- I. Introduction
- II. A New Conceptual Framework
- III. Disparate Treatment Doctrine Asymmetrically Disfavors Workers of Color and Helps Maintain Racial Hierarchy
- IV. Thinking About Solutions
- V. Conclusion
Calling social dominance by more palatable names, pretending that it is only a feature of other people’s societies, assuming that it is due only to the actions of a “misguided” few, and presuming that it is merely a dying legacy of the past not only are exercises in self-delusion, but also contribute to the tenacity of group dominance by obfuscating its very existence, and thereby making it that much more difficult to change.
We judge it as considerably more harmful to the cause of equality and the fulfillment of democratic ideals to pay too little than too much attention to the dynamics of group dominance.
Imagine that you are a black employee who is vying for a promotion. You have worked hard trying to rise through the ranks. You have assiduously applied for more attractive jobs, but you have been out of luck for a promotion because those jobs were at other facilities, and the employer has a policy of filling vacancies from within the same facility where possible. But now, a position has opened up in your facility, and it needs to be filled quickly. You think it is finally your turn: you are the only person on the candidate list who is already working at the facility, and you are clearly qualified. If everything goes by the book, you have finally earned yourself a promotion. You are also the only black applicant for the job.
But you don’t get the job. Instead, you learn, a white male got the job. What’s more, the person who got the job was not on the initial candidate list. He was not even qualified, and thus not eligible, for the job based on the original job description. Yet, you learn that, after finding out that you would have gotten the job under existing procedures, the hiring manager ordered the job description to be rewritten. Magically, the other candidate is now the only person remaining on the new eligibility list and will be hired. You are frustrated, upset. You can’t help but think that your race was the reason that you did not get the promotion. You know Title VII of the 1964 Civil Rights Act prohibits race discrimination in employment decisions. Surely, you think, if Title VII was written to protect anyone, it would be you. Fool me once: you are wrong. The district court judge agrees with you that you were more qualified. However, the court tells you that it was not your race that caused you not to get the promotion; it was that the manager engaged in favoritism for his “fishing buddy”—favoritism ostensibly unconnected from any racial motivation. That is a decision that the employer can legally make because favoritism unconnected to race is not illegal. So, you lose your case. You are confused. At trial, the manager had steadfastly denied any favoritism, so you think he should be held to his testimony under oath. But the district court didn’t believe those denials and concluded that favoritism is the better explanation for what happened. You hope that the appeals court will correct what seems like a clear error. Your hope is misplaced. While the court tells you that it is troubled and might have reached a different conclusion on a clean slate, it also says that it is bound by what it sees as acceptable judgment calls by the district court in interpreting the evidence. “Title VII does not have a limitless remedial reach,” the court says, and you are not within its existing reach.
Now imagine that you are a black employee of an employer in financial distress, hoping not to get laid off. You are somewhat hopeful. You are the only black employee in an administrative position, and your employer has committed itself to racial equity, inclusion, and diversity through an affirmative action plan that aims to achieve and protect an equitable racial composition of its workforce. But the financial situation is bad. The employer must lay off numerous people in your job category. You are initially one of them, but like all of the laid-off employees, you get a hearing to contest your layoff. After further review, the employer decides that your layoff would improperly set back its goals of racial equality in the workforce and decides to retain you after all. A white employee, who was in some respects senior to you but was laid off, sues, claiming that the decision to retain you but not her was illegal race discrimination under Title VII. Surely, you think, it must be possible for an employer to voluntarily try to retain its only black administrative employee in an effort to have a more racially equitable workforce. Fool me twice: you are wrong again. The court decides that retaining you is not a decision that your employer can legally make under Title VII. The employer could only do so if it had discriminated against black employees in the past, and there was not enough evidence of that. For the court, because there are few blacks in the local labor force, it really wouldn’t be significant for the employer to have no black employees at all. What’s more, retaining you to protect the employer’s affirmative action progress was illegal because it “unnecessarily trammel[ed]” the interests of the white employee, who the court viewed as the one who was really entitled to your position.
Imagining yourself as the worker in the first scenario above puts you into the shoes of one of the most common types of litigants in the American civil justice system. Employment discrimination cases are consistently among the most frequently litigated types of civil cases in the federal courts. By imagining your frustration about not receiving any relief, you also share in the feeling most commonly experienced by plaintiffs in these cases. Employment discrimination cases are notoriously difficult to win for plaintiffs. Employment discrimination plaintiffs fare more poorly than other civil plaintiffs in the district courts, and even if they succeed at the trial court level, they are extremely likely to have their victories overturned on appeal. This pattern appears to have been in place for decades now. Claims of race discrimination are perhaps the most frequently litigated type of employment discrimination case, and there are some indications that race discrimination plaintiffs fare particularly poorly in the courts. Racial minorities, in particular black plaintiffs, are the prototypical plaintiff in such cases. Black plaintiffs bring most employment discrimination cases claiming race discrimination. Thus, the apparent antiplaintiff bias of federal judges in employment discrimination cases disproportionately affects black plaintiffs, who lose most of their cases. This is so even though a wide variety of data shows that racial inequality and racial discrimination against workers of color continue to pervade the American economy. In the face of such continued discrimination, one might think that employment discrimination law should be equally, perhaps more, hospitable to claims by workers of color compared to other kinds of civil plaintiffs, not less. Yet, it is decidedly not.
The shape of employment discrimination doctrine under Title VII of the 1964 Civil Rights Act, the federal statute governing race discrimination in employment, significantly contributes to this lack of success by discrimination claimants. Scholars have criticized the doctrine as a complicated “morass” that is in significant part the outcome of restrictive judicial interpretation of broad statutory language (especially since the 1980s), notwithstanding intermittent legislative intervention meant to broaden workers’ antidiscrimination protections. This restrictive interpretation has created a doctrine that in many of its specifics, including in the race discrimination context, discredits discrimination claims of plaintiffs, who are predominantly workers of color.
This is deeply troubling. After all, the Supreme Court has repeatedly reminded us that one of the main goals of Title VII was to create lasting improvements in the working conditions of racial minorities, in particular black Americans, and to ensure equal opportunities for minority workers to succeed in the workplace. In interpreting and applying Title VII race discrimination law, the federal judiciary (including the Supreme Court itself) has not delivered on these goals for decades now. The first vignette above, taken from a real case, is illustrative of the resulting “fool me once” problem for racial minority plaintiffs: despite being the primary intended beneficiaries of Title VII’s prohibition of race discrimination, they have terrible chances of success even when they have strong evidence of such discrimination.
Given the crucial importance of employment to most people’s livelihoods, especially for racial minorities who already disproportionately suffer from social and economic stratification, it is important to uncover the reasons underlying this longstanding disconnect between asserted statutory goals and interpretive practice so that effective countermeasures can be developed. Yet, comparatively few scholarly analyses, particularly outside of the implicit bias literature, have tried to do so by combining traditional doctrinal and theoretical analyses with systematic empirical insights from social science.
The first major goal of this Article is to offer an analytical framework that takes on this task in a novel way and explains the federal judiciary’s cramped interpretation of Title VII race discrimination law based on conclusions from social psychology. This framework combines findings from Social Dominance Theory (SDT)—a social-psychological theory with a considerable footprint in social science journals but that has yet to significantly enter legal scholarship—with insights from Critical Race Theory (CRT) and conceptualizes the interpretation of race discrimination law as a mechanism through which racial hierarchy is regulated and, in the main, (re-)produced. SDT researchers have uncovered robust evidence that basic human psychological preferences for society to be organized as a group-based (including racial) hierarchy can influence different groups of people to think and behave in ways that help to create and maintain such hierarchy. This Article argues that there are good empirical reasons to think that federal judges as a group (with individual variation to be sure) have comparatively strong such preferences, and that they will act in accordance with those preferences when interpreting hierarchy-relevant laws such as Title VII’s race discrimination provisions. These preferences influence, among other things, people’s baseline assumptions and perceptions of the prevalence of discrimination against different groups, their views on the desirability of antidiscrimination intervention, as well as the ideologies people use to explain the world around them. This Article analyzes Title VII race discrimination doctrine with reference to these psychological processes and shows how they explain the doctrine’s suspicion towards claims of discrimination by workers of color, its many rules that discredit their allegations, and why it rejects the vast majority of their claims as a result. Uncovering the interpretation of employment discrimination law as a process at least partially driven by the judiciary’s preferences for group-based social hierarchy, in turn, has important implications for potential reform, including illustrating the likely benefits of increasing judicial diversity of various kinds.
The second major goal of this Article is to show that its analytical framework not only helps us better understand the ongoing judicial interpretation of employment discrimination law, but also illustrates that popular descriptions and conceptualizations of the overall logic of Title VII jurisprudence are inaccurate and potentially unhelpful. Taking a hierarchy-centered view of Title VII doctrine counsels against accepting what seems to be a relatively broad consensus that (1) Title VII in general, and its race discrimination provisions in particular, are “symmetrical” (i.e., they protect all groups covered by the law in the same general fashion); and (2) to the extent that the doctrine is asymmetrical (i.e., provides certain protections to some but not other groups), it is asymmetrical in favor of racial minorities because it allows for affirmative action. This Article argues that, instead, Title VII race discrimination jurisprudence, its affirmative action doctrine included, should be understood as being fundamentally asymmetrical to the disadvantage of workers of color while protecting the interests of white workers to a greater extent.
This is because, consistent with what this Article’s analytical framework predicts, the federal judiciary has interpreted Title VII in ways that make it comparatively easy to prevail for litigants whose interest is to preserve existing racial hierarchy in the workplace (most employers and white workers) and comparatively difficult for litigants whose interest is to reduce existing hierarchy (employers who pursue affirmative action and workers of color). As a result, Title VII race discrimination law is neither always hostile to plaintiffs nor always friendly to employers. Rather, it is hostile to plaintiffs in “traditional” race discrimination cases like the first vignette above, which prototypically challenge racial hierarchy because they are predominantly brought by workers of color. By contrast, it provides greater comparative protection to plaintiffs in affirmative action cases (prototypically white workers) in which plaintiffs aim to prevent interference with existing hierarchy, as shown in the second vignette above. Similarly, employers are treated deferentially when defending “traditional” cases, but with greater hostility when defending affirmative action programs. One common source for all of these asymmetries is a “baseline error”—an often unspoken assumption, counter to the weight of the evidence but consistent with Social Dominance Theory, that discrimination against workers of color is not sufficiently frequent to warrant particular attention, and that discrimination against white workers warrants just as much (if not more) judicial concern. In all cases, however, workers of color receive the short end of the stick because the doctrine consistently undervalues their interests. Thus, such workers disproportionately miss out on employment opportunities, and racial hierarchy in the workplace persists.
Given the Supreme Court’s repeated command that one of the main purposes of Title VII is the elimination of racial hierarchy in the workplace, and the fact that Title VII grew out of the Civil Rights Movement’s efforts to end the subordination of workers of color, this is highly problematic. Research findings from SDT and the framework proposed in this Article help us explain why we should nevertheless not be surprised to see the existing doctrinal landscape. They also help us better evaluate potential unintended negative consequences of reform proposals that proceed from a view of Title VII doctrine as symmetrical or, if anything, asymmetrical in favor of workers of color.
Overall, then, there are several gaps between Supreme Court rhetoric about what Title VII is meant to achieve and the actual interpretive practice of both the Supreme Court itself and the lower courts. These gaps manifest themselves in doctrinal asymmetries that consistently under-protect the employment opportunities of workers of color and can be uncovered by a close interrogation of doctrinal rules that apply to different types of cases. It is important to uncover these asymmetries and their social psychological origins so that we can begin to develop countermeasures that help us move closer to the stated goals of federal employment race discrimination law. This Article speaks to all of these steps.
Part II provides an overview of this Article’s main analytical framework. It begins in Section II.A with an overview of Social Dominance Theory and lays out SDT’s claim that human societies develop as, and predictably remain, group-based social hierarchies. These hierarchies are aided in their stability by a complex interplay between individual psychology, individual and institutional behavior, and social and cultural ideologies. The law can be understood to play a critical role in this process. I propose two analytical heuristics that place the interpretation of antidiscrimination law by the courts into the SDT framework. I call these heuristics asymmetrical narrowing and proof asymmetry. I also discuss important contributions that SDT can make to a more comprehensive understanding of employment discrimination law. Section II.B gives a brief overview of the concepts of “framework critique” and “baseline error” that scholars in Critical Race Theory have developed, and which integrate productively with SDT.
Part III applies the framework of Part II in analyzing Title VII doctrine and uncovering its racial asymmetries. In particular, I focus on disparate treatment doctrine regulating race discrimination claims. I first analyze the doctrine that applies to the claims prototypically brought by racial minorities as plaintiffs, i.e., cases like the first vignette above. In particular, I focus on the burden-shifting regime established by McDonnell Douglas Corp. v. Green and, using the framework from Part II, show how its current doctrinal logic is fundamentally asymmetrical to the detriment of workers of color. I then compare such “traditional” disparate treatment cases to Title VII affirmative action doctrine, i.e., the doctrine that applies to cases like the second vignette above. This doctrine is based on a modification of the McDonnell Douglas scheme, but an analysis of its rules shows how Title VII operates much more favorably to plaintiffs when they are prototypically white claimants challenging interference with existing hierarchy. Both areas combine to preserve existing racial hierarchy in the American workplace. The framework developed in Part II provides a powerful aid in explaining why this is so.
Finally, Part IV begins a conversation about the implications of my analysis for potential solutions to the problematic state of employment discrimination law. I discuss some of the proposals for doctrinal reform that scholars have already offered and argue that they risk unintended consequences that may further solidify aspects of existing racial hierarchy. Their implementation could also undermine support, and a coherent ideological basis, for more effective remedies—most prominently broader affirmative action programs. Accordingly, Part IV takes a first cut at offering some alternative suggestions for reform. Some could be implemented more short term, such as modifying previous doctrinal choices. Others recognize that any lasting progress towards reducing the negative effects of racial hierarchy, and the law’s complicity in it, will have to include more structural changes. Such changes could include, most prominently, broadening the pipeline into the federal judiciary in a hierarchy-conscious manner.
II. A New Conceptual Framework
A. An Overview of Social Dominance Theory
1. Basic Conceptual Structure.
Social Dominance Theory starts from the observation that, historically, all human societies producing economic surplus have organized as systems of group-based hierarchy. While the specific form and ideological foundation may differ between societies and over time, the fact that human societies are structured as group-based hierarchies in which one or more social groups are dominant, and one or more social groups are subordinated, is a general feature of societies throughout history.
According to SDT, this group-based hierarchical structure is based on three different stratification systems: (a) an age system in which adults have disproportionate social power over children; (b) a gender system in which men have disproportionate social power over women; and (c) an “arbitrary-set” system which implements social hierarchy along any number of socially constructed group categories that may be salient in a given society’s context, such as social class, caste, race, and others. In the United States, “for most of U.S. history, race . . . has been and remains the primary basis of social stratification,” and is also the primary focus of this Article. Arbitrary sets are typically associated with the most extreme forms of violence against subordinated groups and are more flexible in how they manifest themselves and change over time.
SDT sees societies as dynamic systems in which group-based hierarchy is the net effect of behavior and decision-making at various societal levels, including aggregated individual discrimination, aggregated institutional discrimination, and behavioral asymmetry among groups at different levels of the hierarchy. Behavioral asymmetry refers to group differences in behavior between dominant and subordinated groups as a result of their fundamentally different social circumstances, such that dominant groups “behave in ways that are more beneficial to themselves” compared to subordinate groups. These processes are mediated by legitimizing ideologies or “myths,” i.e., “attitudes, values, beliefs, stereotypes, and ideologies that provide moral and intellectual justification for the social practices that distribute social value within the social system.”
People’s behavior within any given society is partially driven by individual differences in people’s support and preference for group inequality and for social systems to be structured as group-based hierarchies. This individual difference is captured by a construct called Social Dominance Orientation (SDO). Although SDO is an important analytical concept within the theory, SDT does not explain societal inequality only based on individual differences or considers SDO as a “pathological” condition that permanently divides “good” from “bad” people. Rather, SDO is seen as an attribute “thought to reflect normal human variation and to be influenced by a combination of socialization experiences, contextually sensitive material and symbolic interests (e.g., high social status), and dispositional differences in factors such as aggression and lack of empathy.” In other words, SDO is a complex and interactive variable that is sensitive to context and acts as “both an effect and a cause of intergroup attitudes and behaviors.” Still, SDO has been found to be a “relatively stable cause of prejudice against outgroups and legitimizing ideologies helping to justify opposition to social policies beneficial to subordinate groups.” Moreover, SDT has made a number of robust findings regarding the general distribution of SDO among different groups in society. Relevant to this Article, SDT has consistently found that members of dominant social groups, particularly dominant racial groups (whites in the United States) and men, on average have significantly higher levels of SDO than subordinate racial groups and women. Thus, on the whole, members of such dominant groups will be more supportive of group-based hierarchy than women and racial minorities. As discussed below, this has significant implications in analyzing the federal judiciary and its decision-making in hierarchy-relevant contexts such as employment discrimination.
SDT is fundamentally concerned with the dynamic interaction between individual differences, social ideologies, discrimination at the individual and institutional level, and behavioral asymmetries between dominant and subordinate groups, and how the interaction of these factors creates, influences, and stabilizes group-based hierarchy. In this dynamic model, the development and maintenance of group-based hierarchy are not seen as stemming from a single root cause, but rather as the result of a system in which individual parts are both a cause and effect of other parts, mutually influencing and reinforcing one another.
Finally, in the SDT framework people’s behavior, societal roles, legitimizing myths, and social policies can take on two forms, depending on whether they manifest support for the maintenance or enhancement of group-based hierarchy (“hierarchy-enhancing” or “HE”) or whether they manifest support for the reduction or elimination of this hierarchy (“hierarchy-attenuating” or “HA”). The level of hierarchy in a society depends on the net balance of hierarchy-enhancing and hierarchy-attenuating forces (see Figure 1).
Within this general conceptual structure, SDT has developed more specific claims which are relevant to the arguments in this Article.
2. A Closer Look at Legitimizing Ideologies.
Legitimizing ideologies or “myths” play an important role in SDT because they operationalize the idea that power and hierarchy are more effectively maintained (especially over the long-run) not with brute force, but through consensual ideology. Legitimizing myths “consist of attitudes, values, beliefs, stereotypes, and ideologies that provide moral and intellectual justification for the social practices that distribute social value within the social system.” A legitimizing myth can be hierarchy-enhancing when it “enhance[s] social hierarchy by justifying it or the practices that sustain it,” or hierarchy-attenuating, when it “reduce[s] social hierarchy by delegitimizing inequality or the practices that sustain it, or by suggesting values that contradict hierarchy (e.g., inclusiveness).”
Legitimizing myths “mediate the relationship” between people’s individual preference for group-based hierarchy (i.e., SDO levels) and support for social policies. That is, the higher a person’s SDO level, the more likely that person is to support a HE social policy, and this relationship is explained by the person’s support for the HE legitimizing myth(s) that is mobilized to justify this policy. Similarly, the lower a person’s SDO level, the more likely that person is to support a HA social policy, and this relationship is explained by the person’s support for HA legitimizing myth(s) that justify the policy (see Figure 2).
Further, given that the content of ideologies is often indeterminate to some extent, the type of rationale underlying a particular ideology could be either hierarchy-enhancing or hierarchy-attenuating. SDO levels should correlate positively with HE rationales and negatively with HA rationales—even for rationales ostensibly describing the same ideology. What is important about legitimizing myths is not their “objective” truth, but rather the “degree to which people believe the statement to be true, and the role that this belief plays in providing moral and intellectual support for [group, e.g.] racial domination.” SDT argues that the relationship between SDO and legitimizing myths differs between social groups and that “the social attitudes and policy preferences of dominants are more strongly motivated by the desire to establish and maintain social hierarchy than those of subordinates are.” In other words, on the whole, members of dominant groups, such as whites and men in the United States, are more likely to subscribe to ideologies that provide support for social practices that help perpetuate their social dominance, and to do so more strongly.
3. Sorting People into the Hierarchy.
The dynamic processes that lead to systems of group-based hierarchy also include the sorting of different types of people into different social roles. For example, through a process of self-selection, hierarchy-enhancing roles or roles in HE institutions—which include profit-maximizing institutions and corporations and the police—are more likely to be filled by people expressing greater support for group-based hierarchy (i.e., those high in SDO). The opposite is true for HA roles and institutions—which include civil rights groups, welfare organizations, or public defender’s offices. SDT also argues that social institutions will in general be organized such that power and authority will disproportionately be wielded by dominant group members. That is, according to the so-called increasing disproportionality hypothesis, “the more political authority [is] exercised by a given political position, the greater the probability that this position will be occupied by a member of the dominant group.”
Importantly for purposes of this Article, SDT researchers have uncovered evidence that processes of institutional socialization can affect people’s views regarding social dominance and hierarchy. There is even some evidence (though not from the United States) that relates such processes to lawyers and the study of law. In particular, Guimond and colleagues found that (1) law students in France had significantly higher levels of SDO than psychology students in general; and (2) there was evidence of a socialization process resulting in upper-level (third or fourth year) law students to have higher SDO levels than first-year law students, while upper-level psychology students had lower SDO levels than first-year psychology students. In other words, there was evidence for both self-selection and institutional socialization, so that exposure to a HE environment in law school increased law students’ preferences for group-based hierarchy, which were comparatively high to begin with. Further, the authors found experimental evidence suggesting that being assigned or found qualified for a position of authority and high status might itself increase SDO, which in turn is related to higher prejudice towards outgroups. For example, in one of Guimond’s experiments, study participants, assigned at random to receive feedback that they “have the profile of a person who is able to lead and to hold position of high responsibility,” showed significantly higher levels of SDO than participants told that they had average such ability, as well as significantly higher antiblack prejudice. In another experiment, people, assigned at random (and told that they had been assigned at random) to a position of director in a company, showed higher levels of SDO than those assigned at random to a position of secretary, as well as higher bias against North Africans. Thus, the authors found the evidence “consistent with the hypothesis that being promoted to a dominant social position has a definite impact on SDO.” Still, the authors also noted that “this is not to suggest that leadership as such, or any position of power in itself will produce an increase in prejudice,” but, rather, that “it is only socialization in a position in an [HE] environment that is expected to increase [HE] legitimizing myths (e.g., racism, sexism, conservatism).”
4. SDT, the Law, and Judicial Interpretation.
While the law’s general position within the SDT framework has not been theorized extensively by social psychologists, the interpretation and application of the law can be understood as a critical component that contributes to the existence and stability of systems of group-based hierarchy. As SDT scholars have noted, in societies that have “democratic and egalitarian pretensions,” but, nevertheless, are characterized by clear group-based hierarchy, such as the United States, the law will be one of the major mechanisms that gives this hierarchy “plausible deniability, or the ability to practice discrimination, while at the same time denying that any discrimination is actually taking place.” I argue that within the SDT framework, laws and legal doctrines occupy a similar position to that of “social policy” in Figure 2, since they also implement rules regarding the distribution of resources and boundaries of appropriate behavior based on ideological justifications. In other words, taking Figures 1 and 2 together, I argue that the law mediates the relationship between personality and ideology on the one hand, and individual and institutional behavior, which in turn structures societal outcomes, on the other.
Thus, given the historical stability of group-based hierarchies, on balance and over time the law should function to preserve group-based hierarchy. This is not to say that the law will, without fail and in all cases, be structured in a way that supports group-based hierarchy. As with other factors in the SDT framework, the law can also be either hierarchy-enhancing or hierarchy-attenuating. In circumstances in which those with lower SDO can mobilize HA forces to a greater degree than countervailing HE forces, for example, legal regimes may emerge that partially reduce group-based hierarchy. However, on balance, HE legal regimes should predominate. With respect to race in the United States, as many CRT scholars have argued, the law (including employment discrimination law) has been a primary site in which racial inequality and hierarchy has been both contested but also, in the main, legitimated.
I argue that the conceptual framework of SDT and its empirical findings can, and should, be applied to inform our analysis of judges’ ongoing interpretation of the law. Specifically, I argue that a synthesis of the above findings suggests that the interpretation of the law by judges, particularly when it involves hierarchy-relevant issues such as employment discrimination disputes involving claims of race discrimination, will be guided by the process of asymmetrical narrowing, and the complementary sub-process of proof asymmetry.
Asymmetrical narrowing means that judges should, all else being equal, interpret legal provisions in a way that takes a narrow view of substantive rights likely to be claimed by members of subordinate groups, and a comparatively broad view of rights likely to be claimed by dominant groups. This could take the form, for example, of developing a relatively “defendant-friendly” doctrine in types of cases that are most often brought by members of subordinate groups, but a relatively “plaintiff-friendly” doctrine in cases most often brought by dominants. Alternatively, this could take the form of favoring the interests of dominant groups when substantive law requires a balancing of interests between dominant and subordinate groups. Proof asymmetry is one of the primary ways in which asymmetrical narrowing will be achieved. It refers to the idea that judges will develop legal tests and regimes that make it comparatively easy for hierarchy-enhancing litigants to prove or defend their case, particularly when it supports a HE outcome, and comparatively difficult for hierarchy-attenuating litigants to prove or defend their case, particularly when they are seeking a HA outcome. Importantly, this means that the relative treatment of one and the same litigant, such as a corporate employer, should depend to a significant extent on the relationship between the employer’s litigation position and the maintenance of racial hierarchy—i.e., whether the litigation position is HE or HA.
Expectations that judges should engage in asymmetrical narrowing and proof asymmetry in hierarchy-relevant areas, such as the interpretation of race discrimination law, can be derived from various aspects of SDT’s framework.
For example, per the increasing disproportionality hypothesis, the judiciary should be composed mainly of members of dominant groups, for whom the re-creation of group hierarchy is most in their self-interest. A federal judgeship is a powerful and prestigious position, and thus can be expected to be occupied mainly by members of dominant groups—particularly white males. This is, and always has been, the case in the federal judiciary. While undoubtedly there are dominant group members that are opposed to group-based social hierarchy, and some will be part of the judiciary, SDT research suggests that, on the whole, their “social attitudes and policy preferences . . . are more strongly motivated by the desire to establish and maintain social hierarchy.” This suggests that, as a group, federal judges will have more of an interest to preserve hierarchy than to subvert it.
More specifically, as most judges are from dominant groups, have been assigned to a significant position of leadership, and, through the appointment process, have been given feedback that they are, in fact, highly qualified for it, they can also be expected to have—on the whole, and with individual variation to be sure—comparatively high levels of SDO. As discussed above, this should be particularly true for white male judges. These higher SDO levels should lead judges to interpret race discrimination law in hierarchy-enhancing ways. This is because higher levels of SDO are not only associated with greater general support for group-based hierarchy, but researchers have also found a strong direct relationship between SDO and opposition to antidiscrimination policies, including in the employment context. For example, researchers have found that people with higher SDO levels are more likely to disagree with statements like “Society should make sure that minorities get fair treatment in jobs” and “We need to take more action to help stamp out the subtle discrimination that members of certain social groups still face.”
As discussed in more detail below, this relationship between SDO and opposition to antidiscrimination policy is influenced by the fact that people higher in SDO perceive less inequality to exist in society. Higher levels of SDO have also been connected to perceptions that greater racial progress has been achieved since the 1960s, and that racial remediation is a zero-sum affair in which gains for minorities must involve losses for whites. Higher SDO also seems to lead people to dislike black, but not white, workers claiming to have been discriminated against, likely because “discrimination claims made by members of racial minority groups potentially question the legitimacy of the racial hierarchy, [and thus] perceivers relatively high in anti-egalitarian sentiment react in a particularly negative manner toward such claimants.” As Part III explains, such hierarchy-enhancing views about the “facts” of discrimination underwrite significant parts of why current workplace race discrimination doctrine is so hostile to the interests of workers of color, but comparatively friendly to the interests of white workers and most employers.
Finally, SDT research suggests that motivations to maintain hierarchy should play a significant role in how judges choose ideological justifications to support their interpretation of the law. Judges, especially appellate judges, support doctrinal rulings with legal principles and ideologies that are meant to give consistent guidance on how similar types of cases ought to be resolved. SDT suggests that these ideologies will tend to be hierarchy-enhancing rather than hierarchy-attenuating, particularly when the hierarchy is being threatened.
In the context of race discrimination law, the ideology of colorblindness, in particular, plays a significant role. Research relating SDO and colorblindness suggests a complex relationship between general support for hierarchy maintenance and endorsement of colorblindness, in part because the meaning of “colorblindness” is highly indeterminate. Colorblindness could be interpreted as a distributive justice commitment to more equal (i.e., colorblind) societal outcomes, in which case it would function as a HA ideology; but it could equally be interpreted as a procedural justice commitment to formally equal treatment, irrespective of its outcomes, in which case it has the tendency to perpetuate existing hierarchical distributions of social resources and would function as a HE ideology. Thus, the particular rationale underlying a commitment to colorblindness, the way in which colorblindness is framed, is important to its relationship with SDO.
What existing research suggests is that people who are higher in SDO endorse procedural understandings of colorblindness strategically to protect existing hierarchy, particularly when that racial hierarchy is under threat. For example, Knowles and colleagues found that (1) white Americans with high SDO levels were more likely to construe colorblindness as a procedural justice mandate when experiencing racial intergroup threat; (2) they were more likely to endorse colorblindness when feeling under threat; and (3) it was a desire for procedural justice that made high-SDO whites more likely to endorse colorblindness when feeling under threat. In other words, “rather than rejecting color blindness—an ideology widely accepted as a moral imperative—when the status quo is threatened, antiegalitarian White people construe it in a fashion that furthers their hierarchy-enhancing goals.”
All of these findings converge towards suggesting that, overall, judges should be more motivated to maintain a hierarchical status quo—influenced by perceptions of low existing social inequality, a zero-sum view of remediation, a negative view of discrimination claimants of color, and ideologies of procedural colorblindness. This Article argues that these dynamics manifest themselves in asymmetrical narrowing and proof asymmetry. Part III will show how this process is reflected in disparate treatment race discrimination law, how this has made existing doctrine fundamentally asymmetrical to the detriment of workers of color, and how doctrinal forks along the road were taken to arrive at this outcome.
5. Important Contributions of SDT.
Before I engage in this analysis, however, I do want to state clearly that in making my arguments in this Article, my point is not that SDT concepts can provide all the answers or fully explain the incredibly complex process of judging. I don’t believe there is a single theory, social psychological or otherwise, that can do so. Nor do I believe that SDT could reliably predict the decision-making of individual judges in individual cases with all of their complexity. Nevertheless, in trying to understand the complex web of reasons for how laws operate, and how they are interpreted and applied, SDT has a significant role to play.
In looking to SDT to help explain the contours and outcomes of employment discrimination law, the goal is not to displace other useful insights, but rather to “add an additional explanatory layer to [obtain] the deepest understanding of persistent inequalities among social groups.” There are clear benefits that come with mobilizing SDT in this effort.
For one, there is good reason to believe that SDT is particularly well-suited to inform our analysis of how employment discrimination law is interpreted. SDT foregrounds an important motivational dimension of how people behave when relations between salient social groups are at issue: they often act to a significant extent in accordance with general human tendencies to want to establish and maintain group-based hierarchical social systems. As noted above, research in the SDT tradition has provided powerful evidence that this psychological tendency exists as such; that it motivates the behavior of dominant group members to a greater extent; and that it manifests itself in support for a wide range of ideologies and behaviors that serve to maintain group-based hierarchy, particularly when the hierarchy is threatened. The interpretation of employment discrimination law, and antidiscrimination law more broadly, revolves in significant part around whether existing group-based hierarchy, and the position of different groups within it, is appropriate. This interpretation is done mostly by dominant group members; is based on ideologies and assumptions about the extent and appropriateness of existing hierarchy; and allegations of discrimination threaten the legitimacy of such hierarchy, particularly when raised by members of disadvantaged minority groups. Thus, SDT’s findings are highly relevant and directly tailored towards helping to explain judicial behavior in this space.
Moreover, the judicial interpretation that has led to the racial asymmetry in employment discrimination law that I analyze in Part III does not involve the kind of behavior for which cognitive explanations, including implicit bias, are most persuasive. Cognitive explanations such as implicit bias are particularly applicable to situations where the decision-maker “lacks . . . motivation, time, or cognitive capacity,” is “physically or mentally fatigued,” or has to make an “inherently ambiguous” decision. By contrast, the judicial decision-making that I analyze in this Article often involves considered choices about baseline assumptions that the doctrine should make regarding the continued persistence (or not) of discrimination, the relative need for antidiscrimination intervention, and the relative strength of the interests of different participants in employment discrimination cases (workers of color, employers, white workers). SDT researchers have made specific findings as to how the motivation to create and maintain group-based hierarchy influences such choices. Thus, SDT-based research is a particularly helpful tool when analyzing legal decision-making in hierarchy-relevant contexts like employment discrimination law.
Taking hierarchy-related motivational psychological processes seriously is even more important when race discrimination is at issue. As numerous scholars have illustrated, the history of race relations in the United States is one of creating, maintaining, managing, enforcing, and contesting racial hierarchy, including through law. If there is an area of American law in which the findings of SDT are especially likely to have explanatory power, it should be race discrimination law. Thus, since existing approaches, such as implicit bias research, have made important findings on the basic psychological processes that influence judges in particular circumstances, but leave a significant amount of variance in the highly complex area of discrimination-related behavior unexplained; and since we are often unable to directly investigate the psychological tendencies of the judiciary; looking to directly applicable social psychological findings from SDT is highly valuable in explaining the shape of legal doctrine particularly when, as I show in Part III, the doctrine aligns with predictions we can fairly make based on those findings.
Consider one brief example that I believe illustrates how an SDT-informed analysis of employment discrimination law can be complementary to, and go beyond, what has already been proposed. In a 2012 article, Professor Katie Eyer provided a thorough analysis of potential reasons for the poor odds of discrimination plaintiffs, including employment discrimination plaintiffs. Eyer argued that social psychology scholarship on attribution can help explain these poor odds, and in particular two phenomena: the well-documented tendency of people to hesitate to attribute a bad event to discrimination if they strongly believe in meritocracy; and the tendency to believe that discrimination (as generally understood) is relatively rare, making discrimination less cognitively accessible and thus leading people to be less likely to attribute a bad event to discrimination. An analysis grounded in SDT is complementary—and provides additional insights—to Professor Eyer’s helpful work. For one, such an analysis can provide greater specificity in certain contexts. As SDT research has shown, people’s willingness to believe that discrimination and hierarchy are relatively rare is not randomly distributed—rather, it varies by a person’s SDO level. This finding allows us to go beyond Eyer’s analysis, which shows that judges generally should be comparatively unwilling to make discrimination attributions. SDT suggests that particularly those judges who are likely to have high SDO levels, for example because they are from dominant groups, should be even more unwilling compared to judges who are likely to have lower SDO levels, for example because they are from nondominant backgrounds.
As a result, an SDT-informed analysis can more parsimoniously explain some of the findings that have been made regarding the role of a judge’s race in deciding employment discrimination cases. While the findings are not perfectly consistent, there are a number of studies that suggest that judges from racial minority backgrounds, particularly black judges, are significantly more likely to find in favor of employment discrimination plaintiffs than white judges. In other words, judges in these studies were not all equally unlikely to make attributions to discrimination. Rather, those most likely to have low levels of SDO, and thus hierarchy-attenuating views, were significantly more likely to make such attributions. Thus, consistent with SDT-based predictions, these judges were more likely to act in hierarchy-attenuating ways and rule in favor of plaintiffs that challenged employment discrimination, even in the face of a hostile doctrine. This analysis is not necessarily inconsistent with that of Professor Eyer, but it helps build on that analysis and provides even greater specificity.
As I describe in Part III, SDT-infused analysis provides useful insights in helping to explain important aspects of the content and underlying logic of employment discrimination doctrine as well. Thus, it should be considered a valuable analytical tool for lawmakers, judges, academics, and advocates alike. Before moving to my SDT-informed doctrinal analysis, however, Section II.B briefly introduces the concepts of “framework critique” and “baseline errors” that I will also mobilize in evaluating existing doctrine in Part III.
B. Framework Critique and Baseline Errors
Combined with analytical tools from SDT, this Article applies a so-called “framework critique,” a concept that has played an important part in the work of Critical Race Theory scholars. Such a critique recognizes that “the frameworks we use to understand and describe social facts are constructed in the context of particular political and social activities and projects,” and in the context of race and the law, frameworks of “equality” may serve to “reinforce racial privilege as part of a broader conservative project to limit a wide-ranging redistribution of resources along racial lines.” In other words, the frameworks within which we debate questions of racial equality, and the starting points from which we measure the existence of such equality, are ideologically contingent. In many cases, existing frameworks in the law “do conservative political work in the context of particular historical periods to protect white interests.”
Noah Zatz recently applied this kind of critique to Title VII in analyzing the frequent use of the often ill-defined term “special treatment” by courts and scholars when discussing discrimination cases, and in the process explained the important concept of “baseline errors.” Zatz argues that an accusation that a particular employer action, or perhaps a court order, that treats employees from different racial groups non-identically involves “special treatment” (an accusation that generally tries to suggest that the treatment is itself discriminatory) cannot be appropriately analyzed without first establishing the “relevant nondiscriminatory baseline.” Where race discrimination has affected the status quo distribution of employment benefits, this nondiscriminatory baseline should be “the world as it would have been without” the discrimination. In such situations, treating employees from different racial groups non-identically may well be remedial rather than discriminatory, i.e., “not merely nondiscriminatory, but antidiscriminatory: necessary to remedy discrimination” if the non-identical treatment helps to recreate the nondiscriminatory baseline from which discrimination caused a deviation.
To disregard the relevant nondiscriminatory baseline and assume instead that a discriminatory status quo, i.e., “the world created by [the] discrimination,” is the nondiscriminatory baseline is to commit a “baseline error” that confuses remediation with discrimination based on “decontextualization.” Acting on such a baseline error, in turn, not only enables the denigration of remedial action as inappropriate “special treatment,” but also tends to normalize inequalities that stem from discrimination:
When the benefits of discrimination against others are taken as a baseline entitlement, an intervention’s remedial character becomes invisible. Instead, that equalizing intervention looks like special treatment, raw redistribution away from members of a dominant group who earned their place at the top. This dynamic simultaneously shields discrimination’s beneficiaries from acknowledgement of their windfall and derogates discrimination’s victims as undeserving when they receive relief.
This Article argues that a significant portion of current Title VII disparate treatment doctrine proceeds from such a baseline error and consequently leads to asymmetrical outcomes that disfavor workers of color and entrench racial hierarchy in the workplace. This baseline error is to take prevailing—and historically persistent—racial inequalities that subordinate workers of color as normal and presume such inequalities to be the result of nondiscriminatory decisions. As a result, discrimination against such workers is perceived as the exception, rather than a regular feature of employer decision-making; claims of discrimination by workers of color are viewed with suspicion, and employer assertions of nondiscriminatory reasons for employment decisions are taken as more credible than employee allegations of discrimination. At the same time, this erroneous baseline is used to justify the view that white workers should be protected from “reverse discrimination” on facially identical (and perhaps more favorable) terms as workers of color, and that doing so is merely extending the “same” protective treatment to everyone. By contrast, affirmative action programs are construed as deviations from this baseline that require extraordinary justification because they involve highly problematic “special treatment” and unearned “preferences” for minorities.
I argue that the appropriate baseline would instead be to recognize that major racial inequalities persist in the American economy and that these inequalities almost uniformly disadvantage workers of color. This baseline would recognize that racial discrimination—even if conceived of by the basic test that asks whether the outcome for one and the same, say black worker, would have been different if she had been white—continues to be the reason underlying such inequalities in enough cases that it is appropriate to take claims of discrimination by minority plaintiffs at least as seriously as assertions by employers that they acted for nondiscriminatory reasons, and more seriously than claims of white plaintiffs that they suffered “reverse discrimination.”
Consider, for example, that notwithstanding Title VII’s now fifty-year tenure, persistent and large gaps continue to define America’s economy along racial lines. As of the late 2000s, black and Hispanic men on average earn less than three-quarters of the earnings of white men, and black and Hispanic women earn substantially less than white women. Black and Hispanic unemployment has for decades consistently been significantly higher than white unemployment. EEOC data shows that as late as 2015, at a national level, minority workers are strongly underrepresented compared to their share of the labor market in the job categories “Executive/Senior Level Officials & Managers,” “First/Mid Level Officials & Managers,” “Professionals,” and “Craft Workers” but strongly overrepresented in the categories of “Operatives,” “Laborers,” and “Service Workers”. These labor market numbers show a clear racial hierarchy that continuous to disproportionately distribute the most attractive jobs to white Americans and the most unattractive ones to workers of color.
While there are various reasons for such disparities, race discrimination continues to be a highly significant contributor. Audit studies consistently show that racial differences in employment outcomes do not result solely from differences in individuals’ interests, credentials, or potential, but that being a member of a racial minority group (or being perceived as such) significantly and negatively affects one’s access to employment benefits. Such findings have been made across different locations and levels of job selectivity, showing that racial discrimination hurts minority workers from the low-wage economy to jobs requiring college degrees. In certain labor markets, a white job applicant with a criminal record for possession of cocaine can still do “just as well, if not better” than Latino and black job applicants with no criminal background. Such racial disadvantage accumulates over a worker’s career, leaving employees of color more and more behind over time—particularly at higher levels of education. Indeed, a recent meta-analysis of more than thirty audit studies covering a forty-year period from the mid-1970s to the mid-2010s suggested that hiring discrimination against workers of color “neither declined over time nor varied by education level or type of occupation involved.” Overall, scholars have noted that across methods of investigation, “considerable scientific evidence of the persistence of workplace discrimination is evident in audit studies, grounded labor market analyses, social psychological research using experimental designs, and surveys of ordinary citizens.”
Part III shows how significant aspects of current disparate treatment doctrine do not take this baseline of a racially stratified American economy into account. As a result, we see a doctrine that leads to systematic asymmetrical treatment to the detriment of workers of color and helps maintain a system of racial hierarchy in employment. Research from SDT scholars gives us persuasive explanations as to why this is so.
III. Disparate Treatment Doctrine Asymmetrically Disfavors Workers of Color and Helps Maintain Racial Hierarchy
This Part provides three examples of how Title VII disparate treatment doctrine creates asymmetries in how it treats plaintiffs of color and in doing so helps maintain racial hierarchy in the workplace. Two of these examples come from what I call “traditional” disparate treatment cases, in which plaintiffs claim that an employer intentionally took an adverse employment action against them because of their race, while the employer claims that it took the action for nonracial reasons. I focus on cases proceeding under the so-called McDonnell Douglas burden-shifting regime. The third example compares the doctrine governing such traditional cases with Title VII affirmative action doctrine and argues that this comparison illuminates another asymmetry disfavoring workers of color. Throughout, I show how the framework from Part II helps to persuasively explain why we should expect such asymmetries to develop. First, however, I provide some background on relevant disparate treatment doctrine to set the stage for the critical analysis that follows.
A. Disparate Treatment Doctrine Background
The general idea underlying Title VII’s prohibition of disparate treatment in employment decisions is relatively easily stated. When an employee brings a disparate treatment claim, the ultimate fact that she has to prove is that the employer intentionally discriminated against her in a relevant employment decision because of a protected characteristic covered by the statute, including race. In other words, the general elements of a disparate treatment claim are (1) an adverse employment action; (2) discriminatory intent; and (3) causation, or a linkage between intent and employment action. Various difficulties arise, however, when this inquiry is applied in individual cases because (1) it calls for an evaluation of the decision-makers’ mental states—their motive(s) or intent—which is difficult, if not impossible, to observe, and thus prove, directly; (2) the employer usually controls most of the relevant evidence; and (3) human behavior, and thus employment decisions, are usually motivated by many different factors, leading to questions about how important the consideration of a protected characteristic must be before an employment decision counts as illegal disparate treatment. Trying to respond to these issues, disparate treatment law has developed a complicated array of legal tests and rules.
The most common way in which plaintiffs can attempt to prove unlawful disparate treatment is through one of the more formalized burden shifting mechanisms that courts have developed over time. The basic dividing line in this context is between the analysis established by McDonnell Douglas Corp. v. Green, also sometimes called the “pretext” regime, and the so-called mixed motives analysis, also sometimes called the “motivating factor” regime, that was initially put forth by the Supreme Court in Price Waterhouse v. Hopkins and subsequently codified in modified form by the 1991 Civil Rights Act.
Under McDonnell Douglas, the analysis proceeds in three steps: (1) a plaintiff must first establish a very specific type of prima facie case, which, if shown, creates a mandatory rebuttable presumption that discrimination occurred; (2) the employer may rebut this showing by introducing evidence of a legitimate, nondiscriminatory reason for the challenged employment action; and (3) the plaintiff may then try to establish that the reason offered by the employer was not the real reason for the action, but rather a pretext for illegal discrimination. A plaintiff must succeed in showing that her protected characteristic—for example, her race—was a but-for cause of the challenged employment action. If plaintiffs succeed in proving a case under this regime, they are eligible for the full slate of remedies available under Title VII, including compensatory and punitive damages.
Under a mixed motives analysis, a plaintiff must only establish that her race was at least one of the employer’s motivations for a particular employment decision—in other words, that it was a “motivating factor” for the decision even though there may have also been other, lawful motivations. Race need not have been the but-for cause of the decision. If a plaintiff makes this showing, the employer is liable for disparate treatment. But, if the employer can show that it would have made the same decision even absent the unlawful consideration of race, it can limit the plaintiff’s remedies to declaratory relief, certain injunctive relief, and attorney’s fees.
The way in which these two regimes relate to each other, and to which types of cases they apply, has been disputed for as long as the regimes have existed. This Article focuses its doctrinal analysis on the McDonnell Douglas regime. While the continued relevancy and vitality of the regime have been questioned by some, I make this choice for the following reasons. For one, “courts still rely heavily on the McDonnell Douglas framework in Title VII disparate treatment cases.” The Supreme Court, too, continues to use the framework in the disparate treatment context without any suggestion that it has lost its usefulness. Second, even though plaintiffs may bring claims under both frameworks, there are reasons to think that a significant number of plaintiffs will choose to pursue their cases only under a McDonnell Douglas approach to avoid allowing the employer to raise the same-decision defense and limit the plaintiff’s remedies. Lastly, the McDonnell Douglas framework is frequently used in reverse discrimination cases brought by white plaintiffs, and a modified version of the framework is used to evaluate challenges to affirmative action programs—both of which I analyze in more detail below. Accordingly, to provide a more coherent analysis, and recognizing that there is only so much that can be achieved in a single article, my focus will be on the McDonnell Douglas framework.
I argue that the logic and application of this burden-shifting regime creates doctrinal asymmetries that disadvantage workers of color throughout its various stages, asymmetries that can be explained when viewed through the lens of racial hierarchy and the framework established in Part II.
B. Asymmetry Examples
1. Example 1: The Prima Facie Case and “Background Circumstances.”
The first step for any plaintiff under the McDonnell Douglas regime is to establish a prima facie case of discrimination. The main purpose of this step is for the plaintiff to bring forward the type of evidence that, if left unrebutted by the employer, allows for the inference (indeed presumption) that illegal discrimination more likely than not explains the employer’s adverse action. For example, in McDonnell Douglas, which involved a discriminatory failure-to-hire claim, the Court determined that a prima facie showing could be made by establishing that the plaintiff (1) “belongs to a racial minority”; (2) “applied and was qualified” for the job at issue; (3) was rejected “despite his qualifications”; and (4) “the employer continued to seek applicants” with similar qualifications. The Court recognized, though, that different factual situations may call for different prima facie elements, and courts have indeed adjusted the elements for different types of claims. For example, in a case alleging discriminatory discharge, a court might require a plaintiff to show that she met the employer’s legitimate expectations for her job, was discharged nevertheless, and the employer sought a replacement. The goal is to “eliminate the most common nondiscriminatory reasons for” the allegedly discriminatory employment decision. After all, when such nondiscriminatory reasons are not applicable, and the employer provides no other reason, one can reasonably infer that, rather than acting in a “totally arbitrary manner,” the employer instead acted “more likely than not based on the consideration of impermissible factors.”
Constant across all formulations of the McDonnell Douglas prima facie case, however, is that membership in a “protected class” is always its first prong. In a superficial way, this makes sense; liability under Title VII requires an adverse employment action that was taken “because of” an employee’s race. Thus, an employee must be a member of a racial group to be discriminated against “because of” this membership. However, this prong becomes more curious if one considers that all racial groups are protected from race discrimination by Title VII, including white workers, as the Supreme Court has clearly held. If all racial groups are protected, every plaintiff can presumably automatically fulfill this first prong. Does the racial group membership requirement of the McDonnell Douglas prima facie case then do any substantive work in deciding whether it is reasonable to draw an inference of discrimination from a particular set of facts?
I argue that the answer depends on what one views as the relevant nondiscriminatory baseline and, relatedly, one’s view regarding the extent to which different racial groups have been, and continue to be, at risk of racial discrimination in the workplace. If all racial groups are at a substantial, or at least nontrivial, risk of race-based discrimination, then it makes sense to use racial group membership as such, together with other elements tailored to “eliminate the most common nondiscriminatory reasons for” a particular type of employment decision, and to conclude that an inference of discrimination is reasonable when all elements have been proved. Based on such a nondiscriminatory baseline, symmetrical treatment, i.e., asking plaintiffs of all racial groups to meet the exact same legal test, would also be equal treatment. Alternatively, one may think that members of some racial groups, particularly white workers, are at a much lower, perhaps negligible, risk of being discriminated against specifically based on their race. Proceeding from this baseline, one will not be willing to assume that the fact that a worker is white makes race-based discrimination the most likely explanation for a negative employment decision against that worker, even if the most common nondiscriminatory reasons for the decision have been eliminated via the remaining prongs of the prima facie case. Instead, one will require additional evidence suggesting that an employer took the unusual step (unusual at least in the American context) to discriminate against a white worker specifically because she was white. Indeed, not requiring such additional evidence would be “special treatment” of white plaintiffs by allowing them to benefit from a mandatory presumption of discrimination that is not justified by the evidence they have presented.
This is precisely the debate that has played out in the lower courts. More so than in some other contexts discussed below, some courts have recognized that there is a debate to be had over what the relevant analytical baseline should be, and what the corresponding implication is for Title VII disparate treatment doctrine. The result is a circuit split between courts that favor an identical legal standard for plaintiffs from all racial groups and those that recognize that equal treatment of all racial groups requires white plaintiffs to meet a different evidentiary standard.
On the one hand, various circuits have modified the first prong of the prima facie case for white plaintiffs and have required them to meet a so-called “background circumstances” test. This test requires a showing of “background circumstances [that] support the suspicion that the defendant is that unusual employer who discriminates against the majority.” Some of these courts specifically recognize that not to apply a modified standard for white plaintiffs would be to commit a baseline error—to assume that discrimination against white workers is as likely as discrimination against workers of color when the realities of the American economy simply do not support this assumption.
This would, in turn, unjustifiably provide greater (i.e., asymmetrical) legal protections to white plaintiffs. These courts recognize that the equal protection of workers of color and white workers requires taking into account an unequal baseline situation in the American economy: given consistent and pervasive racial inequality in employment favoring white workers, it is reasonable to infer discrimination (and require the employer to state its reason for the adverse action) from the combination of the remaining prongs of the McDonnell Douglas prima facie test and the fact that a plaintiff belongs to a racial minority. It is not reasonable, without more, to make the same inference for white plaintiffs.
However, not all courts apply the background circumstances test. While courts have relied on different reasons to reject the test, some circuits reject the test using reasoning that directly contradicts the above analysis. These courts insist that the first prong of McDonnell Douglas must be in all cases, including for white plaintiffs, the generic requirement of being a “member of a protected class.” For example, the Eleventh Circuit in Smith v. Lockheed-Martin Corp. required only a showing that the plaintiff “is a member of a protected class (here, Caucasian).” In a footnote, the court made clear that the circuit had “rejected a background circumstances requirement” and buttressed this conclusion with the statement that “[d]iscrimination is discrimination no matter what the race, color, religion, sex, or national origin of the victim.”
As noted above, courts taking this approach implicitly proceed from a baseline where white workers and workers of color face similar risks of race discrimination and the inference of discrimination from the remaining prongs of the prima facie case is equally strong for workers of all racial backgrounds. I argue that this commits a fundamental baseline error that erases the starting point of continuing racial inequality in the workplace. Through such “decontextualization,” this error doctrinally entrenches the “baseline entitlement” of white workers to a discriminatory status quo by giving them access to a mandatory inference of discrimination based on facts and assumptions that ignore their dominant position in the racial hierarchy and their favored status in the workplace. To put the point slightly differently, to the extent that courts reject the background circumstances test (or, as some circuits have done, begin to soften or question it), they instantiate doctrinal asymmetry that disfavors workers of color.
Findings from Social Dominance Theory connect this doctrinal debate to people’s predispositions towards maintaining existing racial hierarchy in the American workplace. This research shows that a person’s level of support for the maintenance of group-based hierarchy, i.e., a person’s SDO level, strongly influences related views regarding how much inequality currently exists in society—and thus one’s choice of the baseline from which to evaluate claims of discrimination. For example, Kteily and colleagues recently observed that when they asked study participants to rate how much power various racial groups had in American society, people with high SDO scores thought that there was comparatively less racial hierarchy in America than did low-SDO participants. This perception, in turn, made the high-SDO participants significantly less supportive of antidiscrimination policies. The authors concluded that “people come to perceive systematically different levels of inequality depending on how desirable they believe it [i.e., inequality] to be,” which in turn “reinforce[s] their convictions about the types of social policies they tend to favor.” In other words, if people are motivated to establish and maintain hierarchy, they perceive less inequality (including racial inequality) and thus believe that less egalitarian intervention to protect the disadvantaged (e.g., racial minorities) is required.
The flipside of this is, of course, that comparatively more egalitarian intervention is required on behalf of whites, who as participants in a flat hierarchy are perceived to be at an equal risk of racial discrimination. SDT research has made empirical findings in this regard as well. For example, Eibach and Keegan found that white Americans were significantly more likely to think about racial progress in zero-sum terms (i.e., progress for some means losses for others) than racial minorities. In turn, white participants with greater zero-sum thinking judged the level of progress made towards racial equality since the 1960s as the highest among all participants. SDO also predicted both sentiments: higher levels of SDO correlated with perceptions of greater racial progress since the 1960s, and high-SDO participants found it much easier to come up with examples of how “minority progress had harmed Whites” than ways in which racial progress helped whites. In other words, whites and individuals with high SDO appear to see the country further along in creating racial progressprogress which they view as entailing losses for them.
This view can easily translate into a belief that whites are victimized by racial progress, and that there is both less need for the law to intervene to protect minority groups and a greater need to intervene on behalf of whites. Research by Michael Norton and Samuel Sommers confirms that whites increasingly see themselves as victims of discrimination. Asking their participants to rate the level of discrimination that both white and black Americans have faced in each decade from the 1950s to the 2000s, they found that white Americans believed that by the 2000s, whites on average experienced greater levels of antiwhite bias than black Americans experienced antiblack bias.
These research findings are arguably reflected in court decisions that reject the background circumstances test. Judges faced with deciding whether to apply the test are making a doctrinal choice that is not predetermined. The Supreme Court has not spoken on this issue apart from noting that the elements of the prima facie case should require “evidence adequate to create an inference that an employment decision was based on an illegal discriminatory criterion.” Some judges reject the background circumstances test because they think that it inappropriately heightens the burden of proof for white plaintiffs and requires an adjusted approach towards claims of discrimination by whites when they suffer the same discrimination as racial minorities. This is a view of the American workplace in which existing inequality and hierarchy are low, and dominant group members, such as white males, are subject to significant amounts of discrimination. This view proceeds from a baseline error and, as research on SDT shows, is characteristic of attempts to support and maintain group hierarchy. Judges accomplish such hierarchy maintenance through proof asymmetry when they reject the background circumstances test and apply a legal test that makes it comparatively easier for hierarchy-enhancing individuals—white reverse-discrimination plaintiffs—to prove their case.
This doctrinal move to implement proof asymmetry can have downstream effects that directly and negatively affect the ability of workers of color to remedy existing hierarchy in the workplace. Consider Ricci v. DeStefano. In Ricci, a group of mostly white firefighters sued the City of New Haven when the City refused to certify the results of a firefighter promotional exam, which would have disproportionately excluded black firefighters from promotions, because it was worried about the design and fairness of the test and that certifying the results would expose the City to disparate impact liability. The Supreme Court overturned the City’s actions as illegal race discrimination. Importantly, the majority proceeded from “this premise: The City’s actions [in refusing to certify the results] would violate the disparate-treatment prohibition of Title VII absent some valid defense.” The Court cited no precedent nor much argument for this proposition. It simply asserted that because the City had taken its actions in light of the race-based disparity in test results, it had acted “because of race” and thus presumptively violated Title VII’s disparate treatment prohibition.
Yet, this position equates merely paying attention to race (with the intent to avoid unlawful disparate impact on racial minorities) with unlawful and intentional discrimination against whites, without requiring any proof that the City actually intended to discriminate against the higher-scoring firefighters because they were white—for example, by showing that the asserted motivation to avoid disparate impact liability was pretextual. In essence, the Court created a “super inference” of discrimination for the white firefighters based on the strong zero-sum view that avoiding harm to black firefighters necessarily meant race-based harm to whites. As we have seen, such a view strongly corresponds with the motivation to maintain racial hierarchy. In so doing, the Court created “racial asymmetry in the burdens and presumptions of Title VII law” that “skew[s] antidiscrimination law in favor of whites as a group.” After all, “[b]ecause all remedial measures on behalf of racial minorities can at some level be characterized as racially attentive, treating racial attentiveness—attending to the racial consequences of one’s actions—as a form of discriminatory motivation destabilizes virtually all remedial options, even those expressly authorized by settled doctrine and federal statutory law.” In other words, proof asymmetry in disparate treatment law that facilitates discrimination lawsuits by white plaintiffs becomes the tool to doctrinally operationalize the motivation to maintain racial hierarchy by enabling challenges to any act of racial remediation and making egalitarian intervention in favor of subordinate groups much less likely.
Rejecting the background circumstances test for white plaintiffs can be seen as creating similar doctrinal drift by doctrinally exaggerating the risk of race-based discrimination that white workers face. As Angela Onwuachi-Willig has pointed out, courts that have rejected the test have made it harder for courts to screen out, and more expensive for employers to defend, reverse-discrimination cases that clearly seem to lack merit. Thus, a doctrinal choice to reject the test supports the maintenance of existing racial hierarchy in the workplace by causing employers to second-guess employment decisions that allocate benefits to minority workers.
2. Example 2: Employee and Employer Burdens After the Prima Facie Case.
Another example of doctrinal asymmetry disfavoring workers of color can be found in how steps two and three of the McDonnell Douglas regime interact. As noted above, once a plaintiff establishes a prima facie case, the employer can rebut it by introducing evidence of a legitimate, nondiscriminatory reason for the challenged action. If the employer does so, the plaintiff can try to establish that the reason offered by the employer was not its real reason, but a pretext for discrimination.
The Supreme Court has made clear that the employer’s obligation to provide a nondiscriminatory reason is limited. It is only a burden of production, not of persuasion, and the employer need only introduce some evidence that supports its alleged reason, “not persuade the court that it was actually motivated by the proffered reasons.” By meeting this burden of production, the employer “destroys” the mandatory presumption of discrimination created by the prima facie case. Further, even if a plaintiff succeeds in disproving the employer’s asserted reason, they are not guaranteed to prevail. While a successful prima facie case, combined with proof showing that the employer’s alleged reason was not its real reason, can be sufficient for a plaintiff to win, it certainly will not always be. In other words, even if a plaintiff establishes a mandatory inference of discrimination through its prima facie case, and even if there is additional proof that the employer’s rebuttal reason is not credible (perhaps even contrived), the plaintiff only succeeds if they can affirmatively convince the factfinder that the employer’s dishonesty was a pretext for intentional discrimination. To put the point another way: The Supreme Court has rejected the notion that there should be a mandatory finding of discrimination when an employee provides sufficient evidence for a prima facie case together with sufficient evidence that the employer’s asserted nondiscriminatory reason is not credible.
In this context, the doctrinal asymmetry that disadvantages workers of color manifests itself differently than in Example 1. While the asymmetry in Example 1 stemmed from the relative treatment of different types of plaintiffs (white plaintiffs and plaintiffs of color) with respect to the same legal question (what must be shown as part of the prima facie case), in this Example the asymmetry is created by the relative treatment between plaintiffs and employer defendants. This relative treatment is, on the whole, asymmetrical to the detriment of workers of color because it systematically favors defendants in a cause of action that is prototypically and predominantly asserted by plaintiffs of color, rather than white plaintiffs. As with the acceptance or rejection of the background circumstances test, and as I show below, this doctrinal outcome was not predetermined by Title VII’s text or history, nor necessitated as a matter of doctrinal logic. Rather, it is the result of doctrinal choices by the Supreme Court that systematically resolved doctrinal ambiguity through hierarchy-enhancing rather than hierarchy-attenuating approaches and utilized asymmetrical narrowing and proof asymmetry to this end. In doing so, the Supreme Court acted to protect racial hierarchy in the workplace through Title VII doctrine.
Consider how today’s rules in this area came about. The Supreme Court’s first major disparate treatment case, McDonnell Douglas itself, ruled only that after an employer has articulated a legitimate reason for an adverse employment action, the plaintiff must be given “a fair opportunity to show that petitioner’s stated reason . . . was in fact pretext.” The Court used various unclear and perhaps conflicting formulations as to what the precise nature of the pretext inquiry should be. Thus, the Court left open the question whether a prima facie case, combined with a showing that the employer’s asserted reason was not credible, would be sufficient to legally mandate judgment in favor of the plaintiff. In Texas Department of Community Affairs v. Burdine, the Court further muddied the waters by stating that upon the employer’s introduction of its asserted reason, the plaintiff “must have the opportunity to demonstrate that the proffered reason was not the true reason for the employment decision,” yet also adding that at the pretext stage “[t]his burden now merges with the ultimate burden of persuading the court that she has been the victim of intentional discrimination.” Equivocating further, the Court noted that the plaintiff “may succeed in this either directly by persuading the court that a discriminatory reason more likely motivated the employer or indirectly by showing that the employer’s proffered explanation is unworthy of credence.” In U.S. Postal Service Board of Governors v. Aikens, the Court again approvingly cited these two methods of establishing pretext yet also stated that the relevant question at the pretext stage is about “discrimination vel non,” i.e., whether the plaintiff had shown that the “defendant [had] intentionally discriminated against the plaintiff.” Further complicating matters, Justices Blackmun and Brennan concurred to explicitly state their view that based on the above cases, a prima facie case combined with disproving the employer’s asserted reason mandated a finding for the plaintiff.
The Court finally resolved this question in St. Mary’s Honor Center v. Hicks, ruling that workers are not entitled, as a matter of law, to judgment in their favor upon proving a prima facie case and successfully discrediting the employer’s asserted reason(s)—though in some circumstances the evidence may still permit them to succeed upon such proof. The presumption of discrimination created by the prima facie case, which “simply drops out of the picture” upon the employer’s production of a nondiscriminatory reason, does not assert any further influence even if a plaintiff succeeds in subsequently discrediting the employer’s asserted explanation. Though not as hostile in tone, the Court confirmed and solidified this basic holding and set the doctrine as it stands today in Reeves v. Sanderson Plumbing Products, Inc.
Within the framework of this Article, one can analyze this doctrinal history as involving fundamental disputes about the proper baseline view of the prevalence of discrimination against workers of color, about the likelihood of employer complicity or innocence vis-à-vis such discrimination, and about how, and whether, the factual likelihood of discrimination is properly reflected in the prima facie requirements of the McDonnell Douglas regime. Hicks and Reeves resolved these disputes in favor of proceeding from a baseline that assumes that discrimination by employers is insufficiently frequent to justify mandatory judgment in favor of plaintiffs who pit a successful prima facie case against a discredited employer reason. In doing so, the Supreme Court made various doctrinal choices that implement asymmetrical narrowing and proof asymmetry by both weakening plaintiffs’ hands (by limiting the power of the prima facie case) and strengthening defendants’ hands (by extending extreme doctrinal deference to employers, even if they cannot credibly justify their decision-making).
With respect to the prima facie case, the Hicks majority relegated it to a mere “procedural device, designed only to establish an order of proof and production.” In other words, the Court implicitly rejected what it had said fifteen years earlier: that the prima facie case establishes a mandatory inference of discrimination because “we presume these acts, if otherwise unexplained, are more likely than not based on the consideration of impermissible factors.” Under the Hicks view, the mandatory inference of discrimination that a plaintiff earns when successfully proving the prongs of the prima facie case does not actually reflect a substantive judgment that, absent a credible response by the employer, discrimination has likely taken place. Instead, it serves the purely procedural “role of forcing the defendant to come forward with some response,” after which it “simply drops out of the picture.” Through this reasoning, the Court in essence codified an implicit conclusion that, notwithstanding the clear hierarchy that continues to describe the American workplace, enough racial progress has been made that if in, say, a hiring case, (1) a qualified black applicant applies for an open position, (2) is rejected, (3) the employer continues to look for other applicants of similar qualifications, and (4) the employer subsequently cannot come up with any credible reason why they did not offer the job to the qualified black applicant, one still need not assume that racial discrimination was the likely reason for the rejection of the black applicant.
As the dissent in Hicks pointed out, the Court could have just as easily made a different decision by changing its assumptions about the relevant nondiscriminatory baseline. After all, in cases in which the employer’s reason has been rebutted,
[t]he plaintiff has raised an inference of discrimination (though no longer a presumption) through proof of his prima facie case . . . . Such proof is merely strengthened by showing, through use of further evidence, that the employer’s articulated reasons are false, since “common experience” tells us that it is “more likely than not” that the employer who lies is simply trying to cover up the illegality alleged by the plaintiff.
The “common experience” referenced by Justice Souter would be that which recognizes that the realities of continued racial inequality, hierarchy, and discrimination in the American workplace indeed mandate the conclusion that if an employer cannot credibly explain why they refused an employment benefit to a worker of color when the most common nondiscriminatory reasons for such a refusal are absent, this decision was likely the result of race discrimination. The Court could have implemented such a rule of substantive law if it had wanted to.
The doctrinal moves that a majority of the Court made instead proceed from a baseline error about the (lack of a) continued prevalence of racial discrimination in the workplace. Such views about racial progress, the likelihood of discrimination, and the corresponding need (or lack thereof) to provide antidiscrimination protections to workers of color can themselves be seen as expressions of a desire to maintain racial hierarchy. The doctrine created by the court, making it more difficult for workers of color to prevail, is one way in which disparate treatment law has been shaped to contribute to maintaining this hierarchy.
With respect to the employer’s ability to prevail, the Court increased doctrinal asymmetry to the detriment of workers of color by further buttressing the significant deference to employers and their presumed good faith decision-making that already underlies the employer’s limited production burden. The Court put in place a default assumption that even if the employer’s asserted nondiscriminatory reason for its actions is not credible, something other than discrimination, such as cronyism or personal animosity between the employer and the plaintiff that is untainted by race, may well have been the reason for a particular employment decision. Recall the first vignette from the Introduction. In that case, both a district and an appellate court found it consistent with Title VII to reject a finding of race discrimination and credit the nondiscriminatory reason of personal favoritism even though the relevant employees had denied that favoritism played a role. This provides a clear illustration of how far the courts, led by the Supreme Court, have gone in rejecting a baseline view in which racial discrimination against workers of color is considered a strong possibility, persuasive data to the contrary notwithstanding.
While a full discussion is beyond the scope of this Example, it bears noting that courts have created various other sub-doctrines that further fortify the doctrinal asymmetry laid out above and make it very difficult for plaintiffs to prove that an employer’s action was the result of intentional race discrimination. Examples include the so-called “cronyism defense,” the “stray remarks doctrine,” the “same actor inference,”; the “honest belief” rule,; and the practice of deferring to “business judgment,” according to which courts have made clear that “courts must not second-guess an employer’s initial choice of appropriate qualifications . . . .”
These doctrinal moves show asymmetrical narrowing and proof asymmetry in action. They show asymmetrical narrowing because the doctrinal developments take a broad view of the rights of hierarchy-enhancing institutions—corporate employers—to make decisions to the detriment of workers of color as they see fit, even if they may not be able to credibly justify them. By contrast, they take a narrow view of the rights of hierarchy-attenuating individuals—workers of color seeking greater access to workplace benefits—to not have to suffer adverse employment actions under circumstances suggesting discrimination. They show proof asymmetry because the Court adopted doctrinal tests that make it comparatively difficult for HA individuals—i.e., racial minority claimants—to win their cases and achieve a HA result (greater protection for racial minorities in the workplace and compensation for adverse employment actions taken against them), and relatively easy for HE institutions—corporate employers—to win their cases, particularly when the outcome is a HE one (the justification of an adverse employment action against a worker of color).
3. Example 3: Affirmative Action Doctrine.
My third, and final, example of doctrinal asymmetry comes from the area of Title VII doctrine typically invoked to suggest asymmetry of the opposite kind, i.e., in favor of racial minorities: the rules regulating voluntary employer affirmative action programs. In cases deciding the legality of such programs, the tendency of courts (and disparate treatment doctrine) to be suspicious of claims of race discrimination and to defer to the wisdom of an employer’s decision-making (as we saw in Example 2) does not manifest itself. Indeed, the doctrine does quite the opposite. The framework and research discussed in this Article suggest that this is at least partially because the law, and many judges, work (consciously or unconsciously) to maintain existing racial hierarchy. In Example 2, employers were acting in a hierarchy-enhancing role, seeking outcomes that maintain existing racial hierarchy. Title VII doctrine is comparatively supportive of and deferential to such efforts. In this Example, employers who pursue affirmative action for workers of color are working in a hierarchy-attenuating role. As we will see, and as SDT research helps us predict, the law is much more hostile to such employers. The hierarchy-connected tendency to engage in baseline errors again plays a role, but so does the hierarchy-enhancing doctrinal ideology of procedural colorblindness that facilitates the framing of affirmative action as “racial preferences” and “reverse discrimination.”
Cases in which white plaintiffs challenge race-based affirmative action programs under Title VII are treated as a variant of disparate treatment cases, and are, in a broad sense, subject to the same three-step analysis under McDonnell Douglas as other disparate treatment cases discussed above. The specifics of what each party needs to show in affirmative action cases, however, are quite different. To fulfill their prima facie burden, affirmative action challengers only need to show that race “has been taken into account in an employer’s employment decision.” In other words, the doctrine departs from a starting point of strong suspicion towards affirmative action programs by making basically every decision that an employer takes in reliance on an affirmative action plan challengeable by a (white) plaintiff who alleges to have been disadvantaged by it. The employer can rebut this showing by pointing to the affirmative action program as the nondiscriminatory reason for its decision. The plaintiff then has the burden of establishing that the employer’s reason is a “pretext” by showing that the affirmative action program is invalid under prevailing judicial doctrine. For a voluntary affirmative action plan to be valid under Title VII, it generally must meet the following criteria: (1) the plan must have a sufficient factual predicate, which generally means that a plan must be responding to at least a “manifest imbalance” that reflects underrepresentation of racial minority members in “traditionally segregated job categories”; (2) the plan must not “unnecessarily trammel the rights” of majority employees; and (3) the plan must be designed to attain, but not maintain, greater racial balance in the position at issue.
This doctrinal infrastructure provides few incentives, and many risks, for employers willing to take hierarchy-attenuating steps by implementing an affirmative action program. For one, the doctrinal requirements can be difficult to meet, exposing the employer to significant risk of a lawsuit brought by a disgruntled white worker. For example, the “manifest imbalance” that is the predicate for a valid program must show a very significant underrepresentation of a particular group in the employer’s workforce compared with the relevant local labor force. Accordingly, employers who are uncertain whether the underrepresentation of workers of color in their workforce is significant enough to be “manifest,” or whether they used the right comparator population or statistical tools to measure the imbalance, may decide not to take the risk. Further, “manifest imbalance” is generally not enough. The imbalance must exist in a “traditionally segregated job category.” As some commentators have noted, this requirement can inappropriately “narrow the situations in which affirmative action may voluntarily be used,” particularly when a job requires special skills. This is because, in such a job category, the relevant comparator labor force may well have such a small proportion of minority members that it will be difficult or impossible for an employer to show a significant statistical distinction between that labor force and its own workforce. Courts generally have also not allowed employers to avoid the restrictive nature of these requirements by seeking to pursue more forward-looking goals than simply eliminating prior segregation and instead have required that plans must be broadly “remedial.” While details in this regard are beyond the scope of this Article, such courts have thus rejected what they considered non-remedial, forward-looking purposes such as achieving racial diversity in relevant jobs or ensuring that all racial groups receive a fair share of jobs in a developing industry.
In addition, even if meant to address a manifest imbalance in a traditionally segregated job category, a plan is not valid if it “unnecessarily trammels the rights” of majority employees. This could occur, for example, if it imposed rigid quotas, required the discharge of majority workers, created an absolute bar to their advancement, or inappropriately unsettled their legitimate expectations—for example, those stemming from a seniority system. Finally, the courts have made clear that affirmative action programs must be temporary, meant to attain, not maintain, greater racial balance in the workforce.
These requirements further exacerbate the minefield for employers attempting to implement a plan in accordance with the law. The difficulty of meeting them is illustrated by the case that underlies the second vignette in the Introduction. In that case, the court thought the rights of a white employee were illegally “trammeled” when a school district, which was in financial distress and needed to lay off employees, decided to modify its initial layoff plan, firing one less person in order to avoid having to fire its sole black administrator and suffer a setback to its long-term affirmative action goals. According to the court, the law did not allow this remedial action because once the district created an additional position, the expectations of the slightly more senior white employee to be retained trumped the district’s affirmative action goals. Further, the court framed the district’s efforts to keep its only black administrator as one that impermissibly tried to “maintain” racial balance.
Some courts have gone even further and ruled that when a white worker brings a race discrimination claim against an employer that has an affirmative action program, and there is at least a factual question as to whether the employer acted pursuant to the plan in making the decision challenged by the white employee, they will treat the existence of the affirmative action program as “direct evidence” of unlawful discrimination and require a determination of whether the plan is valid. Thus, having an affirmative action plan exposes employers to litigation by disgruntled white employees even for decisions where it is not clear that the plan was even a relevant consideration. Given this risk, an employer might well decide that its commitment to remedying societal race discrimination does not outweigh the legal risks attending the implementation of an affirmative action program.
Superficially, the courts’ hostile treatment of voluntary employer affirmative action is curious. The Supreme Court’s approval of affirmative action under Title VII is explicitly grounded in preserving employer freedoms, and there is reason to believe that there would have never been a majority on the Supreme Court to approve affirmative action if it had not been tied to protecting employer prerogatives. The framework and empirical research discussed in this Article, on the other hand, makes the doctrine look much less curious. It allows us to understand Title VII affirmative action doctrine as another example of asymmetrical narrowing and proof asymmetry that, far from introducing asymmetry favoring workers of color into an otherwise symmetrical Title VII framework, is part and parcel of a legal regime that works asymmetrically to the disadvantage of workers of color. This conclusion becomes clearer when one compares affirmative action doctrine with the doctrine governing “traditional” disparate treatment claims discussed in Examples 1 and 2.
“Traditional” disparate treatment cases are most frequently brought by workers of color and deal with the types of claims most often relevant to their employment situation—greater access to jobs and job security, career advancement through promotion, equal pay, and other employment opportunities and benefits that are still reserved predominantly for white Americans. In other words, in such cases, plaintiffs of color seek a hierarchy-attenuating outcome through an increased allocation of positive social value to subordinate groups. Employer defendants seek a hierarchy-enhancing outcome by stifling this allocation and preserving the existing hierarchy. The doctrine that regulates such cases is very defendant-friendly and, unsurprisingly, plaintiffs of color overwhelmingly lose such cases. They are assumed by some courts to be no more at risk of discrimination than white workers, and this erasure of the continued racial hierarchy in the American workplace can make courts hostile to employer actions taken in their interests. Even if workers of color establish a mandatory presumption of discrimination in their favor and succeed in discrediting the employer’s rebuttal, courts are still suspicious of their claims of discrimination.
By contrast, courts grant significant deference to employers when their actions preserve existing racial hierarchy. Courts ask little more of employers than to state a reason for such actions that is not racial animus simpliciter—it need not even be their real reason, nor is there necessarily a negative consequence if the asserted reason is not credible. Further, through various sub-doctrines the courts have protected employer “cronyism” that hurts racial minorities, diminished the probative value of racial slurs and biased statements, credited employers’ “honest beliefs” even if the employer makes mistakes that disadvantage racial minorities, refused to second-guess employer “business judgment,” and more.
Yet such deference falls away when the same employers take hierarchy-attenuating steps and try to allocate positive social value to workers of color through affirmative action programs. Courts are wary of that kind of “business judgment,” even if it is made voluntarily. Even though ostensibly subject to the same burden-shifting regime that applies in “traditional” cases, white plaintiffs seeking hierarchy-enhancing outcomes in invalidating affirmative action programs face a much more plaintiff-friendly doctrine. They can attack an affirmative action program because the existing allocation of opportunity is perhaps unequal along racial lines, but not “manifestly imbalanced” enough, or not in the right job category, to justify intervention. They can attack the decision-making process by which an employer determined the existence of the “manifest imbalance” it needs to justify voluntary affirmative action. They have a right to demand that their interests are not “unnecessarily trammeled,” and that their expectations for employment advantage are preserved. And even if an employer clears all of these hurdles with a narrowly designed affirmative action program responding to clear inequality, the employer can only use such a program to make steps in the right direction, but not to preserve its gains. Employers, meanwhile, face the risk that their affirmative action programs are treated as direct evidence of their unlawful discriminatory intent.
Through such asymmetrical narrowing and proof asymmetry, disparate treatment doctrine plays its part in maintaining America’s racial hierarchy. And it does so in a highly efficient way: On the one hand, the highly restrictive doctrine governing affirmative action programs ensures that employers think twice about whether implementing a program is worth the inevitable lawsuit—particularly those employers who may not have egregious imbalances in their workforce or a long history of internal discrimination, but who may still be committed to contributing to the alleviation of racial hierarchy in the American economy. Under current affirmative action doctrine, there is little space for such forward-looking programs, and employers may well decide that adopting them is not worth the potential cost. White plaintiffs thus need not file many anti-affirmative action cases to preserve the doctrine’s deterrent effect.
On the other hand, the deferential doctrine that applies in “traditional” disparate treatment cases provides little incentive for employers and managers to change their thinking or day-to-day approach to their business even when it clearly affects workers of color negatively. The burden of making inroads into the continuing racial hierarchy of the American workplace falls on discrimination plaintiffs, predominantly from racial minority groups—but they almost inevitably lose their disparate treatment cases.
Empirical research from SDT researchers is again helpful in explaining this asymmetry. Perhaps the most common objection to affirmative action programs is that they involve “preferential treatment” for racial minorities, and “unearned” preferences at that. Individuals who would not have gotten a job, promotion, or training now do get those benefits (and, by implication, whites, who would have otherwise gotten those benefits, now no longer do), and this is only because they are racial minorities who are “preferred” under an affirmative action program. Providing such preferences is unfair, the argument goes, because they represent “racial spoils” that are violative of the principle that benefits should be distributed on the basis of “merit.” Accordingly, the legal system must pay careful attention to such “preferences,” to ensure that they do not, or only to the smallest extent possible, disrupt the expectations that those who work hard and “play by the rules” of meritocracy have earned for themselves.
Of course, any objection to “preferences” must be based on an implicit starting point, a nondiscriminatory baseline, from which departures can be characterized as “preferential.” The objection to affirmative action as entailing unfair racial preferences relies on a nondiscriminatory baseline at which little hierarchy remains in the American workplace, white workers and workers of color are subject to similar risks of discrimination, and thus “taking sides” between them is arbitrary and discriminatory. As described in more detail above, findings from SDT research show us that taking this baseline view is strongly related to greater motivation to maintain racial hierarchy, which we should expect to see in large parts of the federal judiciary.
SDT researchers have made even more specific findings that connect the differential treatment of white plaintiffs in the affirmative action context and plaintiffs of color in the “traditional” disparate treatment context to human tendencies to create and maintain social, and racial, hierarchy. Consider a series of studies by Unzueta and colleagues. In an initial study, white participants read vignettes about a [black/white] police officer who had been passed over for a promotion in favor of a [white/black] officer and either attributed his failure to be promoted to racial discrimination or not. The participants rated the “likability” of the officer, indicating whether they thought the officer was likable, friendly, had a good personality, and so on. The authors found that the higher the participants’ SDO levels, the less they liked the black officer claiming discrimination.
In a follow-up study with a multiracial pool of participants, the authors investigated similar effects of SDO on hirability determinations. Asked to imagine themselves as part of an HR team, participants read about a [white/black] applicant for employment who had left his previous employer—in one condition after filing a formal discrimination claim because all promotions had been given to [black/white] co-employees, and in the control condition without any mention of his reasons for leaving. The participants then again rated the applicant’s likability and also rated the applicant’s “hirability” by indicating how likely they would be to interview and hire the applicant, and how likely it was overall that the applicant would be hired. Once again, participants with higher SDO levels liked the black discrimination claimant significantly less, and they also deemed the black applicant who had previously filed a discrimination claim significantly less hirable than participants with lower SDO levels.
To sum up, high SDO participants did not simply consider all potential workers who had previously made a discrimination claim to be less likable and hirable. They could have viewed all such applicants as potential troublemakers. Instead, they were racially specific—it was black, but not white, discrimination claimants that suffered these negative effects. The authors suggest that this may well be because “discrimination claims made by members of racial minority groups potentially question the legitimacy of the racial hierarchy, [and thus] perceivers relatively high in antiegalitarian sentiment react in a particularly negative manner toward such claimants.”
Disparate treatment doctrine reflects such negative reactions to workers of color. For instance, recall Example 2 above. Courts are not willing to hold that a successful prima facie showing, combined with discrediting the employer’s rebuttal reason, should mandate a finding of discrimination in part because regardless of the credibility of the reason an employer actually articulated, it is still possible that personal animosity or another, nonracial, nondiscriminatory reason “really” explained why the worker of color lost out. This strongly reflects the findings by Unzueta and colleagues: higher SDO judges (of which there are likely numerous) are likely to view discrimination claimants of color as less likeable, and thus should be more likely to assume that a reason such as personal animosity, not race discrimination, caused the worker to lose out on an employment opportunity. This was precisely what happened in Hicks itself. In that case, a black plaintiff claimed that his discharge was racially motivated. Even though the district court determined that the defendant’s asserted reasons for the discharge—disciplinary violations—were not credible, the court still ruled against the plaintiff. The court did so because it thought that the plaintiff had not sufficiently shown that racial, rather than purely personal, animosity had motivated the discharge—even though the employer’s agents involved in the discharge explicitly disclaimed such personal animosity. Thus, the plaintiff had not met his burden of persuasion on discriminatory intent. The Supreme Court validated this doctrinal interpretation.
Similarly, higher SDO judges are likely to view workers of color as less hirable, and thus may be more likely to assume that a lack of qualifications or poor performance, not race discrimination, caused the worker to lose out. As described above, much of “traditional” disparate treatment doctrine is built around precisely such kinds of judgments and thus reflects efforts to protect existing racial hierarchy. On the flipside, in the affirmative action context, the doctrine explicitly reflects the higher likeability of white discrimination claimants seeking to maintain existing racial hierarchy. The doctrine is explicitly concerned with their plight, making sure that only very significant racial imbalances can justify disturbing their status quo entitlements, and explicitly demanding that their interests be respected and not “unnecessarily trammeled.” If a greater motivation to maintain existing hierarchy makes judges view white discrimination claimants as comparatively more likeable and hirable, those judges should, in turn, be more likely to find that an affirmative action program illegitimately disadvantages such “meritorious” workers by giving to workers of color, less likable and hirable perhaps, illegitimate and unearned preferences.
Lastly, SDT research can help explain an important ideological dimension of the asymmetry unveiled by comparing “traditional” disparate treatment doctrine and affirmative action doctrine. Another major line of criticism of affirmative action programs is that they involve racial discrimination (just the reverse of the “traditional” kind) because they require that race is considered when deciding who should receive employment benefits. Doing so involves treating people differently on the basis of their race—and that is the textbook definition of “racial discrimination.” Unless there is a very good reason for it, we should not allow racial discrimination of any kind, including the “reverse” kind, because racial discrimination does not become any less problematic just because it may seem advantageous for a decision-maker to engage in “benign” discrimination at a particular point in time. Justices Clarence Thomas and Antonin Scalia have been some of the most outspoken proponents of this view over the last three decades and have deeply grounded their arguments in appeals to “colorblindness,” particularly in the procedural form that considers race to be an inappropriate criterion for any decision that distributes social goods.
Of course, as both CRT and SDT research show, appeals to procedural colorblindness are not neutral with respect to their relationship to the maintenance of racial hierarchy. Recall the research of Knowles and colleagues, which found that higher SDO levels are related to construal of, and greater support for, colorblindness as a procedural justice mandate under conditions of racial intergroup threat, including threat induced by minorities pushing for affirmative action. The affirmative action doctrine described above and its characterization as a narrowly limited exception to an otherwise applicable strong norm of formal equal treatment reflect a clear procedural justice understanding of the ideological underpinning of disparate treatment doctrine. This underpinning can, and should, be understood as a hierarchy-enhancing legitimizing ideology that is being mobilized in support of maintaining existing racial hierarchy in the workplace in response to the threat to the hierarchy caused by employers implementing affirmative action programs that benefit workers of color.
IV. Thinking About Solutions
Building on the descriptive argument that existing disparate treatment law features significant doctrinal asymmetries that disfavor workers of color and favor the maintenance of racial hierarchy in the workplace, this Part begins an argument—more detailed development of which will have to occupy much more space in future work—which suggests that “fixing” disparate treatment law might have to take a different, or at least a modified, approach to what has been proposed in some parts of the employment discrimination literature.
As noted briefly in the Introduction, there is a broad consensus that disparate treatment law is in disarray and is overly harsh to plaintiffs, and thus various proposals for doctrinal reform have been made. Rather than focusing on reducing asymmetry that disfavors workers of color, however, such proposals have often centered around reducing the complexity of existing doctrine, and the evidentiary burdens that it places on plaintiffs. While scholars have made strong cases in support of such proposals, I suggest that in the particular context of race discrimination we ought to be cautious of unintended consequences that could flow from strong implementation of such proposals—consequences that may further support the continued maintenance of racial hierarchy in the American workplace. Instead, we ought to focus on reforms that directly help to reduce doctrinal asymmetry and support hierarchy attenuation.
Some proponents of the view that Title VII race discrimination doctrine is (contrary to the claims made in this Article) already largely symmetrical have argued that doctrinal reform should make it even more symmetrical. This would involve, for example, eliminating the “background circumstances” test analyzed in Section III.B.1 as inappropriately requiring white plaintiffs and plaintiffs of color to meet different doctrinal standards. As I have argued, however, viewed against the backdrop of the asserted purpose of Title VII to eliminate racial hierarchy in the American economy, this test is necessary to provide equal doctrinal treatment to the claims of workers of color. Equally important, the test is necessary to avoid the doctrinal ossification of a baseline error that inappropriately eases the burden of establishing a disparate treatment claim for white reverse discrimination plaintiffs who challenge racial remediation—as happened in Ricci, for example. Such ossification increases the risks and costs for, and thus reduces the likelihood of, employers engaging in such remediation when it is profoundly needed. Instead, a robust background circumstances test should be applied to white reverse discrimination plaintiffs. This would have the added benefit of aiding a doctrinal push for a mandatory finding of liability when a plaintiff proves a prima facie case and discredits the employer’s asserted reason(s) for a challenged employment action. With a robust background circumstances requirement, both workers of color and white workers would be putting forth evidence in their prima facie cases that justifies an inference of discrimination, which in turn is further strengthened by having discredited the employer’s rebuttal. This would provide stronger support for a substantive, rather than merely a procedural, reading of the prima facie case as suggested in Section III.B.2. Thus, a universal and robust background circumstances requirement for white reverse discrimination plaintiffs would be a useful doctrinal reform.
Other scholars have proposed assisting employment discrimination plaintiffs by reducing the causation burden that (all) disparate treatment plaintiffs have to meet. Such proposals generally call for a legal standard which requires plaintiff to prove something less than but-for causation to establish liability, such as the “motivating factor” showing that is required in Title VII “mixed motives” cases. Reducing the plaintiff’s evidentiary burden would make it easier for more plaintiffs to succeed, especially if their cases are currently marginal. More specifically, a lower causation burden would allow plaintiffs to succeed in “overdetermined” cases in which there are multiple sufficient reasons for a particular employment action (say the plaintiff’s race and his disciplinary history) but none of the reasons can be viewed as a necessary “but for” cause. Such proposals, too, however, may have significant unintended consequences that could hinder the elimination of racial hierarchy in the workplace.
This time, the problem is one of ideology. The argument that “motivating factor” causation should always be sufficient to establish liability is based on the notion that any consideration of a protected characteristic in an employment decision is inappropriate. As the Supreme Court noted in Price Waterhouse, for example, under this view of causation the protected characteristic “must be irrelevant to employment decisions” and thus “may not be considered in making decisions” at all. In the context of race, of course, this standard amounts to an extreme commitment to procedural colorblindness. As I discussed in Section III.B.3 above, however, this strict procedural view of colorblindness has been the main ideological weapon wielded against voluntary employer affirmative action programs. Indeed, it is what drives the view that affirmative action ought to be treated and policed as a very limited departure in favor of workers of color from an otherwise symmetrical Title VII. For example, this view underlies the decisions finding that an affirmative action plan counts as “direct evidence” of unlawful intentional disparate treatment unless the plan fits within the strict confines of existing doctrine. As discussed in Section III.B.2, it also underlies the “premise” in Ricci that the actions of the City of New Haven in trying to avoid disparate impact against workers of color were presumptively unlawful disparate treatment because race was considered in the decision not to certify the test results.
Accordingly, implementing a universal “motivating factor” causation standard would double-down on procedural colorblindness, magnify the perceived theoretical inconsistency between a “traditional” disparate treatment doctrine that abhors consideration of race and an affirmative action doctrine that requires it to be effective, and put even greater pressure on those asked to justify affirmative action programs viewed as a “special exception.” The plurality in Price Waterhouse avoided having to answer this conundrum by simply carving affirmative action cases out of the reach of the principles announced in the opinion. However, it was promptly called out by Justice O’Connor’s concurrence and the dissent, both of which argued that “motivating factor” causation should be available to all plaintiffs, including those challenging affirmative action programs. Of course, if a motivating factor standard were available to affirmative action challengers, there could be no affirmative action programs. The whole point of an affirmative action program is to take the protected characteristic, including race, into account. Thus, by definition, each decision taken pursuant to an affirmative action program would involve “motivating factor” causation and liability.
This is not to say that there is no place for a lower causation standard in disparate treatment doctrine. Given longstanding racialized inequality in the American workplace, we should be suspicious of employment decisions that even partially rely on race to the disadvantage of workers of color. But in pushing for lower evidentiary burdens for plaintiffs, we should avoid framing the benefits of a lower causation burden in strong colorblind terms, and instead emphasize how procedural colorblindness is the ideological weapon to which those with an interest in maintaining racial hierarchy turn when justifying their policy preferences. The lower causation burden should be used only in “traditional” disparate treatment cases, and only in the service of the statutory purpose of eliminating racial hierarchy in the workplace. To protect against baseline errors, white plaintiffs should have to make a relevant background-circumstances showing before having access to the standard. Affirmative action programs should not be challengeable using motivating factor causation but instead be made easier to implement. While there are complex factors to be considered in this context, including those involving the dynamics of SDO vis-à-vis different formulations of affirmative action programs that require more detailed analysis, and of constitutional law, the framework laid out in this Article should be helpful in proposing useful answers in that regard as well.
Finally, some commentators have urged a move away from proof structures towards evaluating a single determinative question: based on all of the evidence, did the plaintiff succeed in showing that it was more likely than not that the employer intentionally discriminated against her because of a protected characteristic. Here, too, while there is something to be said for removing complexity from the legal inquiry (which may well be a stumbling block for deserving plaintiffs), there is equal reason to believe that such a simplified inquiry could be just as hard to meet for workers of color.
For one, such a standard creates much greater discretion for the finder of fact. As some of the implicit bias literature has shown, decisions with greater amounts of discretion are notoriously prone to be biased against people of color. Research grounded in SDT, too, suggests that judges, many of whom we can suspect to have relatively high levels of SDO, could be influenced in more discretionary decision-making by skewed perceptions of existing racial inequality and racial progress, or differential liking of white discrimination claimants versus those of color. Thus, this Article suggests that if courts were to accept this reform proposal, their analysis should be firmly conscious of the possibility of baseline errors and the hierarchy-relevant outcome of any particular decision. Proof structures can be complicated, but if properly calibrated based on a clear underlying purpose, they may also constrain factfinders and judges who may otherwise be hostile to discrimination claims and inclined to rule against them. A McDonnell Douglas analysis grounded in the statutory purpose to eliminate racial hierarchy in the workplace—with a robust background circumstances requirement, a mandatory finding of liability upon a prima facie showing combined with discrediting the employer’s rebuttal, and a rejection or significant softening of the various doctrinal rules at the pretext stage that derail most disparate treatment cases—could serve such a function.
Ultimately, and in the longer term, perhaps the most effective and lasting way to address the many hierarchy-enhancing aspects of current disparate treatment doctrine would be to push for a judiciary that is more explicitly motivated to attenuate existing racial hierarchy. As SDT research has shown, while the study of law may increase people’s SDO levels, being in an environment with hierarchy-attenuating mores can decrease people’s SDO levels, and people choosing to work in those environments are likely to have low levels of support for group-based hierarchy to begin with. Public defenders, in particular, have been found to have very low SDO levels. There have been calls to diversify the bench with more judges from what the framework of this Article would likely consider to be hierarchy-attenuating positions, such as public interest lawyers, public defenders, and certain lawyers representing individual clients.
The framework and analysis provided in this Article suggest that such efforts may indeed be productive steps that could help address the current hostility of employment discrimination law towards workers of color. Such an approach would be particularly useful when combined with efforts to increase racial and gender diversity on the bench, calls for which have also been made by different groups. SDT research has consistently shown that women and members of racial minorities have lower average levels of SDO than white men. Thus, a judiciary that combines racial and gender diversity with experiential diversity in the practice of law would likely be poised to act in ways, including crafting judicial doctrines, that are more sensitive to the continued subordination of workers of color in the American economy. Such a judiciary, on the whole, might be able to turn the tide and bring us closer to the goal of helping to eliminate racial hierarchy through law than what we see today.
As I have briefly noted earlier, this is not an abstract prediction. Judge characteristics, particularly race and gender, have been found to influence the way in which judges decide employment discrimination cases such that judges of color, and particularly black judges, are much more likely to find in favor of plaintiffs bringing race (and potentially also sex) discrimination claims, even under the cramped doctrines that currently exist. Female judges, too, have been found to be more likely to find for plaintiffs, at least in the sex discrimination context. Importantly for purposes of this Article, existing data suggests that female judges and judges of color can influence appellate panels towards more hierarchy-attenuating outcomes in certain circumstances, even if they are in the minority on a panel. This is particularly important in an area like federal employment discrimination law, where statutory law is broad and vague, and much of the doctrine results from judicial interpretation at the circuit court level. Thus, while the data on judicial diversity and decision-making in employment discrimination cases is neither perfect nor unambiguous, it supports the proposition that a framework like that laid out in this Article, grounded in robust social psychological research by SDT scholars, can help us both better understand and explain judicial decision-making and evaluate the implications of various proposals for change. It is particularly appropriate to analyze federal employment discrimination law, and particularly race discrimination law under Title VII, through a hierarchy-centered lens given the Supreme Court’s pronouncements about the purposes underlying this area of the law. This Article makes a start in this direction, hopefully with much more analysis in this vein to come in the future.
This Article has looked to Social Dominance Theory—a prominent theory in social psychology with a robust body of supporting empirical research—to question the view that Title VII race discrimination doctrine is symmetrical, protecting all racial groups equally except for those instances, most notably affirmative action, that create limited preferences in favor of workers of color. Digging deeper into the doctrinal logic, armed with research findings on how judges are likely to act in hierarchy-relevant cases, this Article shows how current Title VII disparate treatment doctrine is fundamentally asymmetrical to the detriment of workers of color and maintains the racial hierarchy that continues to pervade the American economy. Taking a hierarchy-centered view helps us uncover unintended negative consequences that may result from various law reform proposals that have been suggested to address the harsh nature of current law. This Article has provided initial suggestions for what reform grounded in one of Title VII’s main asserted purposes—to eliminate racial hierarchy in the workplace—could look like. There is much more to be said and analyzed in this context. But if we are to move towards a more racially egalitarian society, we must be conscious of, and challenge, the human tendency to perpetuate group-based hierarchy at every step along the way.
Jim Sidanius & Felicia Pratto, Social Dominance: An Intergroup Theory of Social Hierarchy and Oppression 309 (1999).
The following description is taken from the facts of Foster v. Dalton, 71 F.3d 52 (1st Cir. 1995).
See Foster, 71 F.3d at 56; see also Neal v. Roche, 349 F.3d 1246, 1251 (10th Cir. 2003) (noting case law holding “that an employer’s actions based on loyalty to a friend or relative (particularly an unemployed friend or relative) are not considered ‘discriminatory,’ even where they benefit the nonprotected friend or relative at the expense of a more qualified, protected person”).
The following description is taken from the facts of Cunico v. Pueblo Sch. Dist. No. 60, 917 F.2d 431 (10th Cir. 1990).
Id. at 440 (quoting Johnson v. Transp. Agency, 480 U.S. 616, 617 (1987)).
See, e.g., 1 Barbara T. Lindemann et al., Employment Discrimination Law 1‑1 n.1 (5th ed. 2012) (noting that data from the Administrative Office of the U.S. Courts in the year 2011 shows that “[t]he only statutory category of civil litigation that produces a greater volume of cases is habeas corpus petitions”).
See generally Kevin M. Clermont & Stewart J. Schwab, Employment Discrimination Plaintiffs in Federal Court: From Bad to Worse?, 3 Harv. L. & Pol’y Rev. 103 (2009); see also Michael Selmi, Why Are Employment Discrimination Cases So Hard to Win?, 61 La. L. Rev. 555, 558–61 (2001).
See, e.g., Clermont & Schwab, supra note 7, at 127 (noting that “[t]he most significant observation about the district courts’ adjudication of employment discrimination cases is the long-run lack of success for these plaintiffs relative to other plaintiffs”).
See, e.g., id. at 111–12 (noting that the “spread between defendants’ and plaintiffs’ reversal rates in jobs cases is more extreme than the spread in non-jobs cases, with jobs defendants doing better and jobs plaintiffs doing worse than their non-jobs counterparts”).
See, e.g., Katie R. Eyer, That’s Not Discrimination: American Beliefs and the Limits of Anti-Discrimination Law, 96 Minn. L. Rev. 1275, 1287 (2012) (noting that “discrimination litigants have encountered difficult odds since at least the late 1970s (when comprehensive data [was] first available)”).
See, e.g., Ellen Berrey et al., Rights on Trial: How Workplace Discrimination Law Perpetuates Inequality 54–56 (2017) (discussing how, in a randomly selected sample of over 1,700 employment discrimination cases, race discrimination cases were the most commonly filed type, accounting for forty percent of all cases).
This historical lack of success has led at least one scholar to claim that federal courts have an “anti-race plaintiff ideology.” See Wendy Parker, Lessons in Losing: Race Discrimination in Employment, 81 Notre Dame L. Rev. 889, 931–33 (2006); see also Selmi, supra note 7, at 562 (noting that “[r]ace discrimination claims are generally thought to be the most difficult employment claim to succeed on”). The empirical data on this point is somewhat mixed, however, and may depend on the sample of cases studied. For example, in their analysis of a sample of over 1,000 employment discrimination cases filed between the late 1980s and early 2000s, Laura Beth Nielsen and various colleagues found that sex discrimination cases are comparatively more likely to be dismissed at the pleading stage, see Berrey et al., supra note 11, at 66–68, and to be lost on summary judgment, see Jill D. Weinberg & Laura Beth Nielsen, Examining Empathy: Discrimination, Experience, and Judicial Decisionmaking, 85 S. Cal. L. Rev. 313, 340–42 (2012). Relative comparisons aside, however, both race and sex discrimination cases are overwhelmingly unsuccessful.
See, e.g., Laura Beth Nielsen et al., Contesting Workplace Discrimination in Court 21, 48 (2008), http://www.americanbarfoundation.org/uploads/cms/documents/nielsen_abf_edl_report_08_final.pdf [https://perma.cc/LD3K-355C] (noting that in a randomly selected sample of employment discrimination cases filed in seven federal judicial districts between 1988 and 2003, eighty percent of race discrimination claims were brought by African-American plaintiffs, and only eight percent by white plaintiffs); Parker, supra note 12, at 897–98, 906 (noting that in a national sample of Title VII disparate treatment cases published in 2003, about sixty percent were brought by African-American plaintiffs, and only ten percent by white plaintiffs).
See Clermont & Schwab, supra note 7, at 113 (noting the possibility that “district courts process employment discrimination cases with a neutral or even jaundiced eye toward plaintiffs” and that “unconscious biases may be at work at the appellate level”).
According to at least one scholar, various other explanations that have been proposed, including that there has been a change in societal biases from open prejudice to the subtle and structural bias that makes discrimination harder to prove, that comparatively many employment discrimination claims lack merit, and that the best cases settle rather than being resolved in court, cannot fully explain the terrible record of success of employment discrimination plaintiffs. See Eyer, supra note 10, at 1286–91.
See infra Section II.B (discussing illustrative data and research studies in more detail). Such continued discrimination also helps explain why the vast majority of employment discrimination plaintiffs claiming race discrimination continue to be workers of color, as noted above.
See, e.g., Martin J. Katz, Unifying Disparate Treatment (Really), 59 Hastings L.J. 643, 645 n.8 (2008) (“Courts and commentators have routinely referred to current disparate treatment doctrine as a ‘swamp,’ a ‘morass,’ and a ‘quagmire.’” (citations omitted)). Other scholars have, even more pessimistically, concluded that this area of the law is fundamentally broken. See, e.g., William R. Corbett, Fixing Employment Discrimination Law, 62 SMU L. Rev. 81, 116 (2009) (arguing that employment discrimination law “is now so bad that judges do not know how to analyze motions for summary judgment or properly instruct juries”).
In particular, the Civil Rights Act of 1991 not only expanded the remedies available under Title VII to include damages but specifically responded to restrictive decisions of the Supreme Court that had narrowed the reach of the statute. See Civil Rights Act of 1991, Pub. L. 102–166, §§ 3, 102, 105 Stat. 1071, 1071–72 (codified as amended at 42 U.S.C. § 1981) (noting one purpose of the legislation was “to respond to recent decisions of the Supreme Court by expanding the scope of relevant civil rights statutes,” which included Title VII).
See generally, e.g., Katz, supra note 17, at 655–58; Corbett, supra note 17; Victor D. Quintanilla & Cheryl R. Kaiser, The Same-Actor Inference of Nondiscrimination: Moral Credentialing and the Psychological and Legal Licensing of Bias, 104 Calif. L. Rev. 1, 6, 39–40 (2016); Kerri Lynn Stone, Taking in Strays: A Critique of the Stray Comment Doctrine in Employment Discrimination Law, 77 Mo. L. Rev. 149 (2012); Natasha T. Martin, Pretext in Peril, 75 Mo. L. Rev. 313 (2010); Chad Derum & Karen Engle, The Rise of the Personal Animosity Presumption in Title VII and the Return to “No Cause” Employment, 81 Tex. L. Rev. 1177 (2003); Ann C. McGinley, The Emerging Cronyism Defense and Affirmative Action: A Critical Perspective on the Distinction Between Colorblind and Race-Conscious Decision Making Under Title VII, 39 Ariz. L. Rev. 1003 (1997).
See, e.g., Local 28, Sheet Metal Workers’ International Association v. EEOC, 478 U.S. 421, 448 (1986) (“[I]t was clear to Congress that ‘[t]he crux of the problem [was] to open employment opportunities for Negroes in occupations which have been traditionally closed to them,’ . . . and it was to this problem that Title VII’s prohibition against racial discrimination in employment was primarily addressed.” (alterations in original) (quoting United Steelworkers v. Weber, 443 U.S. 193, 203 (1979))); Weber, 443 U.S. at 202 (“Congress’ primary concern in enacting the prohibition against racial discrimination in Title VII of the Civil Rights Act of 1964 was with ‘the plight of the Negro in our economy.’” (citation omitted)); cf. Regents of the Univ. of Cal. v. Bakke, 438 U.S. 265, 285 (1978) (noting that in enacting Title VI of the 1964 Civil Rights Act, the “problem confronting Congress was discrimination against Negro citizens at the hands of recipients of federal moneys” and that “the legislators were dealing with the real and pressing problem of how to guarantee those citizens equal treatment”).
In particular, the first vignette illustrates how existing doctrine credits as nondiscriminatory even highly suspicious (and perhaps factually inaccurate) justifications for negative actions taken by employers against workers of color. See supra note 2 and accompanying text. See also infra Sections III.A, III.B.2.
For two well-known examples of implicit bias literature that address employment discrimination law, see, for example, Jerry Kang et al., Implicit Bias in the Courtroom, 59 UCLA L. Rev. 1124 (2012), and Linda Hamilton Krieger, The Content of Our Categories: A Cognitive Bias Approach to Discrimination and Equal Employment Opportunity, 47 Stan. L. Rev. 1161 (1995). Another example is Victor D. Quintanilla, Critical Race Empiricism: A New Means to Measure Civil Procedure, 3 UC Irvine L. Rev. 187 (2013), which mobilizes implicit bias research related to the phenomenon of “aversive racism,” as well as the concept of “lay theories of racism,” to analyze the effect of changes in pleading rules on employment discrimination cases brought by minority plaintiffs.
See Eyer, supra note 10, at 1286, 1291 (noting that “the overwhelming majority of scholarly attempts to posit a cause for the difficult odds that discrimination litigants face are based on a loose, intuitive approach to understanding the phenomenon, with little or no empirical foundation” and proposing an explanation grounded in psychological research on attribution).
For example, searches for the term “Social Dominance Theory” in the Law Journal Library on HeinOnline and Westlaw’s “Law Reviews and Journals” database conducted on April 24, 2019, resulted in only fifty-nine and forty-seven hits respectively. By contrast, similar searches for another concept from social psychology that has infiltrated legal scholarship to a significant degree, “implicit bias,” result in about 3,000 hits on each. When Social Dominance Theory has been referred to in legal scholarship, this has often been in the criminal law context. See, e.g., Darren Lenard Hutchinson, “Continually Reminded of Their Inferior Position”: Social Dominance, Implicit Bias, Criminality, and Race, 46 Wash. U. J.L. & Pol’y 23, 92 (2014); Devon W. Carbado & Patrick Rock, What Exposes African Americans to Police Violence, 51 Harv. C.R.-C.L. L. Rev. 159, 175–79 (2016). But see Michael Selmi, Subtle Discrimination: A Matter of Perspective Rather Than Intent, 34 Colum. Hum. Rts. L. Rev. 657, 673–76 (2003) (briefly summarizing certain tenets of SDT in the employment discrimination context, but concluding that while SDT “offers important insights into how we arrived where we are today, [it does so] without providing much guidance into how we might obtain a greater understanding of the continued pervasiveness of discrimination”); Darren Lenard Hutchinson, Preventing Balkanization or Facilitating Racial Domination: A Critique of the New Equal Protection, 22 Va. J. Soc. Pol’y & L. 1, 40–42 (2015) (applying SDT in an analysis of constitutional equal protection doctrine). This Article argues that this gap in the legal literature ought to be filled.
See infra Section II.A.
See infra Part III.
See infra Part IV.
Two recent law review articles take specific interest in the notion of “symmetry” as an animating principle in employment discrimination law. See Naomi Schoenbaum, The Case for Symmetry in Antidiscrimination Law, 2017 Wis. L. Rev. 69, 76 (stating that “the term symmetrical . . . describe[s] an antidiscrimination law that does not limit the scope of persons who are protected on the basis of a given trait, nor does it limit the direction of discrimination prohibited”); Bradley A. Areheart, The Symmetry Principle, 58 B.C. L. Rev. 1085, 1088 (2017) (describing the “symmetry principle” as “mandat[ing] that once certain attributes or characteristics are identified as worthy of antidiscrimination protection, all groups within that universal ground must be protected”). Both list Title VII race discrimination provisions as an example of a symmetrical law. See Schoenbaum, supra, at 79; Areheart, supra, at 1088–89. Both Schoenbaum and Areheart argue that, as a general matter, symmetrical antidiscrimination laws are preferable to asymmetrical ones, though Areheart concludes that for some protected characteristics, including race, asymmetry may be justified in certain contexts. Areheart, supra, at 1133–35; Schoenbaum, supra, at 73–75.
See, e.g., Schoenbaum, supra note 28, at 131 (suggesting that affirmative action programs could be made symmetrical if they allowed privileged groups to take advantage of them, implying that current approaches to affirmative action are asymmetrical in favor of disadvantaged groups); Areheart, supra note 28, at 1126, 1135 (characterizing affirmative action as a “limited departure” from the norm of symmetry, but ultimately concluding that this departure is warranted); see also infra Section III.B.3 (discussing the attack on affirmative action as illegitimate “preferences” in favor of racial minorities in greater detail).
My argument in this Article does not dispute the broad claim that under Title VII both white workers and workers of color are included in the groups protected by the statute. That much has been settled. See McDonald v. Santa Fe Trail Transp. Co., 427 U.S. 273, 280 (1976). I contest the claim that, to the extent that Title VII features asymmetries, it favors workers of color.
As I explain more in Part III below, by “traditional” cases I mean cases in which workers claim that an employer took an adverse action against them on the basis of their race while the employer claims to have taken the action for nonracial reasons (i.e., independently of any affirmative action program). See infra Part III.
See infra Section III.B.2.
See infra Section III.B.3. In other words, my argument is more nuanced than simply to say that employment discrimination law always favors white plaintiffs or that federal judges are overtly biased against workers of color. The asymmetry in employment discrimination law that I describe in this Article can be viewed instead as the outcome of judging and interpreting over time in which some types of cases (“traditional” disparate treatment claims) are associated with challenges to existing hierarchy because they are most often brought by workers of color. As a result of the psychological processes that I discuss in more detail below, the doctrine that develops to adjudicate such cases becomes comparatively harsh to plaintiffs because the federal judiciary, on the whole, prefers to maintain existing hierarchy. When white plaintiffs (though more infrequently) bring such types of cases, they thus also face a hostile doctrine and may frequently lose as well. But see infra Section III.B.1 (showing how even in “traditional” cases there are asymmetries comparatively favoring the claims of white workers in some circuits). Based on existing empirical data, for example, it is not clear that white employment discrimination plaintiffs have significantly higher success rates than nonwhite plaintiffs in all instances. Compare Berrey et al., supra note 11, at 70 (noting that white plaintiffs are more likely to avoid dismissal of their pleadings than plaintiffs of color, but that thereafter, white plaintiffs and plaintiffs of color fare similarly), and Pat K. Chew & Robert E. Kelley, The Realism of Race in Judicial Decision Making: An Empirical Analysis of Plaintiffs’ Race and Judges’ Race, 28 Harv. J. Racial & Ethnic Just. 91, 99, 100 tbl.1 (2012) (in sample of reported racial harassment decisions, white plaintiffs were successful slightly more often than black plaintiffs, but less often than Hispanic plaintiffs), with Parker, supra note 12, at 907 (noting that in one sample of published race discrimination cases, whites had a lower success rate as plaintiffs than workers of color). It is somewhat difficult to disentangle those numbers. One could argue that, given the continued dominance of white workers in the economy overall, see infra Section II.B., and comparatively low levels of self-reporting by white workers that they are experiencing bias in the workplace, see Berrey et al., supra note 11, at 47, white workers are much less likely to be discriminated against based on their race (as opposed to other personal characteristics, for example); that the race discrimination claims that white plaintiffs bring thus may well include a greater proportion of weak claims; and that, as a result, white plaintiffs should be comparatively less successful over the run of cases. Therefore, the fact that they appear to do similarly as well as minority plaintiffs may itself suggest that federal judges apply employment discrimination law comparatively more favorably to white plaintiffs. It is also possible that affirmative action cases, in which the doctrine is friendlier to (white) plaintiffs, see infra Section III.B.3, play at least some role in raising white plaintiffs’ win rates to comparable levels with minority plaintiffs. Unfortunately, existing empirical studies do not distinguish between “traditional” and affirmative action cases or give an indication of how many affirmative action cases are being brought. Additionally, it is difficult, if not impossible, to know how many affirmative action programs are not adopted because employers are wary of testing a more hostile doctrine.
See infra Section III.B.2.
See infra Section III.B.3.
See infra Section II.B.
In the Court’s most well-known disparate treatment case, McDonnell Douglas Corp. v. Green, 411 U.S. 792, 800 (1973), the Court described this purpose as being “to assure equality of employment opportunities and to eliminate those discriminatory practices and devices which have fostered racially stratified job environments to the disadvantage of minority citizens.” In the Court’s most well-known disparate impact case, Griggs v. Duke Power Co., 401 U.S. 424, 429–30 (1971), the Supreme Court declared that Congress’ objective in enacting Title VII was “to achieve equality of employment opportunities and remove barriers that have operated in the past to favor an identifiable group of white employees over other employees.”
See infra Part IV.
Cheryl Harris has recently made a similar point, that doctrinal asymmetries that disfavor workers of color can be uncovered through a comparison of Title VII doctrine and constitutional equal protection analysis. Cheryl I. Harris, Limiting Equality: The Divergence and Convergence of Title VII and Equal Protection, 2014 U. Chi. Legal F. 95, 98 (2014) (“[C]ompeting views have come to cohere in endorsing interpretations that have weakened antidiscrimination protection for non-whites, while enabling whites’ challenges to those same remedial measures. The Court’s comparative reasoning often is not racially neutral: The interplay between Title VII and equal protection has functioned asymmetrically to (re)produce an unequal doctrinal terrain.”). This Article undertakes a similar project, though with a different target and conceptual framework.
See Erik J. Girvan, On Using the Psychological Science of Implicit Bias to Advance Anti-Discrimination Law, 26 Geo. Mason U. C.R. L.J. 1, 75–76 (2015) (noting that psychologists have “found strong evidence of personality traits that relate to people’s political ideology and whether they support policies that implicate equality” including Social Dominance Orientation (SDO), an aspect of SDT further discussed below, and that “[p]roposals for and arguments supporting expanded anti-discrimination rights and duties that fail to take these fundamental differences of values and worldview into account are far less likely to be accepted”).
There are two broad types of employment discrimination claims under Title VII: disparate treatment claims and disparate impact claims. Disparate treatment claims involve cases in which an employer is alleged to have intentionally discriminated against a worker based on a legally protected characteristic, such as race, sex, age, etc. See infra Section III.A. Disparate impact claims involve cases in which an employer’s business practice, which was not used or adopted to intentionally discriminate based on a protected characteristic, nevertheless has a disproportionate negative impact on a protected group (for example, an employment test that screens out disproportionately more women than men). I focus on disparate treatment cases because this theory of discrimination is by far the most frequently asserted. See, e.g., Nielsen et al., supra note 13, at 11, 48 (noting that in a randomly selected sample of 1,779 employment discrimination cases filed in seven federal judicial districts between 1987 and 2003, ninety-eight percent of cases involved disparate treatment claims and only four percent involved disparate impact claims). Disparate impact law also raises a number of distinct analytical questions, and thus requires separate analysis.
I focus on race discrimination because, as much research shows, “for most of U.S. history, race . . . has been and remains the primary basis of social stratification.” Sidanius & Pratto, supra note 1, at 61. Thus, when it comes to issues of social power and hierarchy in the United States, and the law’s involvement in both, it is critical to understand the dynamics of race and racial ideology. Indeed, the 1964 Civil Rights Act and Title VII emerged out of one of the most contentious periods of race relations in the United States. See generally David Benjamin Oppenheimer, Kennedy, King, Shuttlesworth and Walker: The Events Leading to the Introduction of the Civil Rights Act of 1964, 29 U.S.F. L. Rev. 645 (1995); Clay Risen, The Bill of the Century: The Epic Battle for the Civil Rights Act (2014). As noted above, race discrimination cases are also one of the—if not the—most frequently litigated type of employment discrimination case. See supra note 11. Increasing our understanding of the dynamics driving race discrimination cases should improve our understanding of employment discrimination law generally. The limitation to race and racial discrimination is also necessary to keep the scope of an already complex argument manageable and accessible. Analyzing other axes of power and hierarchy also covered by antidiscrimination law, particularly sex and gender, and the intersections between them, are also crucially important, though sufficiently complex to require separate focused analysis.
411 U.S. 792 (1973).
Because Social Dominance Theory has not been used much in the legal literature to date, this Section provides a more detailed overview of the theory.
See, e.g., Sidanius & Pratto, supra note 1, at 31–32. “Group-based” hierarchy broadly means that in such a society, “one’s social status, influence, and power are also a function of one’s group membership and not simply of one’s individual abilities or characteristics.” Id. at 32.
Id. at 33.
Id. at 61.
See, e.g., Felicia Pratto et al., Social Dominance Theory and the Dynamics of Intergroup Relations: Taking Stock and Looking Forward, 17 Eur. Rev. Soc. Psychol. 271, 273–74 (2006).
See, e.g., Jim Sidanius et al., Social Dominance Theory: Its Agenda and Method, 25 Pol. Psychol. 845, 847–48 (2004).
Sidanius & Pratto, supra note 1, at 227.
Id. at 45–46.
SDO is a survey-based measure in which people indicate their agreement with various statements such as: “[s]ome groups of people are simply inferior to other groups.” Id. at 67. Researchers have attempted to improve and fine-tune this measure over time. See generally id. at 61–102. Recent research has shown that SDO has two subdimensions, a Dominance dimension (SDO-D) reflecting a person’s support for active suppression, domination, and aggression towards subordinates; and an Anti-Egalitarianism dimension (SDO-E) reflecting a person’s support for more subtle means of creating and sustaining hierarchy, such as cultural ideologies. See, e.g., Arnold K. Ho et al., The Nature of Social Dominance Orientation: Theorizing and Measuring Preferences for Intergroup Inequality Using the New SDO7 Scale, 109 J. Personality & Soc. Psychol. 1003, 1004–05 (2015). SDO-D and SDO-E are highly correlated with each other.
See Sidanius et al., supra note 50, at 849 (“Whereas SDO is an important component of the broader model, social dominance theory is neither primarily nor exclusively concerned with this individual-difference construct.”); see also Jim Sidanius et al., Social Dominance Theory: Explorations in the Psychology of Oppression, in The Cambridge Handbook of the Psychology of Prejudice 152 (Chris G. Sibley & Fiona Kate Barlow eds., 2017).
See Sidanius et al, supra note 50, at 858; Felicia Pratto, The Puzzle of Continuing Group Inequality: Piecing Together Psychological, Social, and Cultural Forces in Social Dominance Theory, in 31 Advances in Experimental Social Psychology 191, 252 (Mark P. Zanna ed., 1999) (explaining that “we have not theorized that people high on SDO are abnormal and fundamentally different”).
Sidanius et al., supra note 50, at 858.
Nour S. Kteily et al., Social Dominance Orientation: Cause or ‘Mere Effect’?: Evidence for SDO as a Causal Predictor of Prejudice and Discrimination Against Ethnic and Racial Outgroups, 47 J. Experimental Soc. Psychol. 208, 209 (2011).
Id. at 213 (citations omitted).
This finding is most comprehensively addressed in the 2011 meta-analysis of I-Ching Lee and colleagues, which reviewed data from 206 samples in 118 independent studies. I-Ching Lee et al., Intergroup Consensus/Disagreement in Support of Group-Based Hierarchy: An Examination of Socio-Structural and Psycho-Cultural Factors, 137 Psychol. Bull. 1029, 1036 (describing studies reviewed and findings). The authors found that average SDO levels of whites and males were consistently and significantly higher than those of women and of nonwhite racial groups in the United States (though perhaps with the exception of Asian American groups). See id. at 1038 tbl.3 (men versus women), 1043 tbl.5 (whites vs. nonwhites), 1057–64 (listing all studies and effect sizes). See also Ho et al., supra note 53, at 1020–21, 1020 tbl.10, 1021 tbl.11; Pratto, supra note 55, at 230, 231 tbl.VI.
See, e.g., Pratto et al., supra note 49, at 275–81; infra Figure 1.
See Pratto, supra note 55, at 248. As a result, “there is some redundancy in the system” making it “unlikely that interrupting one process will break the whole system, adding to the system’s resiliency.” Id.
Taken from Sidanius & Pratto, supra note 1, at 40.
See, e.g., Sidanius & Pratto, supra note 1, at 103.
Id. at 45.
Id. at 104.
See, e.g., Pratto et al., supra note 49, at 287–88; Sidanius & Pratto, supra note 1, at 104–05, 118–23. By extension, as illustrated by minus signs for the diagonals in Figure 2, higher levels of SDO and resulting support for HE legitimizing myths also lead to comparatively lower support for HA social policies. Similarly, lower levels of SDO and greater support for HA legitimizing myths lead to lower support for HE social policies.
Taken from Pratto et al., supra note 49, at 282 (citing Sidanius & Pratto, supra note 1, at 105).
See, e.g., Sidanius & Pratto, supra note 1, at 89; see also Eric D. Knowles et al., On the Malleability of Ideology: Motivated Construals of Color Blindness, 96 J. Personality & Soc. Psychol. 857, 858 (2009). For example, as discussed further below, the ideology of “colorblindness” could be interpreted as a distributive justice commitment to equal (i.e., colorblind) societal outcomes, in which case it would function as a HA ideology; or it could be interpreted as a procedural justice commitment to formally equal treatment, irrespective of outcomes, in which case it has the tendency to perpetuate existing hierarchical distributions of social resources and would function as a HE ideology. See infra notes 109–14 and accompanying text.
See Jim Sidanius et al., Legitimizing Ideologies: The Social Dominance Approach, in The Psychology of Legitimacy 310–11 (John T. Jost & Brenda Major eds., 2001).
Id. at 319. For example, in one random sample of Los Angeles County residents, researchers found that the relationship between SDO and support for the death penalty (a HE social policy) was much higher for white Americans than it was for African-Americans. Id. at 321–22.
See, e.g., Sidanius et al., supra note 50, at 851–52.
See Pratto et al., supra note 49, at 276.
See, e.g., Sidanius et al., supra note 50, at 852.
See, e.g., id.; Hillary Haley & Jim Sidanius, Person-Organization Congruence and the Maintenance of Group-Based Social Hierarchy: A Social Dominance Perspective, 8 Group Processes & Intergroup Rel. 187, 189–90 (2005).
See, e.g., Pratto et al., supra note 49, at 277; Sidanius et al., supra note 50, at 852.
Sidanius & Pratto, supra note 1, at 52. This “law of increasing disproportion operates within all three forms of group-based stratification (i.e., age system, gender system, and arbitrary-set system).” Id.
See Haley & Sidanius, supra note 74, at 195.
Serge Guimond et al., Does Social Dominance Generate Prejudice? Integrating Individual and Contextual Determinants of Intergroup Cognitions, 84 J. Personality & Soc. Psychol. 697, 701–06 (2003).
See id. at 705–06. Follow-up research (unfortunately not involving lawyers) confirmed this socialization effect in psychology and suggests that part of this effect may be related to the fact that exposure to a HA environment reduces beliefs in “genetic determinism,” or the belief that genetics (rather than environmental forces) are the main driver of human behavior and personality. See Michaël Dambrun et al., Why Does Social Dominance Orientation Decrease with University Exposure to the Social Sciences? The Impact of Institutional Socialization and the Mediating Role of “Geneticism,” 39 Eur. J. Soc. Psychol. 88, 97 (2009).
Guimond et al., supra note 78, at 707–09. The authors suggested that these effects may have been the “result of [participants] learning that they have what it takes to occupy a dominant social position. Seeing themselves at the top leads them to perceive inequality in a different manner. They are now ready to accept that certain groups should be at the top and others at the bottom.” Id. at 712.
Id. at 712–13, 713 fig.5.
Id. at 715.
Id. at 716.
See, e.g., James Sidanius, The Psychology of Group Conflict and the Dynamics of Oppression: A Social Dominance Perspective, in Explorations in Political Psychology 183, 201 (Shanto Iyengar & William J. McGuire eds., 1993) (“[T]he legal and criminal justice system will be one of the major instruments used in establishing and maintaining the hierarchical caste system.”).
See Sidanius & Pratto, supra note 1, at 42–43. I call this idea the “plausible deniability expectation.”
See supra note 45 and accompanying text.
In the context of race, one particularly clear example of the law playing a hierarchy-enhancing role in service of protecting a racial hierarchy grounded in white supremacy is the Naturalization Act of 1790, which restricted naturalization, and thus the rights attaching to citizenship, to “free white person[s].” ch. 3, § 1, 1 Stat. 103. While this statute was amended during the Reconstruction era to also allow naturalization by “aliens of African nativity and . . . persons of African descent,” Act of July 14, 1870, ch. 254, § 7, 16 Stat. 254, 256, outside of these categories immigrants generally had to continue to claim whiteness to be able to naturalize deep into the 20th century. The U.S. Supreme Court interpreted whiteness very restrictively in this context—for example, rejecting claims to whiteness by immigrants from India and Japan. See United States v. Thind, 261 U.S. 204 (1923); Ozawa v. United States, 260 U.S. 178 (1922).
The Civil Rights Act of 1964, of which Title VII is a part, is an example. Civil Rights Act of 1964, Pub. L. 88‑352, 78 Stat. 241. It arose out of a period of mass struggle by members of oppressed groups who, along with various allies, were able to build up enough pressure to get Congress to outlaw an existing caste system of racial segregation and explicit race-based discrimination in areas ranging from public accommodations to public facilities and private employment. See generally, e.g., Risen, supra note 42; Oppenheimer, supra note 42. While there are disputes about how much interference with existing hierarchy lawmakers really intended to allow, see infra note 230; and while this Article shows how some of the Act’s employment provisions have come to be interpreted in a hierarchy-enhancing fashion; it is clear that, upon its passage, the Act represented a step towards hierarchy attenuation. For an argument that the law is likely to operate in ways beneficial to minorities only when this also serves the interests of whites, see Derrick A. Bell, Jr., Brown v. Board of Education and the Interest-Convergence Dilemma, 93 Harv. L. Rev. 518, 523 (1980) (proposing that important race cases have arguably been guided by a principle of “interest convergence” by which the “interest of blacks in achieving racial equality will be accommodated only when it converges with the interests of whites” and “the [F]ourteenth [A]mendment, standing alone, will not authorize a judicial remedy providing effective racial equality for blacks where the remedy sought threatens the superior societal status of middle and upper class whites”).
See, e.g., Harris, supra note 39, at 143–44 (“The point is not simply that the divisions among the justices mirror the political divisions within the country regarding racial reform; rather, it is that the law also is a site where racial ideology is constituted, divisions are constructed, and claims are legitimated.”).
In Parts III and IV, I show how doing so helps explain why employment race discrimination doctrine looks the way it does today—and how it can be reformed.
Both asymmetrical narrowing and proof asymmetry are terms that I have coined for purposes of this Article and derived from my own review of SDT literature. They do not represent principles developed by SDT scholars in social psychology but represent my application of relevant concepts to the analysis of race discrimination law.
See, e.g., Pratto, supra note 55, at 232 (noting that various studies “suggest that people in dominant groups will discriminate to their own group’s advantage more than people in subordinate groups”).
While Obama administration appointments created a significantly more diverse judiciary, see Josh Katz, Older Judges and Vacant Seats Give Trump Huge Power to Shape American Courts, N.Y. Times (Feb. 14, 2017), https://www.nytimes.com/interactive/2017/02/14/upshot/trump-poised-to-transform-american-courts.html?_r=0 [https://perma.cc/PKJ5-Y492], the federal judiciary as a whole is still filled mostly with dominant group members. As of 2017, 75% of active circuit judges were white (45.6% white males) and slightly less than 75% of active district court judges were white (49.3% white males). See Barry J. McMillion, Cong. Research Serv., R43426, U.S. Circuit and District Court Judges: Profile of Select Characteristics 8 fig.4, 20 fig.12 (2017), https://fas.org/sgp/crs/misc/R43426.pdf [https://perma.cc/MDD4-T842]. While these numbers reflect a much more diverse current federal judiciary (overwhelmingly due to President Obama’s nominations) than has historically been the case, see Jonathan K. Stubbs, A Demographic History of Federal Judicial Appointments by Sex and Race: 1789-2016, 26 La Raza L.J. 92, 99–112 (2016), the Trump administration has moved decisively to reverse this trend and has almost exclusively nominated white males for federal judgeships. See, e.g., Catherine Lucey & Meghan Hoyer, Trump Choosing White Men as Judges, Highest Rate in Decades, Chi. Trib. (Nov. 13, 2017), http://www.chicagotribune.com/news/nationworld/politics/ct-trump-blacks-judges-20171113-story.html [https://perma.cc/6G59-JVDT?type=image] (noting that ninety-one percent of Trump administration’s nominees were white and eighty-one percent male).
Sidanius, supra note 69, at 319.
For the relationship between dominant group status and SDO, see supra note 59 and accompanying text.
See supra notes 80–82 and accompanying text.
While I am not aware of research directly measuring the SDO levels of American judges, SDT researchers have suggested that judges should have comparatively higher SDO levels. See Jim Sidanius et al., Consensual Racism and Career Track: Some Implications of Social Dominance Theory, 12 Pol. Psychol. 691, 713 (1991) (suggesting—though study dataset did not allow a direct conclusion in this regard—that “there should be power differences for different institutions even within the general fields of law and business, and we would expect that consensual racism, social dominance orientation, political ideology, and the like should differ with those as well” so that “students preparing for careers as criminal prosecutors, corporate lawyers, politicians, and judges should be more discrimination-prone than students preparing for careers in civil rights or labor law or as public defenders”). While this Article discusses much research showing how people with comparatively high levels of SDO are more likely to behave in hierarchy-enhancing ways, arguments in this regard are not dependent on some absolute measure of SDO or branding such people as fundamentally “different.” SDO is one aspect of life in group-based social hierarchies. While SDO provides important explanatory value in contexts such as the topic discussed in this Article, it acts in concert with other causes of inequality, such as more structural forces of institutional discrimination. See supra notes 50–61 and accompanying text. Still, because social psychology research often focuses on individual psychological differences, SDO has been the SDT concept most-often used in empirical research, and this research has focused on how people that are comparatively high versus low on SDO behave differently from each other. Thus, this Article relies on much of this research. This should be seen not as a claim that there are “good” and “bad” people that drive social outcomes solely as a function of their SDO levels. Rather, it reflects the fact that in complex systems of group-based hierarchy, SDO will interact as both cause and effect with other aspects of the system to give it a high level of stability. See Pratto, supra note 55, at 252 (explaining that SDT does not situate “the cause of group dominance in a few deviant, abnormal, exceptionally high SDO people,” because “to do so would paint the psychological states that sustain group dominance as far more aberrant and unusual than they are” given that “the everyday normal discriminatory social patterns of group dominance societies and tacit tolerance of oppression by everyone accomplishes much of the work that sustains group dominance”). What also follows from the contextual nature of SDO is that the effect of SDO on HE behavior is not dependent on some absolute threshold level of SDO. Rather, those with comparatively high levels of SDO in a given context will tend to act to preserve the hierarchy to a greater degree. See id. at 205. Still, while contextual factors may influence absolute SDO levels, people’s tendency to be comparatively high vs. low in SDO is generally fairly stable over time. See Sidanius et al., supra note 50, at 850–51.
See supra note 70 and accompanying text.
See, e.g., Nour S. Kteily et al., Hierarchy in the Eye of the Beholder: (Anti‑)Egalitarianism Shapes Perceived Levels of Social Inequality, 112 J. Personality & Soc. Psychol. 136, 142 (2017); Ho et al., supra note 53, at 1014 tbl.5; Arnold K. Ho et al., Social Dominance Orientation: Revisiting the Structure and Function of a Variable Predicting Social and Political Attitudes, 38 Personality & Soc. Psychol. Bull. 583, 594 tbl.5 (2012).
Kteily et al., supra note 99 (Supp. Pt. 10); Id. at 142 & tbl.1. These statements were part of an index scale that also included statements about integration in society and schools. Id. This index score was strongly and negatively related to SDO—i.e., higher SDO meant more opposition to antidiscrimination policy. Id. at 142 & tbl.1. Previous research had found consistent results with similar scales. See, e.g., Ho et al., supra note 53, at 1014 tbl.5.
Kteily et al., supra note 99, at 142.
See Richard P. Eibach & Thomas Keegan, Free at Last? Social Dominance, Loss Aversion, and White and Black Americans’ Differing Assessments of Racial Progress, 90 J. Personality & Soc. Psychol. 453, 461–62 (2006).
See Miguel M. Unzueta et al., Social Dominance Orientation Moderates Reactions to Black and White Discrimination Claimants, 54 J. Experimental Soc. Psychol. 81, 85 (2014).
Id. at 87. Such pro-hierarchy tendencies need not manifest themselves consciously to be operative. SDT research has also found that people with comparatively high SDO levels may be especially likely to favor dominant groups and their interests implicitly when the hierarchy is under threat. See, e.g., Felicia Pratto & Margaret Shih, Social Dominance Orientation and Group Context in Implicit Group Prejudice, 11 Psychol. Sci. 515, 516–18 (2000) (finding that high SDO study participants are more likely to implicitly discriminate against outgroups when their status is under threat); see also Pratto, supra note 55, at 234–35. The limited direct evidence that exists on actual judges shows that judges, like the rest of the population, have implicit racial biases (generally in favor of whites) and that these biases can impact their decision-making—though the exact relationship between implicit judicial bias and judicial decision-making requires further investigation. See Jeffrey J. Rachlinski et al., Does Unconscious Racial Bias Affect Trial Judges?, 84 Notre Dame L. Rev. 1195, 1210–26 (2009) (examining implicit bias against blacks and its potential influence on criminal sentencing for offenses such as shoplifting, robbery, and battery); Justin D. Levinson et al., Judging Implicit Bias: A National Empirical Study of Judicial Stereotypes, 69 Fla. L. Rev. 63, 103–11 (2017) (examining, among other things, implicit bias against Asian Americans and its potential influence on white-collar crime sentencing); cf. Eyer, supra note 10, at 1325 (“[E]xisting findings from the judge context . . . strikingly support the conclusion that most of the time psychology studies will provide a helpful basis for understanding real-world adjudicative behavior.”).
See supra notes 64–67 and accompanying text (explaining that in the SDT framework ideologies serve as justifications that mediate the relationship between people’s motivations to maintain hierarchy and their support for social policies). See also supra note 85 and accompanying text (arguing that law can be conceptualized as a form of social policy within the SDT framework).
See supra Section II.A.4.
See supra note 104; infra notes 110–14 and accompanying text.
For insightful analyses of the role of conservative judicial appeals to colorblindness in shaping equal protection doctrine in a way that subordinates the interests and claims of racial minorities, for example, see the works of Ian Haney-López infra notes 260 and 298.
See, e.g., Knowles et al., supra note 68, at 859–60.
By contrast, if colorblindness is framed and understood as a distributive justice concept, support for colorblindness may be higher among people with low levels of SDO.
Conversely, high and low SDO participants were equally likely to construe colorblindness as procedural versus distributive justice mandates when not under threat. See Knowles et al., supra note 68, at 860–62, 862 fig.1. Notions of threat could be effectively induced even by simply asking white Americans to self-identify their race. See id. at 860.
Id. at 861–64.
Id. at 864–66. Significantly, in the study that provided evidence for point (3), the threat to the hierarchy was induced by telling readers that “contrary to popular opinion, recent research has found that affirmative action policies have resulted in fewer economic opportunities for Whites.” Id. at 864–65. The trend among low SDO participants was in the opposite direction to that of high SDO participants. While overall more supportive of ideas of colorblindness as a general matter, they were less likely to construe colorblindness in procedural justice terms and became less supportive of colorblindness and procedural justice when the hierarchy was under threat. Id. at 864 fig.2, 865 fig.3, 866 fig.4.
Id. at 867. Similarly, Chow and Knowles found in one of their other studies that participants high on the SDO-E subdimension, which is predictive of preferences for subtle means of protecting existing hierarchy, see supra note 53, were more likely to endorse colorblindness (there defined as opposing government collection of racial data and governmental classification of individuals by race) when they were made to feel that the hierarchy was under threat by reading about an organization described as wanting to achieve equality between blacks and whites through the promotion of strong affirmative action programs and reparations for slavery than when reading about a nonthreatening organization. See Rosalind M. Chow & Eric D. Knowles, Taking Race Off the Table: Agenda Setting and Support for Color-Blind Public Policy, 42 Personality & Soc. Psychol. Bull. 25, 28–29, 29 fig.1. Low SDO-E participants, on the other hand, were less likely to support colorblindness under those circumstances compared to the control condition. Id.
As legal scholar Gary Blasi and social psychologist John Jost have stated in support of an argument for applying a different social-psychological theory, that of System Justification Theory, to the analysis of law: “To posit the existence of a motive is not to presume a specific result in every case.” Gary Blasi & John T. Jost, System Justification Theory and Research: Implications for Law, Legal Advocacy, and Social Justice, 94 Calif. L. Rev. 1119, 1137 (2006). I concur with their assessment, and my arguments should not be read to suggest that predictions that can be made on the basis of SDT, or factors that have been identified by SDT, will always be relevant or will be determinative in all circumstances.
Jerry Kang, Implicit Bias and the Pushback from the Left, 54 St. Louis U. L.J. 1139, 1147 (2010).
See, e.g., Erik J. Girvan, When Our Reach Exceeds Our Grasp: Remedial Realism in Antidiscrimination Law, 94 Or. L. Rev. 359, 374–75 (2016) (citations omitted).
See, e.g., Winthrop D. Jordan, White Over Black: American Attitudes Toward the Negro, 1550–1812, at 321–25 (1968) (discussing compromises made on slavery in the framing of the Constitution); Ira Katznelson, When Affirmative Action Was White: An Untold History of Racial Inequality in Twentieth-Century America (2005) (describing the role of Southern senators in structuring New Deal legislation such that it would least interfere with ability of Southern states to maintain system of institutionalized and legalized segregation); Nell Irvin Painter, The History of White People 65–66, 71 (2010) (providing examples of racial classification schemes of “race scientists” inevitably ordering racial groups into clear hierarchical systems); David R. Roediger, The Wages of Whiteness: Race and the Making of the American Working Class 144–50 (rev. ed., 2007) (discussing the dynamics between two subordinated groups in racial hierarchy, blacks and the Irish, in the context of the formation of working class identity in white-supremacist nineteenth century United States); Richard Rothstein, The Color of Law: A Forgotten History of How Our Government Segregated America 43–48 (2017) (analyzing the role of law and government-sponsored residential segregation in maintaining racial hierarchy in United States); Cheryl I. Harris, Whiteness as Property, 106 Harv. L. Rev. 1707, 1715–16, 1724, 1731 (1993) (discussing whiteness as a concept akin to traditional property interests that is grounded in white supremacy, the development of slavery, and the dispossession of Native Americans, and how it is maintained in various ways through law).
See, e.g., Rachlinski, supra note 104, at 1201–02 (reviewing research findings that implicit bias may account for roughly six percent of variation in white-black interracial behavior); see also Anthony G. Greenwald et al., Statistically Small Effects of the Implicit Association Test Can Have Societally Large Effects, 108 J. Personality & Soc. Psychol. 553, 560 (2015) (reviewing findings of different meta-analyses, finding that regardless of approach in selecting relevant studies, more than four percent of discrimination-related variation can be accounted for and explaining how even small statistical effects can have large social consequences when affecting large numbers of people or the same person repeatedly).
Cf. Eyer, supra note 10, at 1322 (“[I]n the absence of (rarely available) direct studies on judges or jurors—there are few more persuasive indicators of a common underlying cause [for the difficulties faced by employment discrimination plaintiffs] than . . . phenomenological overlap between the findings of psychology scholars and the pattern of decisions in the courts.”); Girvan, supra note 40, at 75–76 (encouraging analysis of how SDO may influence judges).
Eyer, supra note 10.
See, e.g., id. at 1291–93.
Id. at 1293–318.
See infra notes 201–09 and accompanying text.
See Eyer, supra note 10, at 1324–27.
See, e.g., Christina L. Boyd, Representation on the Courts? The Effects of Trial Judges’ Sex and Race, 69 Pol. Res. Q. 788, 793–94, 794 fig.2 (2016) (finding that in employment discrimination cases brought by the EEOC, black judges were significantly more likely to find in favor of plaintiffs in both race and sex discrimination cases); Weinberg & Nielsen, supra note 12, at 342 (finding minority judges significantly less likely to grant summary judgment for defendant in broader set of employment discrimination cases); Chew & Kelley supra note 33, at 100 tbl.1 (showing that African-American judges are significantly more likely to rule in favor of racial harassment plaintiffs).
See Devon W. Carbado & Daria Roithmayr, Critical Race Theory Meets Social Science, 10 Ann. Rev. L. & Soc. Sci. 149, 157–58 (2014).
Id. at 157.
Id. at 158.
See Noah D. Zatz, Special Treatment Everywhere, Special Treatment Nowhere, 95 B.U. L. Rev. 1155, 1158 (2015).
Id. at 1156.
Id. at 1158.
Id. at 1162–63.
Id. at 1158–59.
Id. at 1158–64 (giving examples of such remedial scenarios, including cases in which making distinctions between workers based in part on their race is necessary to identify the victims of prior discrimination who are entitled to a remedy for that discrimination).
Id. at 1162.
Id. at 1164.
Id. at 1157–59.
Id. at 1157.
See generally Nancy Gertner, Losers’ Rules, 122 Yale L.J. Online 109, 110–11 (2012) (“[J]ust as empirical studies highlight the stubborn persistence of discrimination at all levels of jobs and in salaries, federal discrimination law assumes the opposite. . . . It is as if the bench is saying: ‘Discrimination is over. The market is bias-free.’”).
These arguments are presented and analyzed in greater detail in Part III.
See infra Section III.B.2.
See infra Sections III.B.1, III.B.3.
See Victoria C. Plaut, Diversity Science: Why and How Difference Makes a Difference, 21 Psychol. Inquiry 77, 79 (2010).
See, e.g., U.S. Bureau of Labor Statistics, Report 1062, Labor Force Characteristics by Race and Ethnicity, 2015, at 6–7, 73 tbl.12, 76 tbl.12A (2016), https://www.bls.gov/opub/reports/race-and-ethnicity/2015/pdf/home.pdf [https://perma.cc/3M4Y-QULS].
See 2015 Job Patterns for Minorities and Women in Private Industry (EEO-1), EEOC, https://www1.eeoc.gov/eeoc/statistics/employment/jobpat-eeo1/2015/index.cfm#select_label [https://perma.cc/PF7Q-SV3R] (select “National Aggregate, All Industries” from the drop-down menu, then click “Go”) (last visited Jan. 26, 2019).
See, e.g., Marianne Bertrand & Sendhil Mullainathan, Are Emily and Greg More Employable than Lakisha and Jamal? A Field Experiment on Labor Market Discrimination, 94 Am. Econ. Rev. 991, 996–1006, 997 tbl.1, 1002 tbl.5, 1004 tbl.7 (2004) (conducting a “correspondence” audit study sending fictitious resumes to employers in Chicago and Boston and finding, inter alia, that white applicants received callbacks at 1.5 times the rate of black applicants, that black applicants received less benefit from better resume credentials compared to whites, that callback differences did not vary much by occupation or industry involved, and that neither self-declared “Equal Opportunity Employers” nor larger employers treated black applicants significantly better than other employers).
See, e.g., S. Michael Gaddis, Discrimination in the Credential Society: An Audit Study of Race and College Selectivity in the Labor Market, 93 Soc. Forces 1451, 1451, 1453, 1464–70 (2015) (conducting audit study to investigate role of college eliteness and racial discrimination in labor market outcomes for college graduates in three separate geographic regions and finding that “although a credential from an elite university results in more employer responses for all candidates, black candidates from elite universities only do as well as white candidates from less selective universities” and that “race results in a double penalty: When employers respond to black candidates, it is for jobs with lower starting salaries and lower prestige than those of white peers”); Devah Pager et al., Discrimination in a Low-Wage Labor Market: A Field Experiment, 74 Am. Soc. Rev. 777, 781–82, 784–86 (2009) (conducting in-person audit study using black, Latino, and white testers applying for low-wage entry level positions in New York, finding “clear racial hierarchy” so that “relative to equally qualified blacks, employers significantly prefer white and Latino job applicants” and “a black applicant has to search twice as long as an equally qualified white applicant before receiving a callback or job offer from an employer”).
Pager, supra note 148, at 782, 785.
See Donald Tomaskovic-Devey et al., Race and the Accumulation of Human Capital Across the Career: A Theoretical Model and Fixed-Effects Application, 111 Am. J. Soc. 58, 70, 82 (2005) (finding, in a large-scale longitudinal panel study, “substantial career differences in the earnings of black and white men” that “grow across the career” and are greater at higher levels of education).
Berrey et al., supra note 11, at 32 (discussing findings from Lincoln Quillian & Ole Hexel, Trends and Patterns in Racial Discrimination in Hiring in America, 1974-2015 (Apr. 1, 2016) (unpublished conference paper), https://paa.confex.com/paa/2016/mediafile/ExtendedAbstract/Paper6646/paa_2016_complete_submitted.pdf [https://perma.cc/XRT3-5ZCC]).
See, e.g., Tex. Dep’t of Cmty. Affairs v. Burdine, 450 U.S. 248, 256, 258–59 (1981) (plaintiff has the “ultimate burden of persuading the [factfinder] that she has been the victim of intentional discrimination”).
See Michael J. Zimmer, A Chain of Inferences Proving Discrimination, 79 U. Colo. L. Rev. 1243, 1243 (2008) (“There are three elements in a plaintiff’s prima facie case of individual disparate treatment discrimination: (1) the plaintiff suffered an adverse employment action, (2) the action was linked to the defendant, and (3) the defendant’s action was motivated by a protected characteristic of the plaintiff.”).
See Martin J. Katz, The Fundamental Incoherence of Title VII: Making Sense of Causation in Disparate Treatment Law, 94 Geo. L.J. 489, 515–16 (2006).
See id. at 515–16, 516 n.107.
Senator Case, a major proponent of the 1964 Civil Rights Act, pointedly noted this problem during the debates leading up to the passage of the law when he stated: “If anyone ever had an action that was motivated by a single cause, he is a different kind of animal from any I know of.” See Derum & Engle, supra note 19, at 1199 (citing 110 Cong. Rec. 13,837 (1964) (statement of Sen. Case)).
Plaintiffs can technically also prove unlawful disparate treatment through so-called “direct evidence” of discrimination, that is, evidence that so clearly and directly shows that an employment decision was made with discriminatory intent that the conclusion of discrimination does not require any inferences by the factfinder. See, e.g., Swierkiewicz v. Sorema N. A., 534 U.S. 506, 511–12 (2002). However, such evidence is not only highly unlikely to exist in most cases given the near impossibility of showing the mental state of an actor “directly” in most non-extreme scenarios, but courts have also been mired in endless disputes on what actually counts as “direct evidence” in the first place. See, e.g., Deborah C. Malamud, The Last Minuet: Disparate Treatment After Hicks, 93 Mich. L. Rev. 2229, 2321 n.290 (1995) (“[T]he various circuits have about as many definitions for direct evidence as they do employment discrimination cases.” (quoting Tyler v. Bethlehem Steel Corp., 958 F.2d 1176, 1183 (2d Cir. 1992))). Thus, I do not focus on this relatively uncommon theory.
411 U.S. 792, 802–06 (1973).
490 U.S. 228, 249 (1989).
See 42 U.S.C. § 2000e-2(m) (2012) (“[A]n unlawful employment practice is established when the complaining party demonstrates that race, color, religion, sex, or national origin was a motivating factor for any employment practice, even though other factors also motivated the practice.”).
Tex. Dep’t of Cmty. Affairs v. Burdine, 450 U.S. 248, 254 n.7 (1981).
McDonnell Douglas, 411 U.S. at 802–04. This burden-shifting structure is analyzed in much greater detail below.
See, e.g., McDonald v. Santa Fe Trail Transp. Co., 427 U.S. 273, 282 n.10 (1976).
See 42 U.S.C. § 2000e-5(g)(1) (2012) (“[T]he court may . . . order such affirmative action as may be appropriate, which may include, but is not limited to, reinstatement or hiring of employees, with or without back pay . . . or any other equitable relief as the court deems appropriate.”); see also id. § 1981a(a)(1) (“[T]he complaining party may recover compensatory and punitive damages . . . .”); Ponce v. Billington, 679 F.3d 840, 845 (D.C. Cir. 2012).
See 42 U.S.C. § 2000e-2(m).
See Ponce, 679 F.3d at 844.
See 42 U.S.C. § 2000e-5(g)(2)(B).
For example, until the Supreme Court decision in Desert Palace, Inc. v. Costa, 539 U.S. 90 (2003), most, but not all, courts followed Justice O’Connor’s concurrence in the splintered decision in Price Waterhouse, and held that a plaintiff could only use a mixed motives analysis upon a showing of direct evidence that a protected characteristic was a motivating factor in an employment decision. See William R. Corbett, Mike Zimmer, McDonnell Douglas and “A Gift That Keeps Giving,” 20 Emp. Rts. & Emp. Pol’y J. 303, 308–09 (2016). In Desert Palace, however, the Supreme Court held that the mixed motives analysis was available regardless of the type of evidence used by the plaintiff, and subsequently many scholars suggested that the McDonnell Douglas regime had thus become obsolete. See id. at 310; see generally Jamie Darin Prenkert, The Role of Second-Order Uniformity in Disparate Treatment Law: McDonnell Douglas’s Longevity and the Mixed-Motives Mess, 45 Am. Bus. L.J. 511, 511 n.2 (2008) (collecting sources making this argument). However, this prediction did not turn out to be true; courts continue to apply both frameworks, and it has not yet been resolved under which circumstances each type of analysis should apply. See Sandra F. Sperino, Beyond McDonnell Douglas, 34 Berkeley J. Emp. & Lab. L. 257, 260–61 (2013) (“Even though it has been more than 20 years since the 1991 amendments, there has been no satisfactory agreement regarding how the McDonnell Douglas test intersects with the mixed-motive rubric.”); Corbett, supra note 17, at 82 (“There are two proof structures under the disparate treatment theory of discrimination: 1) McDonnell Douglas or pretext; and 2) mixed motives. Among several problems with these proof structures, the most significant is that no one knows which one applies in any given case.”); see also Chadwick v. WellPoint, Inc., 561 F.3d 38, 45 & n.8 (1st Cir. 2009) (“The Desert Palace decision has proved ripe terrain for scholarly debate over how that decision interacts with the McDonnell Douglas framework. . . . Suffice it to say that the two decisions have not been definitively disentangled or reconciled . . . .” (citations omitted)).
I return to the mixed motives analysis in greater detail in Part IV below. See infra Part IV.
See supra note 169 and accompanying text; see also Sperino, supra note 169 at 257 (noting a “gradual weakening of the framework over the past two decades” and that “[r]ather than casting this test into oblivion, courts are slowly chipping away at its preeminent place as a proof structure”).
Prenkert, supra note 169, at 538; see also Judge David F. Hamilton, Address to the Association of American Law Schools, Section on Employment Discrimination Law 2013 Annual Meeting; On McDonnell Douglas and Convincing Mosaics: Toward More Flexible Methods Of Proof in Employment Discrimination Cases, 17 Emp. Rts. & Emp. Pol’y J. 195, 196 (2013) (“[I]n terms of practical effects, the last time I checked, a few weeks ago, McDonnell Douglas had been cited about 40,000 times and is cited hundreds more times every month.”); see, e.g., Chavez v. Credit Nation Auto Sales, LLC, 641 Fed. App’x 883, 885 (11th Cir. 2016) (“There is more than one way to show discriminatory intent using indirect or circumstantial evidence. One way is through the burden-shifting framework set out in McDonnell Douglas Corp. v. Green, and Tex. Dep’t of Cmty. Affairs v. Burdine . . . .” (citations omitted)). Anecdotally, multiple colleagues of mine who served as federal judicial clerks recalled applying only the McDonnell Douglas regime when reviewing employment discrimination cases that came before them.
See, e.g., Young v. United Parcel Serv., Inc., 135 S. Ct. 1338, 1345 (2015) (“We have also made clear that a plaintiff can prove disparate treatment either (1) by direct evidence that a workplace policy, practice, or decision relies expressly on a protected characteristic, or (2) by using the burden-shifting framework set forth in McDonnell Douglas.”); id. at 1353–55 (holding that a plaintiff may prove a disparate treatment claim asserting pregnancy-based discrimination using a modified McDonnell Douglas approach).
See, e.g., Ponce v. Billington, 679 F.3d 840, 845 (D.C. Cir. 2012). The court noted that they “expect that plaintiffs often will allege, in the alternative, that their cases are both [a pretext case or a mixed-motives case.]” Id. (quoting Price Waterhouse v. Hopkins, 490 U.S. 228, 247 (1989)).
See supra note 168 and accompanying text. As some courts have expressly noted, there is a risk for a mixed motives plaintiff that a jury in a close case might “split the baby” and find in favor of the plaintiff on liability but in favor of the defendant with respect to remedies under the same-decision defense, leaving the plaintiff without meaningful compensation. See, e.g., Coe v. N. Pipe Prods., Inc., 589 F. Supp. 2d 1055, 1097–98 (N.D. Iowa 2008); see also Zimmer, supra note 154, at 1293–94. While plaintiff-side attorneys may still have an incentive to bring mixed motives cases since attorney’s fees are available in such cases even if the defendant proves the same-decision defense, at least one commentator has noted that many courts have significantly restricted what they permit attorneys to recover in such cases. See Katz, supra note 155, at 534–36.
See Tex. Dep’t of Cmty. Affairs v. Burdine, 450 U.S. 248, 253–54 (1981).
McDonnell Douglas Corp. v. Green, 411 U.S. 792, 802 (1973).
Id. at 802 n.13.
See Lindemann et al., supra note 6, at 2–4 (“Courts adapt the [McDonnell Douglas] framework to conform to the facts of the particular case.”).
See, e.g., Flores v. Preferred Tech. Grp., 182 F.3d 512, 515 (7th Cir. 1999).
Burdine, 450 U.S. at 254.
Furnco Constr. Corp. v. Waters, 438 U.S. 567, 577 (1978). That this is the goal of the prima facie case is not entirely uncontested, and I will return to this point in greater detail below in Example 2, infra Section III.B.2.
Jackson v. Gonzales, 496 F.3d 703, 707 (D.C. Cir. 2007) (“The McDonnell Douglas framework first requires the plaintiff to establish a prima facie case of discrimination by showing that: (1) he is a member of a protected class . . . .”).
See supra note 153 and accompanying text.
See McDonald v. Santa Fe Trail Transp. Co., 427 U.S. 273, 279–80 (1976).
Burdine, 450 U.S. at 254.
See Angela Onwuachi-Willig, When Different Means the Same: Applying a Different Standard of Proof to White Plaintiffs Under the McDonnell Douglas Prima Facie Case Test, 50 Case W. Res. L. Rev. 53, 79 (1999) (discussing research on widespread racial discrimination in the workplace to support the “argument that it is unreasonable to infer discrimination for white plaintiffs based solely upon the last three prongs of the McDonnell Douglas prima facie case test” and that “[b]ecause white job applicants are generally favored by employers, it makes no sense for courts to infer discrimination for white applicants absent a showing that their employer is that unusual employer who discriminates against the majority”); see also Zatz, supra note 130 and accompanying text (discussing various possible meanings of “special treatment”).
The Supreme Court has not spoken on this issue.
These circuits seem to trace back to the D.C. Circuit’s decision in Parker v. Baltimore & Ohio R. Co., 652 F.2d 1012, 1017 (D.C. Cir. 1981). For examples of circuits that cited and accepted Parker, see Murray v. Thistledown Racing Club, Inc., 770 F.2d 63, 67 (6th Cir. 1985), and Christensen v. Equitable Life Assurance Soc’y of the U.S., 767 F.2d 340, 343 (7th Cir. 1985).
Parker, 652 F.2d at 1017; see also Charles A. Sullivan, Circling Back to the Obvious: The Convergence of Traditional and Reverse Discrimination in Title VII Proof, 46 Wm. & Mary L. Rev. 1031, 1058–59 (2004).
See, e.g., Phelan v. City of Chicago, 347 F.3d 679, 684–85 & n.1 (7th Cir. 2003) (“[Plaintiff] argues that this new and different prong imposes an unfair burden to non-minority plaintiffs since a non-minority plaintiff must establish more facts to create the prima facie case. . . . However, this court has recognized that discrimination by employers against white men is a less common phenomenon than discrimination against minorities. For that reason, in order to gain the substantial benefits conferred by the use of the McDonnell Douglas test, [i.e., an “inference of discrimination . . . in the absence of direct evidence of discrimination”] the non-minority plaintiff must be able to plead facts to show why it is likely in this case, that an employer had engaged in such unusual behavior.” (citations omitted)); Harding v. Gray, 9 F.3d 150, 153 (D.C. Cir. 1993) (“Invidious racial discrimination against whites is relatively uncommon in our society, and so there is nothing inherently suspicious in an employer’s decision to promote a qualified minority applicant instead of a qualified white applicant. . . . [T]he background circumstances requirement merely substitutes for the minority plaintiff’s burden to show that he is a member of a racial minority; both are criteria for determining when the employer’s conduct raises an ‘inference of discrimination.’” (quoting Bishopp v. District of Columbia, 788 F.2d 781, 786 (D.C. Cir. 1986))).
See, e.g., supra notes 144–50 and accompanying text; Onwuachi-Willig, supra note 187, at 71–80.
See, e.g., Iadimarco v. Runyon, 190 F.3d 151, 160–63 (3d Cir. 1999) (noting that the proper standard is a more generic inquiry requiring “the plaintiff to present sufficient evidence to allow a fact finder to conclude that the employer is treating some people less favorably than others based upon a trait that is protected under Title VII” and rejecting the background circumstances test as among other things, inappropriately heightening the burden imposed on white plaintiffs and being “irremediably vague and ill-defined”).
644 F.3d 1321 (11th Cir. 2011).
Id. at 1325 (emphasis added).
Id. at 1325 n.15 (quoting Bass v. Bd. of Cty. Comm’rs, 256 F.3d 1095, 1102–03 (11th Cir. 2001)).
See Zatz, supra note 130, at 1159.
See id. at 1157.
For example, in an unpublished decision, the Sixth Circuit suggested that what one would consider “background circumstances” may well have changed over time so that this showing might have become easier to make compared to the early 1980s, when the test first appeared. See Smith v. City of Dayton, No. 93-3639, 1994 WL 540666, at *4 (6th Cir. Oct. 3, 1994) (noting that “relevant facts may include whether the employer has been under external or internal pressure-for example, from affirmative action goals-to hire more minorities,” that while such “actions are of course not illegitimate in themselves,  they may well indicate that the employer’s situation is different from the ‘background circumstances’ existing at the time of Parker,” and “[t]hus, employment actions affecting a member of the ‘majority’ that may not have created a prima facie Title VII case under Parker might, depending on the circumstances with which the employer is presented, create such a case today”).
See, e.g., Zambetti v. Cuyahoga Cmty. Coll., 314 F.3d 249, 257 (6th Cir. 2002) (applying the standard as circuit law but suggesting that it “may impermissibly impose a heightened pleading standard on majority victims of discrimination” and noting “serious misgivings about the soundness of a test which imposes a more onerous standard for plaintiffs who are white or male” (quoting Pierce v. Commonwealth Life Ins. Co., 40 F.3d 796, 801 n.7 (6th Cir. 1994))).
Kteily et al., supra note 99, at 142.
Id. at 152.
Id. at 153.
See Eibach & Keegan, supra note 102, at 459.
Id. at 459–60.
Id. at 461–62.
See Michael I. Norton & Samuel R. Sommers, Whites See Racism as a Zero-Sum Game That They Are Now Losing, 6 Persp. on Psychol. Sci. 215, 217 (2011).
Id. at 216. Indeed, “[b]y the 2000s, some 11% of Whites gave anti-White bias the maximum rating on our [10-point] scale in comparison with only 2% of Whites who did so for anti-Black bias.” Id. Norton & Sommers also found that whites appear to take a strong zero-sum view of racial bias, so that “within each decade and across time, White respondents were more likely to see decreases in bias against Blacks as related to increases in bias against Whites . . . whereas Blacks were less likely to see the two as linked.” Id. at 217.
O’Connor v. Consol. Coin Caterers Corp., 517 U.S. 308, 312 (1996) (quoting Teamsters v. United States, 431 U.S. 324, 358 (1977)).
See supra note 193.
See supra note 196 and accompanying text.
557 U.S. 557 (2009).
See id. at 563–75.
Id. at 579, 592. In other words, the Court viewed the facts as establishing something akin to a prima facie case, though the majority did not expressly phrase it as such.
Id. at 579–80. The City could defend such race-based action only if it could show “a strong basis in evidence” that it would have been liable for disparate impact had it not certified the results. Id. at 584.
See Cheryl I. Harris & Kimberly West-Faulcon, Reading Ricci: Whitening Discrimination, Racing Test Fairness, 58 UCLA L. Rev. 73, 106–07 (2010).
See supra notes 205–07 and accompanying text.
Harris & West-Faulcon, supra note 217, at 107.
Id. at 108.
Id. at 108–09.
Onwuachi-Willig, supra note 187, at 81–83.
See supra notes 162–63 and accompanying text.
Tex. Dep’t of Cmty. Affairs v. Burdine, 450 U.S. 248, 254 (1981).
Id. at 255 n.10.
See Reeves v. Sanderson Plumbing Prods., Inc., 530 U.S. 133, 148 (2000).
See id.; see also St. Mary’s Honor Ctr. v. Hicks, 509 U.S. 502, 511 (1993).
Reeves, 530 U.S. at 146–47 (“The ultimate question is whether the employer intentionally discriminated, and proof that the employer’s proffered reason is unpersuasive, or even obviously contrived, does not necessarily establish that the plaintiff’s proffered reason . . . is correct.”); Hicks, 509 U.S. at 515–16. As the Court has put it, “[t]he ultimate burden of persuading the trier of fact that the defendant intentionally discriminated against the plaintiff remains at all times with the plaintiff.” Burdine, 450 U.S. at 253.
See supra note 13 and accompanying text. As shown in Example 3 below, the doctrine that applies to cases prototypically brought by white plaintiffs—challenges to affirmative action programs—is much more plaintiff-friendly.
Some scholars have argued that Title VII’s legislative history reveals a strong concern with limiting Title VII’s interference with employer decision-making—that the statute’s purpose “was to balance the prohibition of the most obvious forms of discrimination with the preservation of as much employer decision-making latitude as possible.” Chuck Henson, Title VII Works — That’s Why We Don’t Like It, 2 U. Miami Race & Soc. Just. L. Rev. 41, 42, 53 (2012); see also Daniel B. Rodriguez & Barry R. Weingast, The Positive Political Theory of Legislative History: New Perspectives on the 1964 Civil Rights Act and Its Interpretation, 151 U. Pa. L. Rev. 1417, 1490–97 (2003) (discussing various changes made to the Act while under consideration in Congress. which blunted its impact). Thus, one may argue that the Supreme Court’s decisions on the employer’s burden in the McDonnell Douglas framework merely reflect this purpose by allowing employers to make decisions based on any reason so long as it is not a discriminatory one. Still, this begs the question of how the doctrine should determine whether unlawful discrimination was, in fact, not the reason for a challenged decision when it has been shown that the employer has not offered a credible explanation in response to a successful prima facie case. A general statutory commitment to management prerogatives does not necessarily predetermine the answer to this question.
McDonnell Douglas Corp. v. Green, 411 U.S. 792, 804 (1973).
For example, the Court described the question as one of presenting “evidence that the presumptively valid reasons for [the employee’s] rejection were in fact a coverup for a racially discriminatory decision,” id. at 805; of “competent evidence that whatever the stated reasons for [the employee’s] rejection, the decision was in reality racially premised,” id. at 805 n.18; and of showing that the employer’s “assigned reason for refusing to re-employ was a pretext or discriminatory in its application,” id. at 807.
450 U.S. 248, 256 (1981).
Id. (quoting McDonnell Douglas, 411 U.S. at 804–05).
460 U.S. 711, 716 (1983) (quoting Burdine, 450 U.S. at 256).
Id. at 713–14 (footnote omitted).
Id. at 715 (quoting Burdine, 450 U.S. at 253).
Id. at 717–18 (Blackmun, J., concurring) (joined by Brennan, J.).
509 U.S. 502, 511 (1993).
Id. at 510–11. As the majority noted,
[t]o resurrect it later, after the trier of fact has determined that what was ‘produced’ to meet the burden of production is not credible, flies in the face of our holding in Burdine that to rebut the presumption ‘[t]he defendant need not persuade the court that it was actually motivated by the proffered reasons.’
Id. at 510 (quoting Burdine, 450 U.S. at 254).
530 U.S. 133, 146–47 (2000). Notably, the decision in Reeves (a case under the ADEA but which applied the McDonnell Douglas framework without modification) was a unanimous decision with only a short concurrence by Justice Ginsburg. Ginsburg’s concurrence noted only that she expected that there would be few cases in which a prima facie case combined with discrediting the employer’s offered reason would not be sufficient to find liability, and that the Court should “define more precisely the circumstances in which plaintiffs [would] be required to submit evidence beyond these two categories.” See id. at 154–55 (Ginsburg, J., concurring). While Hicks had been a deeply contested 5-4 decision, in Reeves no other Justice was willing to continue to disagree with the conclusion reached in Hicks.
Hicks, 509 U.S. at 521.
Furnco Constr. Corp. v. Waters, 438 U.S. 567, 577 (1978).
Hicks, 509 U.S. at 510–11.
See supra notes 144–52 and accompanying text.
In Hicks, for example, the employer had argued that their reasons for demoting and discharging the plaintiff were plaintiff’s “severity and . . . accumulation of [workplace] rules violations.” Hicks, 509 U.S. at 507. The trial court “found that the reasons [the employer] gave were not the real reasons for [plaintiff’s] demotion and discharge.” Id. at 508. Nevertheless, the court ruled for the employer because it thought that the plaintiff had not proven that the real reason was racial, rather than personal, animosity. Id. at 508.
Id. at 536 (Souter, J., dissenting) (emphasis added) (citing Furnco, 438 U.S. at 577).
If courts were to universally apply a robust background circumstances requirement for white plaintiffs, as discussed in Example 1 above, such a substantive understanding of the prima facie case would also be appropriate for white plaintiffs who can meet such a requirement. Such plaintiffs will have made a strong showing that the otherwise unusual inference of discrimination against a worker because she is white was appropriate in an individual case. As should be apparent from the discussion in Section III.B.3 below, however, employer affirmative action programs should not support a background circumstances finding.
See Malamud, supra note 158, at 2262 n.110.
See supra notes 144–50 and accompanying text.
See supra notes 201–09 and accompanying text. See also Derum & Engle, supra note 19, at 1220 (“After Hicks, courts are no longer willing to assume that, absent an explanation, discrimination motivated the defendant employer. In McDonnell Douglas, Furnco, and Burdine, the Court was willing to risk that employers would be found liable for discrimination even when the employers really believed there was no discrimination, and when there was only circumstantial evidence to support the claim. It treated the statute as a deterrent mechanism—as a tool to make employers more deliberate and less discriminatory. In contrast, the majority in Hicks was willing to risk that employees who suffered discrimination would find no relief.”).
As Justice O’Connor explained, the limited burden assigned to the employer in cases proceeding under the McDonnell Douglas framework reflects, at least in part, the “presumption of good faith concerning its employment decisions which is accorded employers facing only circumstantial evidence of discrimination.” Price Waterhouse v. Hopkins, 490 U.S. 228, 261–66 (1989) (O’Connor, J., concurring).
See Derum & Engle, supra note 19, at 1224–28.
For a more in-depth discussion, see, for example, Natasha T. Martin, Pretext in Peril, 75 Mo. L. Rev. 313 (2010).
See, e.g., Ann C. McGinley, The Emerging Cronyism Defense and Affirmative Action: A Critical Perspective on the Distinction Between Colorblind and Race-Conscious Decision Making Under Title VII, 39 Ariz. L. Rev. 1003, 1005, 1017–18, 1021–22 (1997).
See, e.g., Kerri Lynn Stone, Taking in Strays: A Critique of the Stray Comment Doctrine in Employment Discrimination Law, 77 Mo. L. Rev. 149, 149−52 (2012).
See, e.g., Quintanilla & Kaiser, supra note 19, at 5–6; Martin, supra note 254, at 359–65.
See, e.g., Flores v. Preferred Tech. Grp., 182 F.3d 512, 516 (7th Cir. 1999).
Jackson v. Gonzales, 496 F.3d 703, 708 (D.C. Cir. 2007).
By “affirmative action” programs, I am referring to programs in which employers overtly take race into account when making decisions about which employees to hire, promote, lay off, admit to training, etc.—usually to increase job opportunities and benefits for workers of color. I focus on Title VII doctrine regulating when employers can voluntarily adopt such measures. This is partly because such plans are ostensibly subject to the same doctrinal analytical structure as “traditional” disparate treatment cases, yet, as shown in this example, this analytical structure is applied differently between the two scenarios. Thus, the comparison of affirmative action cases and “traditional” disparate treatment cases further illustrates the patterns of asymmetry disadvantaging workers of color that are discussed in this Part. This limitation also recognizes that doctrinal limits for when courts can order affirmative action remedies, for example, Firefighters Local Union No. 1784 v. Stotts, 467 U.S. 561, 572–74 (1984); Local 28 of the Sheet Metal Workers’ International Association v. EEOC, 478 U.S. 421, 470–71 (1986); and limits set by the Constitution for nonjudicial state actors, for example, Wygant v. Jackson Bd. of Educ., 476 U.S. 267, 277–78 (1986), are complex and deserve separate treatment. I discuss these cases, and race-conscious remedies such as affirmative action more broadly, in other work. See generally David Simson, Whiteness as Innocence, 96 Denv. L. Rev. (forthcoming 2019), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3173153 [https://perma.cc/HDG7-W6X3]. Notably, though, scholars have made arguments about the historical development of constitutional limits on race-based affirmative action (including outside of the employment context) that can be viewed as similar to the asymmetry framework I discuss in this Article. See generally Ian Haney-López, Intentional Blindness, 87 N.Y.U. L. Rev. 1779 (2012).
It is possible, of course, for workers of color to challenge affirmative action programs as well. Generally, however, such cases are brought by white plaintiffs and affirmative action is publicly (and doctrinally) debated as predominantly affecting the rights of whites. Thus, challenges by white workers are my focus in this Example.
See Johnson v. Transp. Agency, 480 U.S. 616, 626–27 (1987). While Johnson was a case in which an affirmative action program with respect to sex was at issue, the Court made clear that the same standards could be applied to programs involving race. See id. at 635 n.13; see also, e.g., Schurr v. Resorts Int’l Hotel, Inc., 196 F.3d 486, 497 n.11 (3d Cir. 1999).
Johnson, 480 U.S. at 626.
The word “pretext” is somewhat of a misnomer in this context. In “traditional” cases, as described above, the pretext inquiry is indeed focused on determining whether an employer is pretending to have made a decision for a nonracial reason when the “real reason” for their action was racial discrimination. Whether the asserted nonracial reason is a “good” or savory one is largely beside the point as long as it was the actual reason. In affirmative action cases, the “pretext” question is not about pretending. All sides may well agree that the relevant decision was made pursuant to an affirmative action program—though as described below, that can be a risky concession for an employer. Rather, the question is whether the affirmative action program fits within certain doctrinal limits. In other words, the “pretext” inquiry is focused on determining the factual existence (or not) of a particular reason in “traditional” cases, whereas in affirmative action cases it is focused on evaluating the quality and propriety of the reason (i.e., whether the affirmative action program fits within the doctrinal boundaries for such programs).
Johnson, 480 U.S. at 626. The text of Title VII does not speak to the permissible boundaries of affirmative action programs, which have instead developed through judicial interpretation. See id. at 627–31; United Steelworkers v. Weber, 443 U.S. 193, 197, 201–02 (1979). The Civil Rights Act of 1991, which amended Title VII, did include a section, however, which recognized (and arguably approved) the existence of such programs within judicial guidelines by providing that “[n]othing in the amendments made by this title shall be construed to affect court-ordered remedies, affirmative action, or conciliation agreements, that are in accordance with the law.” Civil Rights Act of 1991, Pub. L. 102-166, § 116, 105 Stat. 1071, 1079 (codified as amended at 42 U.S.C. § 1981 note).
Johnson, 480 U.S. at 631 (quoting Weber, 443 U.S. at 197); see also Shea v. Kerry, 796 F.3d 42, 57 (D.C. Cir. 2015) (quoting Johnson, 480 U.S. at 631). This formulation focused on underrepresentation is particularly relevant for employment decisions involving hiring, promotion, and admission to training programs. Courts have, at times, dealt with alleged affirmative action programs in areas such as pay equality and rephrased the standard accordingly. See, e.g., Smith v. Va. Commonwealth Univ., 84 F.3d 672, 676 (4th Cir. 1996) (en banc) (asking whether plan was “designed to eliminate a manifest racial or sexual imbalance” in pay).
Johnson, 480 U.S. at 637.
Id. at 639–40.
This requirement is similar to, though slightly more forgiving than, the factual showing necessary to establish a prima facie case in a disparate treatment case involving a “pattern or practice” of illegal discrimination, which generally requires either “gross” statistical disparities, see Hazelwood Sch. Dist. v. United States, 433 U.S. 299, 307–08 (1977), or significant statistical disparities along with anecdotal evidence of individual instances of discrimination, see International Brotherhood of Teamsters v. United States, 431 U.S. 324, 337–43 (1977). See Johnson, 480 U.S. at 632–33 (“A manifest imbalance need not be such that it would support a prima facie case against the employer . . . .”). For example, where there are substantial, but perhaps not “gross” disparities, an employer implementing an affirmative action program may not have to provide additional “anecdotal evidence” of discrimination, when such evidence may be necessary to prove a pattern or practice case. See id. at 633 n.11.
The relevant local labor force could either be the entire labor market or general population of the area for jobs that don’t require specific expertise or for admission to a training program, or it could be the labor force with the requisite qualifications for positions requiring specific skills. See Johnson, 480 U.S. at 631–32.
See, e.g., Hazelwood, 433 U.S. at 313 (remanding pattern or practice case based on questions regarding proper comparator labor force, the choice of which might determine whether statistical underrepresentation was sufficient to support plaintiff’s case).
See, e.g., Smith v. Va. Commonwealth Univ., 84 F.3d 672, 676–77 (4th Cir. 1996) (en banc) (reversing summary judgment for employer because of questions of fact regarding whether employer’s multiple regression analysis used to establish “manifest imbalance” improperly excluded factors that may have contributed to existing pay disparities).
Johnson, 480 U.S. at 632.
See Sullivan, supra note 190, at 1049 n.79.
See, e.g., Schurr v. Resorts Int’l Hotel, Inc., 196 F.3d 486, 497 & n.11 (3d Cir. 1999).
The most famous case rejecting a “diversity rationale” for affirmative action in the workplace is Taxman v. Board of Education, 91 F.3d 1547, 1563 (3d Cir. 1996) (“While the benefits flowing from diversity in the educational context are significant indeed, we are constrained to hold, as did the district court, that inasmuch as ‘the Board does not even attempt to show that its affirmative action plan was adopted to remedy past discrimination or as the result of a manifest imbalance in the employment of minorities,’ the Board has failed to satisfy the first prong of the Weber test.” (citation omitted)). Developments in constitutional law, where the educational benefits of a racially diverse student body can constitute a compelling interest for affirmative action programs in university admissions under the equal protection clause, have so far not translated to Title VII, and at least some scholars expect this to continue for the foreseeable future. See generally Deborah C. Malamud, The Strange Persistence of Affirmative Action Under Title VII, 118 W. Va. L. Rev. 1 (2015) (discussing the likelihood of the Court extending diversity-based affirmative action to the Title VII context).
See, e.g., Schurr, 196 F.3d at 488–89, 497–98 (rejecting affirmative action plan by private employer adopted in response to regulations by the state gaming regulatory authority because not remedial but instead adopted in recognition of the fact that the legislature “was also aware Atlantic City had and has a large minority population, and sought to ensure that the job creation which would accompany casino developments would benefit all segments of the population”).
Johnson v. Transp. Agency, 480 U.S. 616, 638 (1987); see also United Steelworkers v. Weber, 443 U.S. 193, 208–09 (1979).
See Johnson, 480 U.S. at 639; see also Weber, 443 U.S. at 208.
Cunico v. Pueblo Sch. Dist. No. 60, 917 F.2d 431, 440–43 (10th Cir. 1990).
See id. at 440.
See, e.g., Humphries v. Pulaski Cty. Special Sch. Dist., 580 F.3d 688, 694 (8th Cir. 2009); Bass v. Bd. of Cty. Comm’rs, 256 F.3d 1095, 1110–11 (11th Cir. 2001). The similarity of these rulings to Ricci is striking. See supra notes 213–17 and accompanying text. There, too, race-conscious remedial action was interpreted as presumptive disparate treatment that can only be justified by meeting stringent standards of legal justification.
See, e.g., Weber, 443 U.S. at 206–08 (grounding the argument that Title VII’s legislative history supports permissibility of private voluntary affirmative action in legislators’ goal to preserve management prerogatives).
As Deborah Malamud notes, there is reason to think that it was a distinctly pro-business interpretation of Title VII that secured a full five-vote majority in favor of affirmative action. See Deborah C. Malamud, The Story of United Steelworkers v. Weber, in Employment Discrimination Stories 173, 207–13 (Joel Wm. Friedman ed., 2006).
See Nielsen et al., supra note 13, at 18–22.
See supra notes 7–13 and accompanying text.
See supra Section III.B.1.
See supra Section III.B.2.
See supra notes 255–59 and accompanying text.
See supra notes 271–79 and accompanying text.
See supra note 267.
See supra notes 280–84 and accompanying text.
See supra note 285 and accompanying text.
This view is widespread, accepted even by many who (perhaps begrudgingly) accept that due to past abuses in the treatment of racial minorities in the United States, some remedial action is warranted—but only as an exception, and only temporarily. See, e.g., Ian Haney López, "A Nation of Minorities": Race, Ethnicity, and Reactionary Colorblindness, 59 Stan. L. Rev. 985, 1051–61 (2007). Indeed, even the case that established the permissibility of affirmative action programs under Title VII, Weber framed the issue around whether Title VII should be interpreted to allow “racial preferences” under certain circumstances. United Steelworkers v. Weber, 443 U.S. 193, 200 (1979).
As I have noted above, this understanding does not reflect the reality of the American workplace. See supra notes 144–50 and accompanying text.
See supra notes 201–07 and accompanying text.
See supra Section II.A.4.
See Unzueta et al., supra note 103.
See id. app. A (describing the vignettes; in one vignette, a white officer applied for a promotion for which a black officer was later selected, in the other vignette, all the details were kept the same, except a white officer was promoted over a black officer).
Id. at 83.
Id. There was a trend (though not statistically significant) for participants to like the white officer who claimed discrimination more the higher their SDO level. Id.
Id. at 84–85. The participants in the study self-identified their race as follows: Native American: 1; White: 109; Black: 92; Latino: 5; Asian: 20; and multiracial participants: 3. Id. at 84.
Id. at 85. In this study, participants liked the white discrimination claimant significantly more the higher their SDO level. Id. at 84–85. The white applicant’s hirability was unaffected by a prior discrimination claim. Id. at 85. The race of the participant did not change this effect of SDO on likability and hirability. Id. at 84.
Id. at 87.
St. Mary’s Honor Ctr. v. Hicks, 509 U.S. 502, 508 (1993).
See, e.g., Johnson v. Transp. Agency, 480 U.S. 616, 675–77 (1987) (Scalia, J., dissenting) (noting that as a result of the majority’s decision permitting certain types of affirmative action, a “statute [(Title VII)] designed to establish a color-blind and gender-blind workplace has thus been converted into a powerful engine of racism and sexism, not merely permitting intentional race- and sex-based discrimination, but often making it, through operation of the legal system, practically compelled”); cf. Adarand Constructors, Inc. v. Pena, 515 U.S. 200, 240–41 (1995) (Thomas, J., concurring in part and concurring in the judgment) (“[T]here is a moral [and] constitutional equivalence between laws designed to subjugate a race and those that distribute benefits on the basis of race in order to foster some current notion of equality.” (alteration in original) (citation omitted)); Bass v. Bd. of Cty. Comm’rs, 256 F.3d 1095, 1111 n.7 (11th Cir. 2001) (“The only thing to distinguish an affirmative action plan from any other discriminatory statement (other than the degree of formality involved) is that the discrimination it describes or prescribes is permissible if the plan is valid under Title VII and the Equal Protection Clause. If it is not valid, an affirmative action plan amounts to nothing more than a formal policy of unlawful discrimination.”).
See supra notes 109–14 and accompanying text.
See Regents of Univ. of Cal. v. Bakke, 438 U.S. 265, 407 (1978) (Blackmun, J.) (“I suspect that it would be impossible to arrange an affirmative-action program in a racially neutral way and have it successful. To ask that this be so is to demand the impossible. In order to get beyond racism, we must first take account of race. There is no other way. And in order to treat some persons equally, we must treat them differently. We cannot—we dare not—let the Equal Protection Clause perpetuate racial supremacy.”).
See Schoenbaum, supra note 28, at 135–36 (arguing that an asymmetry in the standards of proof may unfairly limit suits).
The necessity of a universal background circumstances requirement may not be as strong in the context of other protected characteristics. While this topic is beyond the scope of this Article, Schoenbaum raises this possibility for age and certain sex discrimination contexts. See id. at 136–37. I argue that at least in the context of race, for which Schoenbaum acknowledges that some of the proposed benefits of greater symmetry are weaker, pushing for such a requirement is clearly appropriate. See id. at 123.
Given space constraints, I do not address arguments for doctrinal changes that would make it more difficult for all plaintiffs to win their cases. Given the prominence of plaintiffs of color in the disparate treatment context, such a requirement would be counter to the framework proposed in this Article, which is concerned with inappropriate burdens placed on workers of color by the status quo.
See, e.g., Katz, supra note 17, at 651–59; Katz, supra note 155, at 515–16, 531 n.147; Corbett, supra note 17, at 107–08. Recall that under McDonnell Douglas, a plaintiff must make the stricter showing of but-for causation. See generally supra notes 164–68 and accompanying text.
As Professor Katz has noted, such overdetermined cases are likely to be more common in employment discrimination cases than in other areas of the law. See Katz, supra note 155, at 512–14.
Price Waterhouse v. Hopkins, 490 U.S. 228, 242 (1989).
Id. at 239 n.3.
Id. at 279 (O’Connor, J., concurring); id. at 293 n.4 (Kennedy, J., dissenting).
The 1991 Civil Rights Act also punted on answering this ideological problem by codifying motivating factor causation but noting that affirmative action plans “in accordance with the law” are permissible. See Civil Rights Act of 1991, Pub. L. 102-166, § 116, 105 Stat. 1071, 1079 (codified as amended at 42 U.S.C. § 1981 note).
While there might be the same-decision defense to limit remedies, establishing that defense necessitates showing the pointlessness of the program by proving that the same decision could, and would, have been made without it. See 42 U.S.C. § 2000e–5(g)(2)(B).
See supra notes 109–14, 313–14 and accompanying text.
Contra Corbett, supra note 17, at 108–09.
See generally Geoffrey C. Ho & Miguel M. Unzueta, Antiegalitarians for Affirmative Action? When Social Dominance Orientation is Positively Related to Support for Egalitarian Social Policies, 45 J. Appl. Soc. Psychol. 451 (2015) (explaining how different types of affirmative action programs weigh group membership differently in selection decisions and how this weighing may affect judgments of people with varying SDO).
See, e.g., Harris, supra note 39.
See, e.g., Sandra F. Sperino, Rethinking Discrimination Law, 110 Mich. L. Rev. 69, 115–21 (2011); Timothy M. Tymkovich, The Problem with Pretext, 85 Denv. L. Rev. 503, 528–29 (2008).
See, e.g., Jerry Kang et al., Implicit Bias in the Courtroom, 59 UCLA L. Rev. 1124, 1142–47 (2012).
See supra Section II.A.4.
See supra Section III.B.3.
Cf. Malamud, supra note 158, at 2272–74 (contemplating the possibility that the McDonnell Douglas framework may function to constrain hostile factfinders, but ultimately rejecting that this should be its function).
See supra Section II.A.3.
See, e.g., Jim Sidanius et al., Social Dominance Orientation, Hierarchy Attenuators and Hierarchy Enhancers: Social Dominance Theory and the Criminal Justice System, 24 J. Appl. Soc. Psychol. 338, 344–49 (1994).
See, e.g., All. for Justice, Broadening the Bench: Professional Diversity and Judicial Nominations 4 (2016) (“As this report details, the federal judiciary is currently lacking in judges with experience (a) working for public interest organizations; (b) as public defenders or indigent criminal defense attorneys; and (c) representing individual clients—like employees, consumers, or personal injury plaintiffs—in private practice.”), https://www.afj.org/wp-content/uploads/2014/11/Professional-Diversity-Report.pdf [https://perma.cc/7WS3-2TXN].
See, e.g., Weinberg & Nielsen, supra note 12, at 347–48; Chew & Kelley, supra note 33, at 113–14.
See supra note 59 and accompanying text.
See, e.g., supra note 126.
For a review, see Pat K. Chew, Judges’ Gender and Employment Discrimination Cases: Emerging Evidence-Based Empirical Conclusions, 14 J. Gender Race & Just. 359, 366 (2011).
See, e.g., Boyd, supra note 126, at 796 n.3; see also Jonathan P. Kastellec, Racial Diversity and Judicial Influence on Appellate Courts, 57 Am. J. Pol. Sci. 167, 171 (2013).
For example, the intersectional effects of race and gender are comparatively underexplored both in empirical research on judicial decision-making, see for example, Boyd, supra note 126, at 795 (noting project’s “inability to speak directly to the important issue of intersectionality”); and in Social Dominance Theory, see for example, Pratto, supra note 49, at 310 (“We need to know more about the relationships among age, gender, and arbitrary-set inequality . . . .”); Lee, supra note 59, at 1047 (noting difficulty to investigate gender-arbitrary set interaction based on existing research into SDO levels). Such research is important in trying to better understand and explain research findings in some studies showing that black judges are significantly more likely to find for plaintiffs in both race and sex discrimination cases, while female judges were only significantly more likely to find for plaintiffs in sex discrimination cases. See Boyd, supra note 126, at 793–94.