To appreciate the significance of Working Law, it is helpful to revisit the now infamous Wal-Mart v. Dukes Supreme Court decision, which upended how we understand organizational compliance with civil rights law. In Wal-Mart, the plaintiffs challenged systemic discrimination flowing from a policy that gave local supervisors subjective discretion over pay and promotion.Footnote 1 There was evidence that managers relied on gender stereotypes when making these decisions, including telling women who were denied promotion “it is a man’s world,” that the manager “did not want women,” and that “men are here to make a career and women aren’t.” The evidence presented by the plaintiffs showed that at Wal-Mart, women held 70 percent of hourly jobs but only 33 percent of management jobs, and their statistical analysis showed that these differences were not explained by factors other than gender (Wal-Mart v. Dukes 2011).
Despite this evidence, the Court rejected the systemic discrimination claim, holding that there was “no significant proof that Wal-Mart operated under a general policy of discrimination” because “Wal-Mart’s announced policy forbids sex discrimination.” (Wal-Mart v. Dukes 2011, 354). Systemic disparate treatment claims require evidence of discriminatory intent, and the Court apparently treated Wal-Mart’s paper policy as conclusive evidence of the absence of that intent. Nevertheless, an alternative disparate impact claim still seemed viable because the Court had previously held that “an employer’s undisciplined system of subjective decision making [can have] precisely the same effects as a system pervaded by impermissible intentional discrimination” and therefore could subject employers to Title VII liability (Watson v. Fort Worth Bank 1988, 990–91; Civil Rights Act of 1991). But the Wal-Mart Court dismissed the class-based disparate impact claim as well, stating, without citation to any authority, that “left to their own devices most managers in any corporation—and surely most managers in a corporation that forbids sex discrimination—would select sex-neutral performance-based criteria for hiring and promotion” (Wal-Mart v. Dukes 2011, 355, emphasis added).
For many employment discrimination scholars, Wal-Mart was a bit of a “what the heck happened” moment. The Court brushed aside evidence of unequal treatment that had been sufficient to establish discrimination in past systemic discrimination and disparate impact cases, and it accepted without question a corporate antidiscrimination policy as conclusive evidence of the absence of systemic discrimination. For the Wal-Mart majority, it was unimaginable that individual managers in an organization with a paper antidiscrimination policy would engage in systemic sex discrimination in hiring and promotion. Moreover, the Court viewed a policy of standardless subjective decision making as unproblematic, in its words, “a very common and presumptively reasonable way of doing business” (Wal-Mart v. Dukes 2011, 354). This reasoning evidenced a very naïve view of the operations of the firm that saw no potential conflict between the will of the organization as expressed in its policy and the agency of managers to implement those policies to serve their own interests (Reskin Reference Reskin1988; Tomaskovic-Devey Reference Tomaskovic-Devey1993; Nelson and Bridges Reference Nelson and Bridges1999). It also raised an important question about the meaning and reach of antidiscrimination law: when employers delegate subjective decision-making authority to low-level managers, what responsibility should employers bear when gender stereotypes shape the decisions of those managers? Can an employer rely on a paper antidiscrimination policy to absolve it of responsibility for widespread and documented disparate outcomes facilitated by the employer’s institutional policies favoring subjective decision making (Hart Reference Hart2006)? In Wal-Mart, the Court comes down firmly on the side of individual agency rather than organizational responsibility: employers are not responsible for the systemic effects of managers’ biased decisions if antidiscrimination policies are in place, even if evidence of disparate outcomes is abundantly available. Prior precedent had held that discriminatory intent could be inferred from evidence of disparate outcomes coupled with anecdotal evidence of discriminatory bias (Teamsters v. United States 1977; Hazelwood School District v. United States 1977). In Wal-Mart, however, a single policy statement was enough to counter that inference and absolve the employer of responsibility for systemic discrimination.
ENDOGENEITY THEORY
Working Law provides a persuasive explanation for how, more than fifty years after the passage of the Civil Rights Act of 1964, we could find ourselves in such a stunning moment of triumph of form over substance. Lauren Edelman argues that in response to civil rights laws, employers create policies and programs that promise equal opportunity, yet often simultaneously maintain practices that perpetuate the advantages of whites and males. Over time, she argues, organizational policies that symbolize attention to civil rights laws have become widely accepted as evidence of actual compliance with those laws, whether or not those policies are effective. In other words, policies against discrimination like the one in Wal-Mart are no longer a means to comply with civil rights laws; now they define what compliance means.
Edelman supports her argument empirically not by identifying a vast employer conspiracy to evade civil rights protections, but by drawing on neoinstitutional organizational theory to show the mechanisms through which, over time, organizational structures come to confer legitimacy. More specifically, she examines how legitimacy and meaning drive organizational decision making and encourage organizations to produce symbolic structures in response to law (Powell and DiMaggio Reference Powell and DiMaggio2012; Scott Reference Scott2013). Just as social psychologists and others have documented implicit bias at the individual level, Edelman identifies a reflexive response to law that perpetuates well-established biases at the organizational level. In Edelman’s account, organizations respond to ambiguous laws by producing structures such as equal employment opportunity (EEO) offices, policies, and complaint procedures that symbolize compliance. Human resources professionals feed this process by inflating legal risk and offering organizational structures as insurance against those risks. Complaint handlers within organizations reframe disputes as management issues or personality problems, draining civil rights violations of their moral context. And over time, courts come to assume that symbolic structures, like Wal-Mart’s antidiscrimination policy, are effective without investigating whether they actually are.
THE STRUCTURAL TURN IN EMPLOYMENT DISCRIMINATION SCHOLARSHIP
Endogeneity theory addresses a core tension in the Wal-Mart opinion between understanding inequality as a product of individual bias and agency and understanding it as a product of systemic organizational practices. In the Wal-Mart opinion, agency resides solely in individual managers, not the organization. Structure, almost by definition, is neutral, and formal antidiscrimination policies are sufficient to defeat liability for systemic discrimination even in the face of significant evidence that they are ineffective. As Tristin Green argues, Wal-Mart is part of a judicial shift toward viewing discrimination “as a problem of low-level, rogue employees acting on biases … without the influence and against the interests of the organizations for which they work. Organizations are innocent under this view. They provide the venue, the neutral physical architecture for discrimination, but nothing more” (Green Reference Green2016, 1).
This developing judicial perspective rejects the long-standing scholarly concept of institutional bias, even though there has been a “structural turn” in both legal and social science research about the workplace in recent years (Bagenstos Reference Bagenstos2006). These recent structural theories respond to empirical findings in social psychology that document the persistence and pervasiveness of unconscious bias (Banaji, Hardin, and Rothman Reference Banaji, Hardin and Rothman1993; Krieger Reference Krieger1995; Banaji and Hardin Reference Banaji and Hardin1996; Greenwald et al. Reference Greenwald, Banaji, Rudman, Farnham, Nosek and Mellott2002; Kawakami, Young, and Dovidio Reference Kawakami, Young and Dovidio2002). Importantly, these empirical findings indicate that even people whose personal beliefs are relatively free from bias are still susceptible to stereotypes in the same way as people who hold biases toward specific groups (Bielby Reference Bielby2000). Complementary sociological research identifies the organizational structures and conditions that facilitate that bias in workplace decision making (Krieger Reference Krieger1995; Bielby Reference Bielby2000; Fiske and Taylor Reference Fiske and Taylor2013), and in some instances, the structures that limit the operation of that bias (Nelson and Bridges Reference Nelson and Bridges1999; Bielby Reference Bielby2000; Uhlmann and Cohen Reference Uhlmann and Cohen2005; Kalev, Dobbin, and Kelly Reference Kalev, Dobbin and Kelly2006; Cheryan et al. Reference Cheryan, Plaut, Davies and Steele2009). Workplace practices that facilitate the operation of bias include subjective decision making, word-of-mouth recruitment for open positions, and unfettered discretion that facilitates social closure (Krieger Reference Krieger1995; Bielby Reference Bielby2000; Sturm Reference Sturm2001; Albiston and Green Reference Albiston and Green2018). Taken as a whole, this body of research indicates that workplace structures and practices are not merely neutral backdrop, but instead may either facilitate or limit the impact of implicit bias.
In response to this research, legal scholars proposed reforms that focus the law’s attention on workplace structures and practices that facilitate or ameliorate bias (see, for example, Krieger Reference Krieger1995; Green Reference Green2016). These proposals would hold employers accountable for adopting practices, such as subjective decision making, that facilitate bias, and for failing to take structural steps to limit the effects of implicit bias and stereotyping. Legal scholars have argued that such laws would prod human resource professionals to identify and disseminate best practices to create what Susan Sturm calls “a floor of acceptable conduct” that would rise as practices continually improved (Sturm Reference Sturm2001, 6).
At their heart, these proposed structural reforms seek to hold employers accountable for workplace practices that enable the operation of bias. In this view, organizations are not merely neutral physical infrastructure; instead, they make choices about how to structure decision making, job ladders, recruitment practices, and the like that have significant consequences for inequality. Managers may be individual agents, but those agents act through, and are shaped by, the workplace structures and practices that organizations create for them. Although organizations might not choose those practices specifically because they further inequality and bias, structural theorists argue that they still should be held accountable when those choices systemically disadvantage a protected class of workers. On paper, decades of legislation and Supreme Court precedent is consistent with this proposition,Footnote 2 yet in Wal-Mart an apparently ineffective paper antidiscrimination policy was enough to negate widespread evidence of bias facilitated by the structural choices of the employer.
STRUCTURE, NOT AGENCY, UNDERMINES CIVIL RIGHTS
What can we make of this result? Are employers and courts intentionally relying on symbolic compliance to undermine civil rights protections? Unlike other scholars who explain the uneven outcomes in employment cases in terms of judicial and employer hostility to civil rights, Edelman seems somewhat agnostic about the degree of agency and intention involved in creating and deferring to symbolic structures. For example, Edelman states that “when judges defer to symbolic structures without evaluating the efficacy in achieving the goals of civil rights laws … [t]hey usually do so inadvertently” (219). In effect, she locates the problem in inaccurate heuristic decision making, not judicial ideological bias. In this case, however, judges’ heuristic decision making responds to organizational efforts at compliance with law, rather than to individuals’ implicit stereotypes of particular groups of people.
Similarly, in Edelman’s account, organizational complaint handlers are not necessarily biased against civil rights claims. Instead, they are steeped in managerial norms and incentivized by the realization that they may be deposed in a future lawsuit, all of which colors their interpretation of disputes. Human resource professionals are not opposed to civil rights, which of course help justify their legitimacy and status. Instead, Edelman’s findings suggest that they inflate the risk of employment discrimination suits and then propose their services and advice as protection because they are motivated in part by professional advancement and justification. Human resources professionals become indispensable advisors when employment litigation threatens, the logic goes. To maintain their authority within the organization, however, they must advise organizations in ways that do not interfere with managerial norms and objectives.
Thus, Edelman does not suggest that organizations create symbolic structures because they are driven by overt animus. Instead, Working Law provides a convincing empirical account of the structural incentives baked into the social environment of organizational actors and the professions, and how they come over time to define compliance for organizations and eventually the courts. Yet, while overt animus may not be the driving force, the result is as though it were. Symbolic structures that in some instances do nothing, but still transmit legitimacy as an indicium of compliance, invite a heuristic rather than searching look at practices that may perpetuate inequality. They also help legitimize the status quo by appearing to respond to discrimination without meaningfully changing the reality on the ground.
As Sam Bagenstos convincingly argues in “The Structural Turn and the Limits of Antidiscrimination Law” (2006), endogeneity theory explains why the structural turn in legal scholarship will not solve the problem of ineffectual civil rights laws. Indeed, Working Law demonstrates how structural reforms proposed by legal scholars can become symbolic acts that do little to change things on the ground yet convince courts that workplaces are in compliance with the law. Even well-meaning attempts at structural reform, such as implementing written policies, sexual harassment training, and EEO offices have been shown to be largely ineffective (Kalev, Dobbin, and Kelly Reference Kalev, Dobbin and Kelly2006) and in some instances have been shown to backfire (Tinkler, Li, and Mollborn Reference Tinkler, Li and Mollborn2007). Yet, as Working Law demonstrates, courts increasingly accept these structural reforms as legal compliance without meaningful inquiry into their effectiveness. This turns the call for structural reforms on its head because the very structures intended to improve workplace outcomes often justify and legitimize the unchanged status quo instead, much like the paper antidiscrimination policy that absolved Wal-Mart from responsibility for the bias of its managers.
QUESTIONS FOR FUTURE ENDOGENEITY RESEARCH
Under What Conditions Are Organizational Structures Effective?
Here I raise a few thoughts on some of the unexplored corners of endogeneity theory as a framework for future research. The first has to do with how we should understand symbolic structures. Edelman notes that by symbolic, she means that the structures evoke a notion of legality and compliance. She emphasizes that symbolic does not mean, nor does it necessarily imply, ineffective, and it is not the opposite of substantive (Edelman Reference Edelman2016, 101). This raises what Edelman calls the “key question”: how, and under what conditions, do symbolic structures engender more (or less) substantive compliance? One might ask even more directly, under what conditions are symbolic structures meaningfully substantive or merely figurative?
Chapter 6 of Working Law sets out to answer this key question. It addresses at length the “how” question, for example, the mechanisms through which symbolic structures engender more or less substantive compliance. For instance, internalized dispute resolution procedures that rely on organizational complaint handlers tend to managerialize problems and solutions rather than focusing on civil rights principles. Similarly, organizations engage in loose coupling between symbolic structures and everyday practices by creating policies on the books that have little impact on day to day production.
While these examples tell us much about how what could be effective responses to law come to be symbolic window dressing, they tell us less about the conditions under which symbolic structures might be made more effective. Edelman points to one study that shows that practices that establish organizational responsibility for fair representation of women and underrepresented minorities are effective (Kalev, Dobbin, and Kelly Reference Kalev, Dobbin and Kelly2006). But, as she recognizes, more theory and research are needed to examine the conditions under which structures are or are not effective.
Worker Agency in Organizations
A second question has to do with the agency of civil rights holders in this account of endogeneity in the workplace. While symbolic structures certainly give organizations the ability to frame the meaning of compliance, rights themselves also provide a symbol that workers might mobilize in the workplace. In my own work, and the work of Michael McCann, for example, individuals mobilize rights as cultural discourse to gain leverage in interactions with workplace organizations and to frame the legitimacy of their claims (Albiston Reference Albiston2005, Reference Albiston2010; McCann Reference McCann1994). This may be another mechanism through which legal logics interact with managerial and organizational logics, but not necessarily in the way that Working Law describes. Thus, understanding the processes through which workers come to construct and mobilize their rights becomes an important avenue of inquiry to identify possible sources of resistance to endogeneity (Albiston Reference Albiston2005, Reference Albiston2010; Dubal Reference Dubal2017).
More generally, endogeneity theory tends to treat institutionalization as the province of organizations, human resource professionals, and deferential courts. The other major constituency in the workplace is workers. Is there any countervailing pushback by workers? If so, why is it not effective? Are systematic rather than individualist forms of action needed? More on this context would be interesting to explore. Consider, for example, the developing organizational form of worker centers, which enable collective action by workers without the legal restrictions on unions (Fine Reference Fine2006). This is not to suggest that Edelman’s claims are incorrect, only to suggest that agency on the part of workers and their advocates remains somewhat unexplored in Working Law, and that Edelman’s call for more awareness and scrutiny on the part of plaintiffs’ attorneys might be extended to other, more collective, forms of worker advocacy.
Theories of Power
A third question has to do with the theory of power that underlies this model. In neoinstitutional organizational theory more generally, theories of power are often either implicit or absent. To be sure, there are important analyses of power associated with how professionals spread ideas about risk and structures. Their strategies are effective because of the power of professions to claim control over an abstract body of knowledge and to frame solutions within it. Yet power is sometimes hard to pin down in the process of institutionalization. The neoinstitutional literature identifies a gradual process of habituation and consolidation around certain practices (Berger and Luckman Reference Berger and Luckmann1991). This leaves ambiguous why some structures and practices become institutionalized and others less so, and what role power plays in this process. Cost is one factor—easier and relatively costless symbolic responses tend to be adopted first, as Edelman points out. But what else contributes to the mix, and is there room here for powerful actors to intervene to limit ambiguity and encourage organizations to adopt practices that are more effective? Future work should consider comparative analysis to understand the forms of power that give rise to variation in outcomes.
Formal Law
One final suggestion for future research is the role of formal law, doctrine, and politics. In Working Law, formal law is, first, an ambiguous starting point for the process of endogeneity and last, a depressing end point as courts come to defer to symbolic structures over time. What happens in between is less closely examined. During the same period of institutionalization that Working Law explores, however, formal law followed a cyclical process of institutionalization and deinstitutionalization of outcome-based theories. Disparate impact theories, which focus on measurable substantive outcomes for underrepresented minorities and women, became disfavored. Legal developments in the 1980s such as Wards Cove Packing Co. v. Atonio (1989) and its progeny undermined disparate impact theory by watering down what employers were required to show to justify practices that produced disparate outcomes. The Civil Rights Act of 1991 partially restored disparate impact theory by codifying it for the first time and overriding much of the Wards Cove decision, but the Act also codified requirements, such as identifying a particular practice that causes a disparate impact, that make disparate impact claims harder to prove. To the extent that the process of endogeneity through the 1990s shows a move away from the courts carefully examining actual outcomes, this could be seen as reflecting the doctrinal move toward favoring disparate treatment theories that tend to individualize discrimination and, absent evidence of individual intent, paying less attention to evidence of unequal outcomes.
Legal endogeneity is certainly related to that shift and facilitated by it, but the sharpest rise of judicial deference occurs somewhat later than the decline in outcome-based theories of discrimination, suggesting that changes in formal law may have contributed in important ways. It also may be that shifts to the right in politics and in the courts are factors that drive changes in the law, judicial deference to symbolic structures, and organizational responses to law.
Future researchers should consider what doctrinal changes would be desirable to limit the effects of legal endogeneity. The heuristic decision making invited by the Iqbal/Twombly pleading standard is one example that comes to mind (Bell Atlantic Co. v. Twombly 2007; Ashcroft v. Iqbal 2009). Together, these two cases rejected the relatively liberal notice pleading standard, holding instead that “only a complaint that states a plausible claim for relief survives a motion to dismiss. Determining whether a complaint states a plausible claim for relief will … be a context-specific task that requires the reviewing court to draw on its judicial experience and common sense” (Ashcroft v. Iqbal 2009, 679). Such “plausibility” language invites the kind of conclusory reasoning seen in Wal-Mart that surely most managers in a corporation that forbids sex discrimination would select sex-neutral performance-based criteria for hiring and promotion. Some explicit guidance limiting the inferences that can be made from symbolic structures without evidence of effectiveness would be helpful to counter the heuristic tendencies of the Iqbal/Twombly duo.
CONCLUSION
Working Law makes a landmark contribution to understanding the relationship between law and organizations. It also raises vitally important questions for future research. What are the conditions under which symbolic structures can be made effective on the ground to improve the circumstances of women and underrepresented minorities? What is our theory of power in the development, deployment, and institutionalization of symbolic structures as symbols of compliance? What legal reforms might be possible to limit the reach of legal endogeneity? More generally, Working Law strongly suggests that a fundamental shift in the approach of employment discrimination law toward more systemic factors and less focus on the agency of individual bad actors would be helpful. And this, many employment discrimination scholars would argue, is long overdue.
CASES CITED
Ashcroft v. Iqbal, 556 U.S. 662 (2009).
Bell Atlantic Corp v. Twombly, 550 U.S. 544 (2007).
Connecticut v. Teal, 457 U.S. 440 (1982).
Griggs v. Duke Power Co., 401 U.S. 424 (1971).
Hazelwood School District v. United States, 433 U.S. 299 (1977).
Teamsters v. United States, 431 U.S. 324 (1977).
Wal-Mart v. Dukes, 564 U.S. 338 (2011).
Wards Cove Packing Co. v. Atonio, 490 U.S. 642 (1989).
Watson v. Fort Worth Bank, 487 U.S. 977 (1988).
STATUTES CITED
Civil Rights Act of 1964, Pub. L. No. 88-352 (1964).
Civil Rights Act of 1991, Pub. L. No. 102-166 (1991).
