these issues is rich, nuanced, and warrants further inquiry. But my focus in this Response will not be on the efficacy or autonomy arguments, nor on the specifics of Bartlett and Gulati’s proposal for reform. Rather, this Response offers a perspective that the authors do not seem to consider. Although they discuss at length the arguments against liability for customer discrimination, they never explicitly set forth their understanding of the positive justifications for antidiscrimination law, other than their implicit assumption that the law is concerned with the harms of discrimination. They express puzzlement that the law prohibits discrimination by firms but not by customers, but they spend little time discussing why the law prohibits discrimination by firms in the first place. Is the law’s apparent apathy toward customer discrimination as puzzling as Bartlett and Gulati claim from the perspective of the basic principles that underlie antidiscrimination law?

A preliminary difficulty in trying to understand the theoretical significance of “firm” versus “customer” discrimination lies in determining who constitutes a customer according to Bartlett and Gulati. They do not explicitly define the term, although the examples they discuss are strongly suggestive. They deftly describe numerous everyday situations where the law seems to leave people free to make transactional choices based on discriminatory preferences relating to race, sex, religion, and so on.4 But the authors leave it largely to the reader to infer how these examples conceptually hang together. They presuppose a broad concept of customer discrimination that includes tipping a waiter in a restaurant, choosing a store clerk to interact with, watching a sporting event, hailing an Uber, and filling out a student course evaluation.5 It may indeed be puzzling, as Bartlett and Gulati suppose, that all of these actions are outside the scope of discrimination law, but one might first ask a more basic question: what exactly warrants lumping them together into the category of customer discrimination? Perhaps if we can understand just what it is that makes a customer a customer, we can gain insight into the basis for antidiscrimination law’s apparent lack of concern about customer discrimination.

To begin, I am not so sure that the law’s reticence with regard to some of the forms of discriminatory conduct discussed by Bartlett and Gulati is truly a function of the actor’s status as a customer, as opposed to some other latent feature of the relevant context. The authors point out, for example, that “[c]ustomers may avoid doing business with a gay photographer because they are disgusted by homosexuality” without fear of legal liability.6 This seems quite true; but on the other hand, it is also probably true that the law would have little to say about a self-employed photographer who avoids dealing with same-sex customers for homophobic reasons. And the same could be said of a housekeeper who takes on only white clients; or a door-to-door salesman who chooses to peddle his wares only in white neighborhoods. Legal liability seems no more likely for these actors,7 who are not “customers” in any ordinary sense, than for Bartlett and Gulati’s white diners who prefer cafés with white clienteles or xenophobic shoppers who steer clear of turban‑wearing clerks.8 Bartlett and Gulati are surely right that in many contexts, “the law does not prohibit discrimination by customers.”9 But in some of these contexts, the law also does not seem to prohibit discrimination by the party on the other side of the transaction either.

This suggests the possibility that the law’s approach to customer discrimination might be better understood as a piece of a larger puzzle, rather than something that could purport to make sense on its own. In other words, the most viable explanation of the legal permissibility of customer discrimination may require enlarging our inquiry to consider the broader principles that justify the interventions of antidiscrimination law in the first place. Given that the law does not even pretend to be anything close to comprehensive and universal in its coverage, what exactly is it that the law is trying to accomplish?

One place to begin our broader inquiry is in the employment context, where no one questions the basic legitimacy of antidiscrimination prohibitions. In this context, Bartlett and Gulati do not mention a rather glaring asymmetry in the application of the law: employers are expressly prohibited from discriminating against employees, but employees are not prohibited from discriminating against employers.10 Employees are free to make their own employment decisions based on discriminatory preferences relating to the race, color, religion, sex, or national origin of their potential employers, co-workers, and clientele. It is understandable that Bartlett and Gulati do not discuss discrimination by employees, insofar as employees presumably do not count as customers in their sense,11 but the apparent permissibility of employee discrimination arguably raises the same problem as the puzzle of customer discrimination. Why is the party on one side of the employment relationship subject to antidiscrimination liability while the other side seems free to discriminate at will?

The question may seem contrived, but I think it is worth asking because there are some answers that quickly suggest themselves; and I believe these answers will help us understand the law’s approach to customer discrimination. We are not accustomed to thinking of employers as potential victims of discrimination by employees, but many of Bartlett and Gulati’s insights about the possibility of customer discrimination could be thought to suggest similar concerns about discrimination by employees. It is not difficult to imagine scenarios in which workers engage in discriminatory conduct against minority supervisors and managers because of bias, resentment, or animus. Yet, if an employee were to engage in such conduct, they would not be held liable for any resulting harm under current federal employment discrimination law.12 Following Bartlett and Gulati, we might then ask: if the purpose of employment discrimination law is to reduce the prevalence of discrimination and minimize the social harms that flow from discriminatory conduct, then why should employee discrimination be outside the law’s reach?

Two responses come to mind, both of which may help us see the puzzle of customer discrimination in a broader theoretical context. The first and perhaps more obvious answer is that an employer who is “victimized” by an employee’s discriminatory conduct can avail itself of self-help. If the discriminatory actor is an at-will employee, the employer is free to discharge him; and even in a situation where the employer is contractually or legally more constrained, the employee’s discriminatory conduct will likely provide a basis to terminate for just cause. Thus, as a general matter, there is no need to provide employers with antidiscrimination remedies against employees because it is the employer who holds all of the relevant economic power.

The second response is that the question may rely on a faulty premise. Our legal frameworks aim generally to reduce the occurrence of discrimination throughout society and to prevent or remedy the harms associated with discrimination. But this characterization of the purposes of discrimination law is arguably overbroad if we are trying to identify plausible rationales for our existing laws, which limit themselves to reach some but not all modes of discriminatory conduct. If we are talking about Title VII in particular, it would be more accurate to say that the original purpose of the law was to remove status-based barriers that historically had prevented members of socially disfavored and vulnerable groups from entering or occupying the most desirable positions in the workplace.13 To be sure, Title VII expresses a general commitment to reduce societal discrimination, but the statute’s specific goal is to foster equality of opportunity in employment.

From the perspective of that normative priority, there is a ready answer to the question why Title VII prohibits employer discrimination against employees but not vice versa. The law focuses on employers because it is they who control the distribution of workplace opportunities and the terms and conditions of workers’ employment. As long as jobs are relatively scarce, the primary barrier to equality of opportunity and workplace conditions is employer discrimination, not the conduct of employees.

One might argue that discrimination by employees at the expense of employers could, in certain circumstances, present a hindrance to full equality of opportunity and of workplace conditions. That may well be true, but even so, it is not particularly puzzling why employee discrimination is not a featured concern of the law. Mere employees do not have the power to hire, fire, promote, or provide pay and benefits. Granted, they might, through pervasive discriminatory behavior, create a hostile environment that undermines genuine workplace equality. But in such a case, the employer ultimately has the power, and is indeed charged with vicarious legal responsibility, to eliminate the offending hostile conditions.14 In short, as real as the possibility of employee discrimination may be, Title VII holds only employers liable for discrimination because employers, not employees, have power over opportunities for, and the terms and conditions of, employment in their workplaces.

This explanation of Title VII’s approach is broadly consistent with liberal conceptions of justice in the context of a market-based economy. In a society that allows some degree of economic inequality to exist, certain basic requirements must be satisfied in order for that inequality to be regarded as fair, or consistent with justice. One of those requirements is that the society’s institutional arrangements (including its laws) must ensure fair equality of opportunity for all citizens. In order for that requirement to be satisfied, “conditions must be such that those with equal talent who make equal efforts have the same chance of occupying positions of status and power.”15 The practice of employment discrimination is plainly inconsistent with the requirement of fair equality of opportunity.16 Discrimination by employees, on the other hand, while perhaps no less morally offensive, is not so clearly a threat to the satisfaction of that requirement for the reasons just discussed.17 Thus, a society plagued by the practice of employment discrimination, if committed to the attainment of the necessary conditions of justice, will have compelling reason to place high priority on legal interventions that prohibit such discrimination. But it might not have equally urgent reasons to enact coercive legal measures to address discrimination by employees against employers.

There might of course be other legitimate justifications for legal interventions that specifically address discrimination by employees. For example, according to liberal political philosophy, another basic condition of justice is that no one can be deprived of a social basis for self-respect—i.e., social recognition consistent with regarding oneself as equally worthy of pursuing and fulfilling one’s chosen ends.18 Widespread discrimination within a society could cause members of disfavored groups to feel socially marginalized and to perceive their interests as being subordinate to the good of more dominant groups. Such a society might have strong reasons to enact thoroughgoing measures, beyond discrimination laws focusing on employers, to bolster the universal entitlement to the social bases of self-respect.19 I do not deny that such a society might exist, or that ours might be an example of one. But the relevant point is that this sort of justification for antidiscrimination interventions is distinguishable from the demand of fair equality of opportunity that drives the prohibition of employment discrimination.

Let us return, finally, to the problem of customer discrimination. Why does current law allow customers to discriminate, even when such discrimination is harmful? I believe the most promising line of response is analogous to the answer developed above to the question of employee discrimination. The primary concern of federal antidiscrimination law is not necessarily to eradicate all conduct motivated by discriminatory preference or bias, but to eliminate basic forms of injustice. In the workplace context, employment discrimination law seeks to realize equality of opportunity and equality in the terms and conditions of workers’ employment. In other contexts, antidiscrimination law seeks to ensure equal access to essential goods, services and benefits such as housing and education. The law intervenes as necessary to prevent unjust inequalities in access to these basic opportunities and resources.

In short, the central concern of current antidiscrimination law is injustice in the distribution of essential goods such as jobs, income, education, and housing. The simplest explanation for the law’s apparent indifference to customer discrimination is that customers as such are not agents of distribution. That is what makes them customers.20 And since customers as such do not distribute anything (except their own money), it follows that customer discrimination need not be considered a significant cause of injustice in the distribution of essential social goods. The conclusion is that customer discrimination simply falls outside the central concern of antidiscrimination law. Or so the argument can be understood.

I want to be clear that I am not putting this distributive justice argument forward in opposition to Bartlett and Gulati’s core claim that the law should do more to address the harms of customer discrimination. Rather, I present it as an alternative framework for their views. Bartlett and Gulati characterize the problem of customer discrimination as a puzzle about the lack of parity in the treatment of customers and firms. The weakness of this framing is that from the standpoint of concern for distributive justice, this asymmetry is not really so puzzling after all: the law focuses on firms and employers because they have distributive power, and it ignores “customers” because they do not. Of course, Bartlett and Gulati can be understood as having much to say on that score. Indeed, their assertion that “it is wrong to assume . . . that discrimination by customers and firm practices are distinct from one another”21 comes close to a claim that customers do in fact have significant power to affect distributive outcomes. But this is precisely my point. What I am suggesting is that instead of setting up their analysis in the form of a “what’s-good-for-the-goose-is-good-for-the-gander” argument, it might be more strongly framed as a direct response to the naïve view that customer discrimination is irrelevant to distributive justice.

When spun in this way, I think that it becomes clearer why many of Bartlett and Gulati’s insights have the intuitive power that they do. The distributive justice perspective makes it obvious why discriminatory tipping should be of legal concern: that is a context in which the restaurant patron (customer) is empowered with distributive responsibility, and that power is significantly magnified if tips constitute a significant portion of wages in the restaurant industry. Similarly, in the context of the sharing economy, customers have a great deal of power to affect the success level and incomes of workers and owners, and patterns of customer discrimination in these contexts could cause significant unjust economic distortions. Thus, concerns of distributive justice could justify legal interventions, like those suggested by Bartlett and Gulati, to curb that power.22 To generalize, the reason that antidiscrimination law might want to take notice of customer behavior in these arenas is not just that customer discrimination is harmful, but that customers in these contexts have some of the functional characteristics of actors who have allocative power over goods and resources in traditional economic contexts—actors who reside in the heartland of antidiscrimination law.23

In conclusion, from a certain perspective, it is natural to think of customer discrimination as lying outside the scope of antidiscrimination law. Antidiscrimination law has traditionally been concerned with building fair equality of opportunity and preventing and rectifying injustices in the distribution of essential goods and resources; and customers as such neither provide employment opportunities nor have the power to control distributive outcomes. Indeed, it is arguably the lack of primary allocative authority in the economic market that makes a customer a customer in the first place. Although Bartlett and Gulati do not explicitly frame their argument in these terms, it could easily be recast as a claim that the law is wrong to assume that employers and firms are the only actors relevant to the perspective of distributive justice. Bartlett and Gulati show that behavior by actors who might be thought of as customers can have significant impacts on the value and availability of employment opportunities and on distributive outcomes. This is especially true in the context of the sharing economy, where the traditional lines between customers, employers, and workers are blurred at best. Bartlett and Gulati’s insights may suggest that expanding antidiscrimination law to take account of customer discrimination might not require a radical departure from its existing justifications and, indeed, might be quite consistent with the law’s existing commitment to achieving the conditions necessary for distributive justice.


  1. [1]. See generally Katharine T. Bartlett & Mitu Gulati, Discrimination by Customers, 102 Iowa L. Rev. 223 (2016).

  2. [2]. Id. at 226.

  3. [3]. Id. at 228–41, 249–50.

  4. [4]. Id. at 224–26, 231–34.

  5. [5]. Id. at 224–25, 228, 232, 253.

  6. [6]. Id. at 226.

  7. [7]. As Bartlett and Gulati explain, the language of section 1981 of the Civil Rights Act of 1866 appears to create a civil right of action for racial discrimination in private contracting activity, but there are few, if any, cases reporting the imposition of liability in connection with such an action. Id. at 225 n.12, 225–26.

  8. [8]. Id. at 225–26.

  9. [9]. Id. at 225.

  10. [10]. See 42 U.S.C. §2000e-2(a) (2012).

  11. [11]. Indeed, if the employer–employee relationship were analogized to a customer transaction, the employer might seem to correspond to the role of the customer, insofar as the employer is paying for services to be rendered by the employee. The reason why employers are not customers in the relevant sense will become clearer in the discussion that follows.

  12. [12]. Employers can be held vicariously liable for the discriminatory actions of “supervisors” and might be liable for harassment carried out by non-supervisors if such harassment was facilitated by the employer’s negligence. See Vance v. Ball State Univ., 133 S. Ct. 2434, 2443 (2013). However, employees per se cannot be held liable under Title VII, which prohibits discrimination only by an “employer.” See 42 U.S.C. §§2000e(b), 2000e-2(a).

  13. [13]. See United Steelworkers of Am. v. Weber, 443 U.S. 193, 202–07 (1979); Griggs v. Duke Power Co., 401 U.S. 424, 431 (1971).

  14. [14]. See Burlington Indus., Inc. v. Ellerth, 524 U.S. 742, 765 (1998).

  15. [15]. Seana Valentine Shiffrin, Race, Labor, and the Fair Equality of Opportunity Principle, 72 Fordham L. Rev. 1643, 1649 (2004) (discussing John Rawls’ book, A Theory of Justice).

  16. [16]. See id. at 1650.

  17. [17]. This is not to deny that discrimination by employees is morally objectionable. The claim here is that discrimination by employees against employers is not clearly a threat to fair equality of opportunity.

  18. [18]. See John Rawls, Justice as Fairness: A Restatement 58–60 (Erin Kelly ed., 2001); John Rawls, A Theory of Justice 386 (Harvard Univ. Press rev. ed. 1999).

  19. [19]. Cf. Michael Blake, The Discriminating Shopper, 43 San Diego L. Rev. 1017, 1031 (2006) (offering a Rawlsian analysis of the permissibility of catering to discriminatory customer preferences).

  20. [20]. And this is also why employers are not truly customers in Bartlett and Gulati’s sense. Like customers in some contexts, employers pay for the services of employees, but employers are the opposite of mere customers because, collectively, they play a significant role in the distribution of jobs and income.

  21. [21]. Bartlett & Gulati, supra note 1, at 238.

  22. [22]. Customer discrimination in the sharing economy and in other highly customer-oriented service markets may also be of concern from the standpoint of distributive justice to the extent that such discrimination can affect the social bases of self-respect. Widespread customer discrimination could operate to perpetuate the social subordination and marginalization of disfavored groups, which could be inconsistent with basic conditions of justice. For further discussion of this point, see generally Blake, supra note 19. Although Bartlett and Gulati do not explicitly focus on this argument, it strengthens their claim that antidiscrimination law should not ignore customer behavior.

  23. [23]. By the same token, the perspective of distributive justice explains why certain actors who are not customers in the ordinary sense, such as the door‑to-door salesman and the housekeeper in my earlier discussion, also might not be of particular interest to discrimination law. Their discriminatory approach to selecting their customers may not be of concern insofar as their choices do not tend to cause any unjust distribution of any important goods or resources—which should be the case, provided that other housekeepers and salespeople can come to the market to satisfy any market demand ignored by the discriminatory actors.


Associate Dean and Professor of Law, Suffolk University Law School.