|
COPYRIGHT 2001 Stanford Law School
INTRODUCTION
I have two broad goals in this essay. My first goal is to clarify, refine, and extend an argument I have made on a number of prior occasions that it is appropriate to think of the norm against discrimination in the marketplace as encompassing two quite distinct norms. The first of these norms prohibits "simple discrimination," the second demands "accommodation." Victims of simple discrimination possess what I will describe as a fairly strong, uncircumscribed "right" to be free from such treatment, while those seeking accommodation possess, in essence, a colorable "claim" on social resources that competes with a variety of other claims on such resources, a policy "argument" to be balanced against other prudential arguments.(1) The second and far more central goal is to investigate whether antidiscrimination law ought to be "group-conscious"--whether the norms against discrimination protect individuals qua individuals or protect people only by virtue of their membership in certain social groups.(2) In my view, this question can best be answered if one first recognizes the distinction between simple discrimination and accommodation. My claim is that the norm against "simple discrimination" can reasonably be understood in the first instance as a norm that protects all individuals, without regard to their membership in any particular social group. However, I argue it would nonetheless be imprudent for a variety of reasons to apply the norm to all individuals: That is to say, it may not be prudent to allow every individual, without regard to social group membership, to vindicate his interest in protection from this harm. On the other hand, I will argue that what justifies the very existence of an accommodation norm is the public interest in integrating social groups, and that the norm is therefore impossible to understand properly in the first instance without reference to the existence of social groups.
Here is the crux of the first argument, distinguishing norms proscribing simple discrimination from those demanding accommodation. Norms against "simple discrimination" assume that, at least in relationship to the conduct of market actors, putative plaintiffs possess a well-defined entitlement baseline.(3) The empowered market actors (i.e., employers, sellers of goods and services classed as "public accommodations") are duty-bound to treat those putative plaintiffs with whom they deal (job applicants, employees, would-be buyers) no worse than they treat others who are equivalent sources of money. (In this regard, a worker is essentially just her embodied net marginal product, a customer no more than a source of net receipts.)(4)
I claim, first, that the "simple discrimination" norm establishes a strong entitlement, what rights theorists would consider a side constraint on the conduct of those who would violate the norm.(5) The "right" is not, in my view, "absolute" in its formation or initial articulation. That is to say, we cannot ascertain whether or not a party ought to have the right to be tree from simple discrimination without engaging in conventional policy balancing, weighing the interests of potential defendants against those of rights claimants.(6) But once that policy balancing is done, we establish a scheme of rights that does not demand case-by-case balancing. Claims of right by one plaintiff should not be balanced against competing claims by other plaintiffs seeking similar treatment; instead, we believe all claims to be free from simple discrimination can be vindicated. Plaintiffs' claims are also not conventionally balanced against claims by defendants that it would be unduly costly to meet the plaintiffs' claims.(7) Nor are their claims (implicitly) balanced against claims that could be made by nonparticipants in the suit that they are more worthy recipients of the "resources" the defendant is expected to "expend."(8)
I contrast this norm prohibiting "simple discrimination" with the norm creating "rights" to accommodation. I define the accommodation "right" as a claim to receive treatment from a defendant that disregards some (though not all) differential input costs. (Broadly speaking, entities need not ignore additional input costs that improve service quality for "many" customers or productivity for many workers. Thus, a disabled worker might be entitled to a costly machine that would help her perform the requirements of a job only if the machine would not bolster the productivity of any would-be worker who had access to it.) If the defendant is the owner of a public accommodation, she must ignore the added costs needed to serve a customer. If the defendant is an employer, she must ignore the costs she expends that allow a worker to function in as effective a fashion as workers who produce the same gross output. I will argue that the accommodation norm establishes a distributive claim--what I also described earlier as a policy "argument" on behalf of those seeking social resources--rather than a right.(9) What I mean by that is that those seeking accommodation are making claims on real social resources that compete with all other social-resource claimants; all such claims cannot be met. Thus, any particular plaintiff's claim to have his "right" to accommodation vindicated is subject to claims that his demands are "unreasonable" in the sense that the resources that would have, to be devoted to meeting them could be spent in a better fashion.
I will explore these issues in Part I by, first, attempting to distinguish simple discrimination from nonaccommodation. Next, I will discuss why it might be difficult in both theory and practice to make the clean distinctions I attempt to make, particularly in three classes of cases. I will discuss cases in which customers prefer not to deal with members of a particular subordinated group (so that the employer, if not the bigoted customers, must sacrifice real resources to be more inclusive). I will also review cases in which the net output of the plaintiff might converge over time with the net output of a party who is currently more productive and cases in which it is costly to discover that a particular plaintiff is indeed economically equivalent to a party who has been treated more favorably. I also discuss briefly why simple discrimination claims might be deemed "prior" to accommodation claims.
In Parts II and III, I explore the question of whether the antidiscrimination norm must advert to the existence of social groups. (I will discuss in this regard both what are conventionally thought of as "social identity groups" and what might be called mere statistical aggregates, "groups" with identifiable common characteristics that lack a culture or self-conscious socio-political identity.) Clearly, the political movements that have successfully pressed state entities to use their power to restrict the legal privilege of certain market actors to make "discriminatory" decisions have always been "group"-based. It is common to see coalitions among distinct social groups pressing for antidiscrimination protection--defined by race, gender, ethnicity, sexual orientation, religion, age, or physical or mental disability. But we do not see coalitions of discrete individual consumers and workers seeking protection against such unfair treatment that make no reference to their membership in a disadvantaged social group.
It is not accidental that the political impetus for antidiscrimination legislation comes from movements acting with considerable group self-consciousness. While many of the harms of discrimination will be borne by individuals qua individuals, most people correctly believe that most "unjust" treatment is a result of group-based prejudices. (In this regard, the individualist political activist may say, more or less explicitly, "What is unjust is that I did not get this job, which I deserved, but the cause of this injustice is devaluation of `my kind' by a member of another social group.") Moreover, some of the harms of discrimination are borne by virtue of one's group membership: What is more harmful than not receiving a good or job to which one is entitled as an individual is that the exclusionary practice helps to create a disadvantaged and stigmatized caste.
Though the political impetus behind state efforts to eradicate market discrimination may well come from people acting in a self-consciously group-identified fashion, it is not obvious that the antidiscrimination norms that emerge do or should make reference to the group status of those who claim to be victimized by discrimination. It has become a question of considerable practical import whether one receives antidiscrimination protection only if one has a particular ascriptive identity. First, we now see a good deal of legal struggle played out in the courts (in the disability rights area especially) over whether particular individuals are members of classes that are expressly protected by antidiscrimination statutes. Second, there is a good deal of political struggle, played out in legislatures and referenda, over whether members of certain groups now left uncovered by such laws (especially gays, lesbians, bisexuals, and transgendered persons) ought to be covered.
I use the issues raised most sharply by court cases considering the scope of the coverage of the Americans with Disabilities Act ("ADA") to investigate whether and why norms against market discrimination do or do not need to make reference to "groups." The most critical point in Part II of the essay is that norms forbidding "simple discrimination" could be enforced without any regard to group identity. We could protect each individual against market-irrational treatment, without regard either to whether she was a member of a group that had conventionally been subject to mistreatment or to our ability to determine that the actor who treated her irrationally did so on the basis of a group-based ascriptive characteristic.
In most cases, though, there are good reasons (most administrative, some substantive) that even norms against simple discrimination are administered with groups in mind. What I explore at some length in Part II is why putative plaintiffs might seek state help in eradicating simple discrimination and why putative defendants might resist simple discrimination suits. (On their face, after all, such suits simply seem to demand that an impersonal, rational capitalist calculator engage in profit-maximizing behavior.) What I hope to illustrate is that benefits to putative plaintiffs will be most pronounced if only group members may sue and that it might be less costly for defendants to comply with the dictates of antidiscrimination law if suits can be brought solely by group members. Nonetheless, the plaintiff's most basic claim in a simple discrimination suit does not depend upon group membership.(10) In practical terms, this observation has two important implications. First, in interpreting statutes like the ADA, which both forbid simple discrimination and demand accommodation, it might be appropriate to be extremely expansive in deciding who is eligible for protection insofar as the defendant seeks only to stop the defendant from engaging in simple discrimination.(11) This may be the case even if one decides that fewer people are eligible for accommodation. Second, insofar as a previously-unprotected group seeks only protection against "simple discrimination," its claims will (and should) be heard extremely sympathetically.
However, I will argue in Part III that norms demanding accommodation are considerably harder even to conceptualize without thinking about the group status of those seeking accommodation. Moreover, to the degree to which we can even conceive of establishing an individualistic, universalized, non-group-referential right to be treated without regard to the input costs of service (in the public accommodations context) or the input costs of generating output (in the employment context), it would be nonsensical to do so.
This is so because the main reason we implicitly distribute the resources needed to accommodate otherwise excluded workers or customers to those parties, in-kind, in the form of "inclusionary" services, rather than in-cash, is to break down the hierarchical segregation of social groups. I will argue, first, that we are especially likely to believe that such in-kind transfers are most appropriate when directed to members of certain social groups rather than random individuals seeking higher levels of social inclusion. I will argue, as well, though, that there are certain limits on accommodation, if, as I believe, the basic goal of an accommodation requirement is to transfer resources, in-kind, to increase social inclusion. These limits are relevant in determining not just what accommodations are reasonable but who should be covered under statutes like the ADA that establish rights to reasonable accommodation. People who do not face something resembling social exclusion in the absence of legal intervention perhaps should not be entitled to employer-subsidized accommodation. This may be the case even when they are disabled in the conventional sense that they are plainly suffering from an unambiguous physical malady that restricts their capacity to perform tasks that most other employees could perform without the inputs the plaintiff seeks.
I. THE DISTINCTION BETWEEN ACCOMMODATION AND SIMPLE DISCRIMINATION
A. Basic Definitions
The Americans with Disabilities Act contemplates granting those people whose rights under the statute are violated two distinct remedies, each of which implicitly instantiates a distinct entitlement. First, like all antidiscrimination statutes, it clearly prohibits "simple discrimination" (differential treatment despite equality along "relevant" dimensions), permitting plaintiffs to enjoin improper refusals to serve or refusals to hire in situations in which plaintiffs can convince decisionmakers that they were denied access though "qualified" for such access. Second, it mandates "reasonable accommodation" (of relevant differences), permitting plaintiffs to secure more complex mandatory injunctions demanding that the defendant take particular affirmative steps to permit them to enjoy the relevant public accommodation or to work at the relevant job.(12)
Obviously, people may be unequal in ways some would deem relevant and others would not; thus, figuring out when a party can invoke the right to prohibit simple discrimination requires some consensus on what traits potential defendants can deem relevant. The right to demand reasonable accommodation likewise requires us to decide which distinctions should be accommodated. (As a matter of positive law, a blind lawyer may be entitled to a reader, without bearing the cost of hiring the reader; a lawyer who spells poorly or writes ungrammatically, unless deemed "dyslexic" is not entitled to a free spell-check program or an editor. As I note in a bit more detail later, the reason for the distinction is hardly lucid as a matter of distributive ethics.)(13)
There are relatively precise conventional understandings of when plaintiffs are mistreated in ways that entitle them to these remedies, though, defined in relationship to conventional understandings of the workings of a market economy. A person suffers from simple discrimination insofar as an employer (in the employment discrimination context traditionally regulated by Title VII) or a public accommodation owner (in the public accommodation context traditionally regulated by Title II of the 1964 Civil Rights Act) fails to treat him "impersonally." Insofar as the employer or public accommodation owner fails to give the employee or customer something he desires because of traits that are irrelevant to his economic function, he is breaching the duty to avoid simple discrimination. A public accommodation owner discriminates in this way if he does not treat a potential customer as well as he treats other customers who supply him the same net proceeds (money he will receive to provide a service net of the costs of service provision). An employer discriminates insofar as he treats the plaintiff employee or job applicant worse than he treats statistically typical employees or applicants whose net marginal product is no higher. (A worker's net marginal product is equal to the value of the increase in goods or services the firm will produce if the employee is added to the firm, net of the added costs that the firm will incur if she were employed by that firm.)(14) In this sense, what distinguishes a market actor's claim that there are only certain "relevant" grounds for differentiation from what are clearly closely cognate claims that any actor is entitled to "meritocratic" treatment in any realm (e.g., college admissions) is the relative consensus on how limited the logical criteria for employment and service really are.(15)
In certain circumstances, customers or employees may be entitled to reasonable accommodation in the sense that the public accommodation owner or employer has a duty to treat the customer or worker in terms of her gross, not net, value added to the firm. That is to say, the putative defendant may have an obligation to ignore the incremental input costs associated with serving the customer or insuring that the worker/applicant produces the same gross output as those who have been treated more favorably.(16)
The accommodation obligation, though, is invariably a limited one. First, the added inputs that the plaintiff seeks are unreasonable if they would benefit (large numbers of?) other customers or potential employees (nearly as much?) as they would benefit the plaintiff.(17) In this sense, the accommodation obligation is limited to those who are thought to be as "meritorious" as those who can work without accommodation. Naturally, the concept of merit that the accommodation plaintiff relies on is (at least marginally) more contested than the concept the simple discrimination plaintiff relies on. The plaintiff seeking accommodation does not claim to merit the treatment she asks for because she has the same relevant traits as the person who has received better treatment: She concedes that a business rationally differentiates workers or customers on the basis of the differential input costs associated with serving them. Instead, she argues that her "talent" is defined by her capacity to produce, and that her capacity to produce is measured by the output she can generate without using aids that benefit workers generally. The metaphor is one of athletes competing in a contest: A "disabled" pole vaulter who vaults as high as his competitors using the especially expensive shoes he needs that would not benefit other vaulters is "as good" a vaulter as those using ordinary shoes. One who can do as well only by using an expensive pole made of a strong, flexible material that would improve any pole vaulter's performance is not "as good."(18)
Second, the cost of these atypical inputs must not be unduly high: It must be "reasonable" in that sense. In this sense, the accommodation obligation is limited by the fact that we must expend real social resources to meet it. (I return to discuss the fact that the defendant who wishes to engage in simple discrimination must sacrifice private, psychic utility if asked to desist, but not physical resources. His discriminatory desires are not representative of society's desires; we wish no one had those tastes. The defendant resisting demands to accommodate attempts to save real resources. In that sense, his desire to save resources is representative of general social desires to save resources. We do not wish to abolish the taste to save such resources, even if we believe in a particular case that the best use of the resources is to use them to accommodate.) Because we must expend real resources to meet the demand for accommodation, we compare the value of expending the resources to meet the policy goals of accommodation with the value of expending the resources to meet other social policy aims.
Two straightforward illustrations might help differentiate accommodation and simple discrimination claims. The simple antidiscrimination principle would preclude a dentist (as public accommodations provider) from refusing to treat a hearing impaired patient, so long as his inability to communicate with the patient neither affected the price the patient would pay nor the cost of serving him. The accommodation principle would require that the dentist take steps to be able to communicate with the hearing impaired patient, if necessary to provide her with the same quality care he gives other patients, without charging the patient the incremental costs of treating her.(19) This is true even though a simple nondiscriminating, impersonal, capitalist calculator would refuse to treat a patient who is atypically costly to serve unless permitted to charge more for the services in the absence of a supplementary duty to accommodate.(20)
In the employment context, the conventional antidiscrimination norm forbids an employer from refusing to hire a blind lawyer who can do the same legal work as a sighted one. The accommodation principle demands that the employer not reduce the blind lawyer's pay if he requires a ("reasonably" costly) reader to generate the same work that sighted lawyers do without an aide.(21)
It is vital to note that it is often difficult to determine whether a particular plaintiff claims to be the victim of simple discrimination or whether she claims instead that she is entitled to a reasonable accommodation. When, for instance, an employee asks an employer to adjust his work schedule, he is claiming first that his net output on the adjusted schedule is no lower than that of fellow employees following the more conventional schedule. Thus, the refusal to hire him on the reduced schedule is a form of simple discrimination.(22) He will often argue in the alternative, though, that if his net output is indeed lower, it is nonetheless reasonable to ask the employer to bear the costs associated with the net productivity shortfall because they must be borne if the plaintiff is to work.
B. Problematizing (Yet Maintaining) the Distinction
Determining whether we are adjudicating a simple discrimination claim rather than an accommodation claim is not always easy. There are three classes of cases that pose particularly troublesome difficulties. First, the plaintiff might concede that he is costlier to serve or employ so long as the organization is insensitive to his distinct needs, but the long-run costs of serving him or employing him are not atypically high. In the long run, the failure to include him would seemingly represent simple discrimination, but there may be short-run "accommodation" costs associated with moving to the long-run equilibrium. I have referred to this in the past as the problem of "dynamic discrimination."(23) Thus, for instance, an employee who concedes that she might be less productive, now, because some co-workers have difficulty understanding her given her statistically uncommon accent may argue that coworkers will adjust quickly to distinct accents. She might reasonably believe this is especially true if the employer hires a diverse workforce that becomes accustomed to the presence of a multiplicity of voices.(24)
Second, the defendant might claim that it is costly to employ an ostensibly qualified employee because customers (or co-workers) will refuse to deal with the applicant. Whether the clear rules we enforce forbidding employers from accounting for customer preferences should be thought of as enacting an accommodation requirement in routine Title VII cases or should be thought of as forbidding customers from engaging in simple discrimination is a reasonable question. Third, if it is costly for a putative defendant to discover that the plaintiff is in fact economically indistinguishable from parties he treats better, it appears from the plaintiff's viewpoint that she is a victim of simple discrimination (i.e., she is treated more poorly than those with identical economic virtues) unless we hold her accountable for the costs of discovering her true traits. But it is clear that the gains to the firm from employing her, net of the costs of discovering her virtues, are lower than they are for parties whose virtues are more cheaply revealed. Title VII sometimes permits cost-saving statistical discrimination and sometimes forbids it: When it forbids it, I believe it is demanding accommodation.
1. Dynamic discrimination.
Assume that a plaintiff in a job discrimination suit persuades the fact-finder that she has the capacity to produce just as much product, net of the incremental costs of employing her, as do those people the employer currently employs. She concedes, though, that given current workplace arrangements, her net output is lower. If the employer could shift costlessly from the current regime to the regime the plaintiff favors, it would appear that she is alleging that she is the victim of simple discrimination. (Think in this regard of cases in which prisons refused to hire women guards in men's prisons because all guards performed privacy-invasive full body searches. If the prison could readily concentrate search tasks on a subset of guards, without bearing administrative costs in making appropriate assignments, it could hire some female guards without bearing costs.)
If, however, the employer would have to bear transitional costs in establishing the regime that eliminated productivity disparities, one could classify these transitional costs as accommodation costs, or argue instead that the plaintiff simply sought to establish a regime in which no simple discrimination occurred. Typically, if the need to bear the transition costs were a function of prior employer discrimination, one would be most prone to think of the plaintiff as raising a simple discrimination claim. (Imagine the following sorts of cases in this regard. An employer never hired women, as a result of simple discriminatory policies; if it hires women now, it must build moderately expensive restroom and changing facilities for women. Had it hired women all along, though, the costs of having male and female facilities would not have been higher than the costs of establishing larger all-male facilities. Imagine, as well, that an employer purchased keyboards for word processors that cannot be used by those with repetitive stress injuries. If it hires some workers with such injuries, it must retire some keyboards before their useful life has run. But the keyboards that do not aggravate the disorder are no more expensive than the keyboards the employer in fact ordered.)
If, however, the employer must bear transition costs to undo the discrepancy between the plaintiff's net product and the net product of those who have been hired, though her prior decisions did not exclude the plaintiff without reason, then it would seem that from the defendant's view, we would be dealing with an accommodation case. (Imagine in this regard that the defendant trucking company purchased equipment without power steering, because it was cheaper when purchased. Power steering is now no more expensive however. The defendant will have to retire equipment prematurely if it is to employ more women and more people with back ailments.) The transitional cost (resulting from premature retirement of usable equipment) may appear to be an accommodation. Plaintiffs who are benefited, though, will be just as net productive, prospectively, as those they compete with, something that is not true in the more typical accommodation case in which the gap between the net productivity of the plaintiff and those more favorably treated is ongoing. Because it is possible to describe these cases as accommodation cases (from the defendants' view) or as simple discrimination cases (from the vantage point of the plaintiffs' current qualities), one would expect them to be difficult cases for courts that believe that they are interpreting statutes, like Title VII, that seemingly demand only an end to simple discrimination.
2. Customer preferences.
Assume that a retailer correctly believes that some of his customers will refuse to deal with a woman salesperson, or an African-American, though that employee might perform all "physical" aspects of the job in the precise manner that a white, male employee does. If the employer must hire the woman, it is arguably the case that he is being asked to "accommodate" her, rather than to not engage in "simple discrimination." This is true not simply in the obvious sense that the employer must sacrifice sales volume (and profits) if he hires a salesman with whom customers will not deal. (In this sense, his decision is not grounded in animus or false stereotypes but in private economic rationality. He is treating the woman or African-American applicant solely as a source of profits to the firm.) It is also true in the more relevant sense that he must expend real social resources to produce the same level of services. (Presumably, if he wants to sell as much, he will have to hire more salespeople since the "unwanted" ones will be partly idle.)
As a matter of positive law, it is clearly the case that an employer cannot refuse to hire as a result of such racist or sexist customer preferences.(25) I have long claimed that customer preference cases are not best thought of as accommodation cases but as simple discrimination cases.(26) Customers are duty-bound under antidiscrimination law to treat salespeople with whom they deal in an impersonal, capitalist, rational fashion (with regard only to their economic function, without regard to their status). The customer will bear only psychic losses if he is asked to stop discriminating. He will bear no resource costs if he must deal with an "unwanted" employee; it will take him no longer to get the goods he wants, nor will he be less informed about the nature of the product he is going to use. The employee's correlative right is vindicated, though, not by suing customers, but by forbidding those who hire the people who deal with customers from raising "customer preferences" as a defense. The employer, in essence, acts as an agent of customers. (The "real" employer of a shoe salesman is the shoe store customer; the shoe store manager simply intermediates between customers and salesman.) An employer would, in essence, manifest the customers' impermissible market-irrationality if she were able to say that she refused to hire those that customers would not deal with, not because of her own market-irrationality, but because profits would decline if she hired an unpopular salesperson. If such a claim is not permitted, the customers' irrationality will have no impact.
It may be of little practical moment whether customer preference cases are "really" accommodation cases or "really" simple discrimination cases: It is clear that the employer is forbidden from accounting for such preferences in any case. Still, the alternative interpretations may ultimately have both analytical and practical significance: The question of whether "accommodationism" is a newer, separate (implicit) social spending program or a core aspect of the traditional antidiscrimination regime surely has at least rhetorical bite.
Those who claim that these are accommodation cases are prone to believe that there is no strong affirmative entitlement to "market rational" treatment from those who deal with us in the "public market sphere." Instead, they would claim, the market antidiscrimination norm essentially derives from the norm against status-based decisionmaking by the state. (The liberal democratic state, which is peculiarly prone to capture by dominant ethnic groups, binds itself to be neutral among citizens.) In this view, this norm against group favoritism covers employers because Legal Realists came to believe that powerful private actors, like the state, are thought to have a certain degree of public power that must be controlled.(27)
This view strikes me as highly misleading. Title II of the 1964 Civil Rights Act covers public accommodation owners far too small to...
Read the full article for free courtesy of your local library.
|
Bookmark this article
Print this article
Link to this article
Email this article
Digg It!
Add to del.icio.us
RSS
