In a recent Boston College Law Review article, Employer Liability for Non-Employee Discrimination, Professor Dallan Flake (Ohio Northern) addresses a subject that has generally perplexed me as well as many employees and employers—how courts can develop a cohesive framework under Title VII to address employer liability for employment discrimination actions due to the behavior of company outsiders. In particular, I have always wondered about the usual trope that customer preference cannot be a defense in discrimination claims while recognizing that there is nothing more important to employers than the preferences of their customers. This article catalogues a host of very interesting cases describing the acts of customers and other non-employee harassers or their biased preferences that raised liability concerns for employers in discrimination claims brought by their employees. Flake’s thought-provoking discussion of these cases offers a noteworthy guide for employers developing policies with respect to discriminatory influences from outsiders.
The article argues that increasing employer involvement in the service industry has led to a number of integrated business models, including outsourcing, that pose new legal challenges when considering non-employee actions. To a large extent, the article illustrates initially how the workplace has evolved from a binary employer-employee relationship by triangulating into an employer-employee-customer relationship. As a result, employees are more likely to interact with non-employee customers or clients, vendors, suppliers, temporary employees, and independent contractors all potentially located at the same worksite. Although the analysis discussed could apply to any of these influential non-employee relationships arising within many of the newer business structures, most of the article emphasizes the challenging dynamics posed by discriminatory actions of customers.
Under Flake’s thesis, the law should not use the same standard for discrimination by non-employees as it does for discrimination by fellow employees because an employer can more easily control its own employees’ harassing behavior. Flake offers a unitary standard of reasonableness for all employment discrimination claims involving acts by non-employees that would establish employer liability under a two-pronged approach: “(1) whether [the employer] knew or should have reasonably known about the non-employee discrimination and (2) whether it acted reasonably in response to the discrimination.” (P. 1173.)
In an intriguing classification of the ways in which non-employee discrimination against employees occurs, the article divides these claims into four categories:
(1) “conscious and direct (such as when a customer sexually harassed a waitress)”;
(2) also conscious but “indirect (such as when airlines hired only female flight attendants based on customer preference)”;
(3) “unconscious… directly, such as when restaurant diners unintentionally tip black servers less than white servers”; and
(4) also unconscious but “indirectly, such as when customers give implicitly biased feedback to employers that is then used to make employment decisions.” (P. 1174.)
The article also highlights how different analytical constructs may apply to these categories of claims. For harassment, an employer can be liable if it has actual or constructive knowledge of the non-employee’s behavior unless the employer shows that it promptly and reasonably acted to end the harassment. For an employer to prevail when subjected to a claim of customer-based preference as indirect discrimination, it must show that its actions were justified by a bona fide occupational qualification (BFOQ) necessary to the essence of the business. For unconscious discrimination, either directly from customers’ unintended actions or indirectly based upon customers’ hidden biases as submitted or inferred, an employer would need to establish that its actions were job-related and consistent with business necessity. Flake believes that these different approaches to dealing with non-employee discrimination have created confusion and fragmentation warranting a unitary standard requiring knowledge and reasonableness in responding as components related to establishing employer liability.
The article does not devote much detail to the inability of employers to control outsiders as compared to its own employees. In continuing explorations of these topics, it might be helpful to examine in more depth whether current employment discrimination analysis operates more as a shield for employer liability without the need for more flexible employer defenses. For example, one might question whether employees have the resources to bring claims challenging an employer’s systemic discriminatory decisions as being based upon customer preference or resulting in a disparate impact. Most discrimination claims tend to be focused on the treatment of individuals by their employers with employees facing difficult burdens of persuasion. Nevertheless, this article makes an important contribution by identifying how customers’ discriminatory preferences based upon unconscious bias might, at a minimum, require a more nuanced analysis regarding employer liability.
Furthermore, Flake asserts that any distinctions or difficulties in an employer’s ability to control the different actors may be covered within the reasonableness standard he suggests. That standard applies equally in cases involving both employee and non-employee harassment. The actual analytical changes imposed by the suggested unitary standard and its knowledge requirement would arise in intentional discrimination cases involving customer preferences and BFOQ claims that were not blatantly discriminatory when addressing concerns of privacy, safety, or authenticity. More provocatively, this unitary standard would also add an employer knowledge requirement in cases based on a disparate impact claim related to customer evaluations, a result possibly “unpalatable to some” as Flake has conceded. (P. 1215.) Apparently, he will address this issue more specifically in his next article, cited in footnote 88, When Should Employers Be Liable for Factoring Biased Customer Feedback into Employment Decisions? (forthcoming in the Minnesota Law Review in 2018).
Even if you do not agree with Flake’s assertion that lesser control over non-employees who discriminate and the challenges of identifying unconscious discrimination warrant a more flexible analysis requiring knowledge before employers become liable for intentional customer-preference discrimination or disparate impact, the article is an engaging read because of the creative way in which it chronicles the riveting theories of non-employee discrimination developed from the cases. Also, the article shines an important light on customer preference cases that may involve unconscious bias. Given the lack of scholarly attention to this subject, the article initiates some crucial steps in understanding the analytical development of employer liability related to the discriminatory behavior and preferences of influential outsiders.
Andrew Verstein, The Jurisprudence of Mixed Motives
, 127 Yale L.J.
(forthcoming), available at SSRN
To say that the law of causation in mixed motives cases is a mess would be an understatement, as Andrew Verstein highlights in his article, The Jurisprudence of Mixed Motives. Most antidiscrimination laws require causation. That is, these laws proscribe adverse employment actions when they occur “because of” a protected characteristic, such as race or sex. The problem is that there are several types of causation, particularly where multiple motives are involved – which is almost always. Yet, few of those statutes specify what type of causation is required. Other statutes specify what type of causation is required, but with no clear definition (e.g., “motivating factor” causation, referenced in the Civil Rights Act of 1991). To make matters worse, courts and commentators often throw other undefined or ill-defined terms into the mix. And if we were inclined to look at other legal fields, such as tort law or constitutional law, in order to make sense of causation in employment discrimination law, we tend to encounter yet more ill-defined terms.
One might think – or at least hope – that it would be possible to (1) identify the universe of potentially applicable causal standards; (2) clearly define each of those standards (and their relationship to one another); and (3) attach a universally applicable and accepted label to each causal standard. That is, we might imagine a Rosetta Stone that would allow us to clear up the confusion that reigns in the Babel of causation. Such a tool would allow us to describe the law with precision and engage in meaningful (and perhaps even cross-substantive) discussions about the normative merits of any particular causal requirement.
I attempted a project like this in 2006, in an article entitled The Fundamental Incoherence of Title VII: Making Sense of Causation in Disparate Treatment Law. That piece posited a universe of causal concepts based on core philosophic causal terms (necessity and sufficiency), arranged from least restrictive (most plaintiff-friendly) to most restrictive (most defendant-friendly):
Yet that piece had two important limitations. First, it conceptualized causation solely as qualitative. That is, in my work, a cause could be necessary, or sufficient, or some combination of those qualitative concepts. But I had no way to conceptualize one cause as “stronger” than another, or to conceptualize a cause as being “big” or “small” – concepts that are intuitively important, and to which many courts and commentators have referred. Second, and relatedly, without a quantitative component to my analysis, I undertheorized the concept of “minimal causation” (the best definition for “motivating factor” causation). There may be important variations within minimal causation that I failed to capture.
Now, in The Jurisprudence of Mixed Motives, Professor Verstein has developed a model that truly advances the search for a Rosetta Stone for causation. His model posits a quantitative view, in which motives have varying degrees of causal influence in decision-making. With this approach, he is able to depict the causal force of multiple motives on a graph as follows:
His graph depicts a simplified world in which a defendant bases a decision (such as a decision to fire the plaintiff) on two motives: B-Motive, which is proscribed (such as race), and A-Motive, which is not proscribed (such as tardiness). Once the causal force of either motive alone or the two motives combined reaches a value of 1 – i.e., the combined causal force is sufficient – the decision in question (the firing) will occur.
This model allows us to understand the relationships between the traditional quantitative concepts. Professor Verstein prefers other labels for the quadrants in his graph, but it is easy to translate his terms into traditional causal language as follows:
This provides an incredibly useful way of viewing these causal concepts and the relationship between them.
Three important concepts are readily apparent from this graph. First, we can illustrate the concept of “de minimis” causation: the area on the left side of the graph closest to the A-Axis. This concept of “de minimis” causation might explain the “substantial factor” language used by some courts and commentators: “substantial factor” might mean “more than de minimis.” Second, we can see that B-Motive can be “de minimis” even when it is a necessary cause (on the leftmost side of the gray triangle), as well as where it is neither necessary nor sufficient (on the leftmost side of the upper left quadrangle or the leftmost side of the lower left black triangle). This suggests a need to be more specific about other causal requirements if we want to use a “substantial factor” test. Third, we can see that there are actually two types of motivating factor causation. Recall that motivating factor means that B-Motive has some causal force, but is neither necessary nor sufficient. One type of motivating factor causation is depicted in the upper left quadrant and the other is depicted in the lower left black triangle. Given that the adverse employment decision (firing) will occur only in the upper left quadrant, that is the most likely candidate for the meaning of motivating factor causation in most statutes. However, in the lower left black triangle, which likely depicts what Justice O’Connor referred to as “thought crime,” the defendant still engaged in decision-making based on a proscribed criterion (race), which still raises normative issues. (This is why I noted that I undertheorized motivating factor causation in my 2006 piece.)
Next, Professor Verstein introduces the idea of B-Motive “predomination” (where the causal force of B-Motive is greater than that of A-Motive). We can see this on his graph, as follows:
This concept of predomination might explain the “primary factor” language used by some courts and commentators: “primary factor” might mean “B-predomination.”
Professor Verstein then combines these insights into a multi-zone graph that seems truly to define the universe of potential causal concepts:
From there, he goes on to label those concepts, focus on the most commonly used concepts, and compare those concepts across various fields of law – including not only employment discrimination law, but also fields as diverse as constitutional law and tax law. The potential for courts and scholars in one field to learn from those in other fields is exciting and unprecedented. Moreover, this approach lays the groundwork for a clear and fruitful discussion of the normative merits of various causation requirements that is unprecedented. In fact, Professor Verstein indicates that this is his next project. To say that this is an exciting development in the law of causation would be an understatement.
In closing, two notes seem warranted:
First, although I refer to his model as being about causation, Professor Verstein eschews thinking of motives as exerting causal force. He eschews causal language because he (understandably) does not want to take sides in the ongoing debate about the determinism of motives – i.e., whether we have sufficient self-determination to resist the influence of our motives. But for those of us who are more concerned with the law of causation and less concerned with how that law treats self-determination, his model provides an incredibly useful way of thinking about causation in mixed motives cases.
Second, I have some minor quibbles with Professor Verstein’s piece, and how I think he could make it even better, which I address in a forthcoming piece, A Rosetta Stone for Causation. None of those quibbles detracts from the fact that his piece is one of the most exciting and important ones I have seen in this field. I wholeheartedly recommend it.
Stephanie Bornstein, Equal Work
, 77 Md. L. Rev.
(forthcoming 2018), available at SSRN
In her article, Equal Work, Professor Stephanie Bornstein (Florida) does a superb job of providing a fresh approach to the continuing problem of pay discrimination in the workplace on the basis of gender and race. As Professor Bornstein correctly acknowledges, pay discrimination has remained an ongoing problem in our society for decades. Her article makes two extraordinarily useful contributions: first, it undertakes a comprehensive survey of the latest literature on pay discrimination and its causes, and second, it offers solutions that do not require legislative reform to chip away at this pervasive problem.
Early in her article, Professor Bornstein provides a comprehensive review of the existing data on gender and pay discrimination in the United States. This overview does a nice job of bringing together all of the most up-to-date research in this area. Then, Professor Bornstein makes an effort to explain why the pay gap continues to exist in the course of exploring recent research on occupational segregation and income inequality in this country. To this end, Professor Bornstein also addresses the complexities of the law on pay discrimination in the workplace, closely examining Title VII of the Civil Rights Act of 1964 and the Equal Pay Act (EPA), before discussing recent efforts on the state level to resolve these problems, including statutes passed in California, Massachusetts, and Oregon. The paper also correctly recognizes the many efforts to close the race and gender pay gap that have been made over many decades and at all levels — federal, state, and local. While some of these efforts have helped reduce pay disparities at different points in our nation’s history, the gap remains persistent and efforts to fix the problem have stalled in recent years.
Thus, Professor Bornstein provides an exhaustive descriptive review of the research in this area, the most current data, and an analysis of the basis for the problem. For this reason alone, the paper promises to serve as a seminal work in this area that other scholars can look to when examining different aspects of this topic. The greatest contribution of this work, however, is Professor Bornstein’s unique approach to helping to resolve these unfair pay disparities. Professor Bornstein’s proposal focuses on the first, threshold prong of the statutory test under the EPA — that the statute requires equal pay for “equal work.” This analysis distinguishes her scholarship from much of the other literature on this topic, which often focuses more on the possibility of narrowing the employers’ “any factor other than sex” defense in the statute.
In this way, the article suggests a new approach to addressing the pay problem. Noting the difficulty of bringing legislative reform to this area, this article discusses ways of addressing pay discrimination problems through existing legal frameworks. By way of example, the article explores how employers in the union and governmental sectors have found ways to create equal pay systems that do a better job than non-union private sector workplaces of paying minorities and women equally to what they pay white men. Professor Bornstein’s approach plays off these examples, arguing that employers must make broader comparisons between workers when reaching pay determinations. Her approach thus advocates that “every difference between jobs [created by an employer] need not be wage-determinative.” This approach explains how the current narrow interpretation of the law is outdated, and how a more comprehensive approach — which applies a broader framing of what constitutes equal work — is now critical to the modern workplace. The unique strength of this piece, then, is its detailed explanation of how we can work within existing legal structures to better protect women and minorities from pay discrimination. The paper thus avoids some of the practical pitfalls that can befall proposals for broad legislative reform in this area.
It is difficult to overstate the important contribution Professor Bornstein makes here with Equal Work. The vast collection of data on the topic and the exhaustive background information provided are by themselves quite impressive. However, the novel approach to helping fix the problem advocated by Professor Bornstein makes this work an invaluable contribution to the academic literature.
While some readers may already know this work, legal academics do not always keep up with monographs that focus on history. So, I will try to widen the audience for this excellent book.
A Class by Herself traces the story of “protective legislation” — e.g., laws regulating wages and hours — concentrating on debates over statutes that applied only to women. The story begins in the Progressive Era of the late nineteenth and early twentieth centuries, continues through the New Deal, and ends with questions about modern laws such as the Pregnancy Discrimination Act and Title VII. The book does an impressive job with multiple historical subjects: legal history, history of “worklaw,” history of feminist thought, and history of politics and of the state.
Most prominently, the book features a nuanced take on the tensions between “difference” feminists (who favored protections that applied only to women) and “equality” feminists (who insisted the same rules should apply to both sexes). Both sides get their due. The “Brandeis Brief” from Muller v. Oregon (1908), once romanticized for trying to bring evidence from social sciences into the courtroom and later castigated for its jarring patriarchal assumptions, gets a balanced treatment. Notably, the book covers a wide time period, and it shows the enduring importance of these debates while still putting them in specific political, social, and legal historical contexts. Scholars of modern law will have much to ponder with Woloch’s discussion of UAW v. Johnson Controls, Inc., 499 U.S. 187 (1991), an emphatic expression of the “equality” feminism the New Deal and Great Society had adopted, in which the Court held that an employer rule barring women from working around substances dangerous to pregnant women violated Title VII. While good history doesn’t necessarily need to draw explicit parallels with the present, it’s impressive to see such history done well.
Woloch also has a wonderfully interesting and complex cast of characters. The book features not only famous actors such as Florence Kelly (of the National Consumers League, an advocate of protective legislation for women) and Alice Paul (of the National Woman’s Party, an advocate of equality feminism), but also other fascinating, lesser-known figures, including but not limited to progressive contemporary Elizabeth Baker and, later, Women’s Bureau Director Esther Peterson. In this regard, Woloch adds some fascinating new primary sources, such as Baker’s PhD dissertation.
In telling this story, Woloch shows she understands the internal rules regarding legal doctrine (including debates between “sociological” and “formalist” jurisprudence) without pretending such evolution occurs in a vacuum. Much of the work is told from the point of view of reformers, and the book demonstrates both how political activists shaped the law and how the law shaped the strategies of activists. As with the best history, Woloch stresses how complex and contingent many events were: actual results were far from inevitable, and it is a virtue of the book that its goal is not to crown as “correct” one side or the other in the debates. Woloch rejects simplistic analysis, and her contributions include the memorable line, “[E]quality tends to favor those best equipped to claim it.” I liked this book a lot.
The hottest new subject area for legal academics is privacy law. The field is still in its infancy, which means that many of the foundational issues of categorization and taxonomy remain to be worked out. Even defining privacy is thorny, as the label is applied liberally to all sorts of invasions, intrusions, disclosures, and interests, with questions about how to delineate and frame privacy concerns arising in several sub-categories. In the labor and employment context, the Restatement of Employment Law has divided privacy interests into three categories: interests in the privacy of persons and locations (including electronic ones); interests in the privacy of personal information; and interests in the nondisclosure of information that was disclosed to the employer confidentially. Within these broad contours, a myriad of different types of violations sit uncomfortably next to one another, each raising its own specific issues.
In Limitless Worker Surveillance, Ifeoma Ajunwa, Kate Crawford, and Jason Schultz take on one specific type of privacy invasion: workplace surveillance. Because of employers’ ownership and control of the workplace, courts have generally interpreted the common law to allow employer surveillance of working areas. Moreover, employers can expand their observation into private areas when employees give their consent to the intrusion. Management has generally run into trouble only when it surreptitiously and secretly records employees in private areas, such as bathrooms, medical examination offices, and employee homes. But routinized, disclosed surveillance is commonplace and seen as part of the job. Ajunwa, Crawford, and Schultz, however, want to disrupt this equilibrium. They argue that freedom from surveillance should be a non-waivable right, and they propose three different federal statutes, varying in scope, to enforce variations of this right against employers.
The article does not spend a lot of time exploring why surveillance, at root, is problematic as a practice—the authors assume that sustained and continuous observation is noxious and enervating. Instead, they dig into the technological changes that afford companies cheaper and easier ways of observing, measuring, and tracking their employees than existed in the days of Frederick Taylor and Henry Ford. Now, digital cameras can record an eternity of activity without the cost of film or videotape. RFID tags and GPS tracking can follow workers all over town without significant investment. Computer software can record everything an employee does online without the employee ever knowing. There really has been a sea change in the levels and types of information available to most employers about their employees. The authors single out wellness plans and productivity apps for special attention, discussing how these programs generally work and how they access workers’ private information in the name of health outcomes and improved productivity.
Ajunwa, Crawford, and Schultz survey the existing legal protections for workers and find them wanting. They point out the absence of regulation in privacy statutes such as the Electronic Communications Privacy Act and the Computer Fraud and Abuse Act. Federal discrimination statutes likewise do not address surveillance; the NLRA does address surveillance, but only in limited circumstances. There are state-level regulations of video and audio monitoring, as well as a growing number of state restrictions on employer access to employee social media accounts. But these protections are spotty and often limited, sometimes allowing surveillance with notice (or even without it). (Interestingly, the authors use state statutes protecting off-duty smoking as a proxy for privacy, even when these statutes were often more a result of tobacco-industry lobbying than an interest in workers’ rights.) Although states like California have fairly robust surveillance regulations, many states offer little to no protection.
The solution, according to the authors, comes in the form of a federal statute to regulate surveillance on the job. They begin with the possibility of a comprehensive privacy regime along the lines of the European Union’s General Data Protection Regulation. Ironically, they argue such a wide-ranging approach might fail because it would necessarily be susceptible to too many exceptions based on consent or work-related data collection. Instead, they suggest a federal “Employee Privacy Protection Act” that would prohibit surveillance outside the workplace as a mandatory, non-waivable rule. A third possibility would be a federal “Employee Health Information Privacy Act,” which would fill the existing gap left by HIPAA regarding employer responsibility over personally identifiable health information. This act would give employees the right to consent to use of their health data as well as the right to have data destroyed at the end of the employment relationship.
Limitless Work Surveillance takes on an important area of workplace concern and provides a set of legislative solutions for consideration. And the authors savvily ascertain how technological changes are driving—and will continue to drive—the ways in which employees will find their personal privacy invaded and degraded. By calling attention to the issue of surveillance in the workplace at this particular moment in our history, Ajunwa, Crawford, and Schultz have made an important contribution to the literature. The article will hopefully spur additional research into the ways in which employers and employees can work together to manage their legitimate interests while protecting the core of human dignity.
Much ink has been spilled over whether platform workers — be they Uber drivers, Task Rabbit taskers, or others — are employees or independent contractors, and litigation over alleged misclassification of platform workers is ongoing. Likewise, there is robust debate over whether the rise of such platforms benefits workers by expanding their earning capacity and flexibility, or simply serves to increase income insecurity and income inequality. Beyond employment, similar debates rage over other platforms, such as whether Airbnb and other home-sharing platforms enhance consumer choices and provide individuals with positive ways to monetize underutilized space, or exacerbate shortages of affordable housing and undermine stable residential neighborhoods.
Orly Lobel’s article provides a broad framework in which to analyze such issues. The article is much broader than the work law implications of the platform economy, but it is extremely useful for scholars and policymakers facing work law issues.
Lobel usefully catalogues ten characteristics of platforms: They achieve economies of scale for individual-to-individual transactions. They facilitate utilization of idle capacity, whether it be spare bedrooms or workers’ otherwise unproductive time. They facilitate commerce over extremely small transactional units, even a few minutes. They allow individuals to monetize almost anything, turning them into mini entrepreneurs. They facilitate customization of transactions. They substitute access for ownership in consumption. They reduce overhead. They reduce barriers to entry. They facilitate dynamic pricing driven by demand. They offer dynamic ratings of providers by consumers and consumers by providers.
Lobel discusses what she regards as easy cases for applying traditional regulation to platforms. For example, she sees no difference with respect to taxation, which she regards as designed to extract a percentage of profits wherever they are produced. She sees the issue of taxing platform transactions as a question of efficient collection and suggests that requiring the platform company to collect taxes may be the most efficient solution. (Interestingly, Lobel does not follow through on the implications for collection of employment taxes.) Other easy cases are permitting and occupational licensing. Lobel urges that where these regulations serve primarily to erect barriers to entry, they be replaced by liability rules and insurance requirements.
Lobel regards the employment question as one of the hard cases. Another hard case is zoning regulations. These raise tradeoffs that policymakers must confront. With respect to the employment status of platform workers, Lobel suggests focusing not on classification but on the protections that are valuable to society, such as anti-discrimination, safety and health, insurance, and portable benefits, regardless of classification. Generally, she observes that “[o]ff-on categories such as consumer-business; employee-freelancer; residential-commercial are, in some instances, no longer viable as organizing frameworks…. What this means for regulators is that, rather than a unified legal entity, which has traditionally been the object of regulation, transactions are now shaped by multiple actors, with varying capacities, interests, and needs.”
Consistent with her scholarship concerning the “new governance” approach to regulation, Lobel highlights private approaches to handling risk: platforms providing insurance, platforms performing background checks, and dynamic ratings of service providers by consumers and of consumers by service providers. She recognizes that some of these developed as responses to litigation or were compelled by regulatory authorities. She observes that the advanced technologies employed by platforms enable better data collection and analysis, and urges that rather than devising regulations top-down, lawmakers collaborate with platforms to use the data to analyze problems and develop responses. Examples given include addressing ride-hailing platforms’ shortcomings in serving persons with disabilities and preventing discrimination by hosts on home-sharing platforms.
Lobel’s article raises many questions and provides no definitive answers. But it provides a very useful framework to consult in approaching legal issues raised by the platform economy with respect to treatment of platform workers as well as other social policy issues.
Editor’s note: For an earlier Jotwell review of The Law of the Platform see Margot Kaminski, Disruptive Platforms (July 19, 2017).
Jamillah Bowman Williams, Breaking Down Bias: Legal Mandates vs. Corporate Interests
, Wash. L. Rev.
(forthcoming 2017), available on SSRN
Those working in antidiscrimination law are well-versed in the central role that the business case for diversity plays in shaping policy. Even as enthusiasm for legal interventions in business or education has waned, the business case for diversity has remained persuasive. Courts have even relied on it to find practices that disparately impact certain groups discriminatory, affirmative action plans legal, and accommodations required. In fact, I would submit that the business case for diversity has eclipsed arguments about justice, inequality, or morality as reasons to support such measures.
That is why Jamillah Bowman Williams’ article, Breaking Down Bias: Legal Mandates vs. Corporate Interests, Wash. L. Rev. (forthcoming 2017), is so important. Williams asks the foundational question of whether the business case for diversity actually accomplishes the goal of antidiscrimination law – reducing bias and promoting racial inclusion – and reports on experimental research that tests the relative efficacy of the business case rationale versus a legal case for equity and inclusion. Williams finds not only that the legal case for diversity is more effective for reducing bias and promoting inclusion, but also that it exerts a stronger normative influence on actors than the business case.
Williams starts from the premise that we should work toward reduced bias, and her analysis focuses on the workplace. From that starting point, she traces the shift of strategies from antidiscrimination law to arguments that underrepresented groups should be integrated into organizations because their involvement makes those organizations work better or be more profitable. In this section, Williams describes the literature on why antidiscrimination law works to change behavior and why the focus shifted to diversity and inclusion in the 1990s and then became so entrenched.
Three main theories have developed to explain why antidiscrimination law is thought to reduce bias. The first assumes employers are rational actors who fear sanctions. The second posits that decision makers act based on moral beliefs about the right thing to do, which are in turn shaped in part by law. The third holds that decision makers act without thinking, based on roles and scripts, so that law is effective if it prompts the creation of new roles and scripts. While there is some experimental evidence supporting both the first and second theories, the history of antidiscrimination enforcement as recounted by Williams suggests that the third has a role as well.
Lack of enforcement in Title VII’s early days gave way to expansion and serious enforcement by the Equal Employment Opportunity Commission after the Equal Employment Opportunity Act of 1972. Heightened standards led to adoption of affirmative action policies, which, in turn, led to demand for EEO and management specialists to better shield companies from litigation. Then, in the 1980s, President Reagan cut the EEOC’s budget, opposed affirmative action, and appointed judges hostile to regulation and especially to affirmative action. But, the new cadre of EEO and management specialists remained popular. Employers simply reframed the purposes and goals of their affirmative action practices to focus on diversity management as a way to increase business effectiveness and profitability.
Williams explains the attractiveness of the business case for diversity:
a business case for diversity may be perceived as more legitimate than antidiscrimination law because it offers a connection between increased diversity and inclusion and positive performance outcomes. It may also be favored because it frames the efforts as proactive, to reap financial rewards, rather than reactive, to stop discrimination and avoid punishment.
There is some empirical evidence to suggest that diversity imposed as a result of an externally driven legal reason and not an internally motivated organizational value may result in resistance. Yet, Williams notes that “[t]he business case for diversity may persuade Supreme Court justices and America’s top business leaders, but the question remains whether this rationale is persuasive to the remainder of America’s workforce.”
To help answer that question, Williams conducted a laboratory experiment to look at whether the business case for diversity decreased bias and increased inclusion of racial minorities. Her results showed that white participants exposed to the business case actually reacted more harshly toward minority teammates than white participants who were not. She then conducted a survey-based experiment to look at whether the traditional legal case for inclusion might be more effective. That study showed that emphasizing civil rights law evoked more positive responses than the business case or no rationale at all.
Williams wraps up the paper by discussing some important implications of her results. First, she notes that conventional wisdom emphasizing the business case for diversity might actually undermine efforts toward greater inclusion. She uses social psychology to explain why. Essentially, the business case challenges deeply ingrained stereotypes and hierarchies without countering them overtly. Making people think about race may activate negative stereotypes, which can lead to negative treatment of minorities. The business focus on diversity may paint minorities as competitors for economic and social resources and may thus cause resistance because that competition poses a threat to scarce resources and privileges of the dominant group. Additionally, by emphasizing the benefit of diversity, organizations may send a message that the dominant group is less valuable. Finally, diversity values may generate resistance because they conflict with color-blind “meritocratic” ideology that is becoming more and more pervasive.
Williams stresses that not only may the business case backfire, but civil rights law has greater potential to promote positive beliefs and behaviors about inclusion. Although she does not frame the issues this way, as lawyers, we may be trained to be skeptical of the moral force of law; non-lawyers, though, find the law to be an expression of what is good and morally right. Williams is careful to say that she does not intend to cast doubt on the accuracy of the business case for diversity, which has a reasonable empirical basis, but instead to provide guidance on how organizations might communicate that goal to their members. Her findings offer compelling reasons to heed that recommendation.
One issue left unaddressed by this paper and open for future work is what to do with this information when it comes to the courts. As Williams notes, many legal scholars are concerned about the lack of antidiscrimination enforcement. She suggests that even with this lack of enforcement, the legal case for diversity is more effective than the business case at achieving diversity within organizations. Yet at some point, non-enforcement by courts might lead to the law’s erasure as a practical matter. So what then? If judges and lawyers are not persuaded by the moral foundations of the law apart from the law’s existence, then how can we stop them from undermining it?
I think that Williams’ work offers food for thought there, as well. Perhaps greater focus by advocates and scholars on those moral underpinnings of the law and on the current lived experiences of people of color can help persuade judges that the law should be enforced vigorously. Lawyering is often at least partially storytelling. We would do well to be sure we are telling the stories of the disempowered and focusing on the moral force of robust equality norms.
James Grimmelmann & Daniel Westreich, Incomprehensible Discrimination
, 7 Calif. L. Rev. Online
164 (2017), available at SSRN
I’m a fan of off-beat approaches to legal scholarship, having attempted a couple of efforts myself. And I try to keep up with developments in the real world that threaten to impact our discipline, like the concern about “Big Data” that has begun to appear in the law journals. So it’s no surprise that I was very taken by a particularly creative piece by Professors James Grimmelmann and Daniel Westreich, which combines an amusing conceit with dead-on analysis of an emerging and important question.
Incomprehensible Discrimination explores one aspect of a much longer article that appeared recently in the California Law Review by Solon Barocas and Andrew Selbst on data-driven algorithmic methods of making employment decisions. The aspect Grimmelmann & Westreich explore is the ability of such analyses to find correlations between, say, job performance and any of a variety of data points with no apparent causal connection to better performance. Barocas and Selbst conclude that traditional disparate impact analysis is not likely to invalidate these kind of selection process. Given sufficient data and robust tests of significance, it’s hard to conclude that reliance on such factors is irrational, even in the absence of any articulable explanation for what one has to do with the other. For Grimmelmann and Westreich, that’s exactly the problem.
For example, it is at least conceivable that better job performance can be predicted from characteristics such as the applicant’s favorite kind of music, whether she owns a car, or her zip code. But what confidence can we have in the resulting explanation?
Grimmelmann & Westreich’s method of exploring this question is a “(fictional) opinion of the (fictional) Zootopia Supreme Court of the (fictional) State of Zootopia,” which is designed to explore the disparate impact implications of this use of Big Data. They posit a hiring process that could easily result in racial (or, in the world of Zootopia, species) distinctions without any intent to discriminate. Thus, the Zootopia Police Department
uses a mathematical model to predict which applicants will be successful police officers. Four facts about this model are undisputed. First, its scores are significantly correlated with a reasonable measure of job performance. Second, the model does not explicitly consider applicants’ species. Third, it nonetheless systematically favors carnivorous applicants over herbivores. Fourth, no one has explained how and why the model works at predicting job performance or how and why it disadvantages herbivores.
The second undisputed fact is designed to rule out intentional discrimination. (Although it remains possible for a facially neutral model to be adopted for discriminatory purposes, that is very difficult to prove.) And, while the third raises a prima facie case of disparate impact, the first would seem to necessarily mean that the practice satisfied the business necessity/job relation standard under traditional analysis.
As for the fourth, that consideration does not currently have a place at the table under either disparate treatment or disparate impact analysis and, to be frank, wasn’t especially relevant before Big Data raised exactly that question. But it is this possibility that Grimmelmann & Westreich argue should be determinative in any court challenge to such a practice. According to them, such algorithmic analysis may “both predict job performance and discriminate against herbivores” without anyone being able to provide a causal explanation for the result. (P. 170.) The Zootopia Supreme Court elaborates:
The [plaintiff] League is correct that the factors that the model identified correlate with species, and the [employer] is correct that these factors also correlate with job performance. The problem is that there is no explanation in the record as to which of these two correlations, if either, is causal. It may be that the factors directly measure applicant characteristics that determine success in the challenging and dangerous field of police work, and that those characteristics happen to be unequally distributed in our diverse society. It may also be that these factors are instead measuring applicants’ species and that they measure likely job performance only because they are identifying species in an applicant pool where the relevant characteristics are unequally distributed. (P. 170.)
To deal with this risk, the authors argue that the defendant’s business necessity burden “requires it to show not just that its model’s scores are not just correlated with job performance but explain it.” (P. 170.) They suggest the necessity for “an explanation in terms of the chains of causation by which one state of affairs leads to another,” and elaborate that a good explanation “is one that identifies the hidden and nondiscriminatory variables connecting the observed factors with the predicted target variable.”
The reason, of course, is that, even a model that is scrubbed of any explicit race factors will still yield racially-slanted results if the factors actually used are correlated with race. Nor is this so strange. Although the authors don’t mention it, the Supreme Court has repeatedly had to deal with the question of whether legislative redistricting decisions are race-motivated (unconstitutional absent a compelling state interest) or politically motivated (constitutional) given that African Americans vote very heavily Democratic so there is a very strong correlation between the two.
According to the authors, that kind of inquiry is necessitated whenever impenetrable algorithms yield disparate results for protected classes. They end with a cri de coeur:
Our holding today is simple. Incomprehensible discrimination will not stand. Applicants who are judged and found wanting deserve a better explanation than, “The computer said so.” Sometimes computers say so for the wrong reasons–and it is employers’ duty to ensure that they do not. (P. 177.)
As is often the case, this Jot can’t do justice to the Grimmelmann & Westreich piece, which is well worth a read, as is the Barocas & Selbst article to which it responds.
Nearly sixty percent of American workers are paid on an hourly basis. Despite this reality, wage and hour law typically generates little attention in academic literature. While there has been considerable discussion about “the gig economy” and independent contractors, the nuts and bolts concerning how most Americans get paid goes largely unaddressed in legal scholarship. For this reason alone, a new article by Elizabeth Tippett (Oregon), Charlotte S. Alexander (Georgia State), and Zev J. Eigen (Littler Mendelson) represents a welcome addition to the literature.
The article focuses on workplace timekeeping software and the ways in which employers can use such software to commit wage theft. The authors review the functionality of thirteen different timekeeping software programs and explore how such software can be used by supervisors to effectively cheat employees out of wages for hours worked. Without going into great detail here, these programs vary considerably in their structure. Several of the programs allow – and in some cases, tacitly encourage – supervisors to review an employee’s submissions and then edit those submissions without providing any sort of notification to employees. The authors postulate that these types of programs pose the greatest risk of wage theft. Programs that afford supervisors less discretion present less risk.
Although the authors do not make any sort of empirical claims as to how common employer wage theft is, they do apply behavioral ethics theory to this situation to suggest, at least, that it would be unsurprising if supervisors engaged in unlawful edits to employees’ time and attendance data. For example, the authors discuss how behavioral ethics suggests that people are more likely to cheat when they can distance themselves from the dishonest transaction (such as where a supervisor can adjust an employee’s overtime hours online rather than directly ordering the employee to work off the clock). The authors speculate that, in light of all of this, employers might be underestimating the litigation risks associated with certain timekeeping systems.
These are all interesting observations, but part of what makes them noteworthy is the fact that this type of analysis is so uncommon in legal scholarship involving the law of the workplace. Questions about hours worked and overtime pay are of central importance to most employees. Yet, legal academics rarely discuss the nitty-gritty of how these things are measured in the workplace and how timekeeping systems might be abused.
The other feature of the article that makes it especially noteworthy is the fact that it serves as a reminder of an obvious, yet valuable, point. The Fair Labor Standards Act (FLSA), the primary federal wage and hour law, is an old statute. Like really old. And its main provisions have largely remained the same over time. Thus, as the authors note, it’s important to reflect on the fact that “the main law regulating work hours and pay for most employees in the United States has remained unchanged since before the Second World War.” (P. 9.) Even though the Department of Labor recently updated regulations regarding the overtime exemptions, the associated recordkeeping regulations have not been updated since 1987 and somewhat comically still contain references to “microfilm.” (P. 46.)
The authors ultimately offer a sensible update to the regulations in this area based on Department of Defense contractor guidelines. These guidelines, among other things, impose greater limitations on the ability of employers to modify timesheets without the knowledge of employees. But as part of their attempt to address the specific concerns associated with the regulation of timekeeping software, the authors make an important point regarding the need for workplace regulations to stay current with technological changes. As employers increasingly turn to technological advances to monitor their employees, the Department of Labor also needs to adjust. Nowhere does that fact seem more important than in the wage and hour context.
Perhaps the greatest success story for unions and workers over the last five years is the wave of state and local ordinances that have raised minimum wages, guaranteed sick and safe time, and improved corporate scheduling practices, among other worker-friendly innovations. At the same time, though, union density has remained flat, as unions’ successes in campaigns for better working conditions have not translated into large numbers of successful union elections. Moreover, with the election of Donald Trump, further erosion of union density is likely: unions will face a hostile NLRB possibly compounded by federal labor law reform in the form of a national “right to work” law or evisceration of collective bargaining rights for federal employees. In other words, unions’ campaigns to improve working conditions have been wildly successful so far – they just haven’t led to more enterprise-based bargaining, as contemplated by the Wagner model of labor law.
Where does this leave unions? The pessimistic story is that private employers – aided by an anticipated spate of favorable NLRB decisions – will continue to rebuff union drives, Republicans will continue to weaken public sector unions in states they control and in the federal government, and the Supreme Court will eliminate mandatory dues or fees in public sector unions nationwide. And it goes nearly without saying that prospects for creating conditions in which unions could thrive – for example, by passing legislation to facilitate sectoral-level bargaining – went from “very unlikely” to “completely DOA” with the 2016 election. But Kate Andrias develops a more optimistic story in her pathbreaking article, The New Labor Law. She focuses with precision on exactly how unions have managed to achieve their recent successes, revealing that, in a sense, sectoral and regional bargaining are actually already here, and they can be nurtured and propelled at the state level.
How could a new model of union representation arrive without anyone noticing? The answer seems to lie in our tendency to place labor law and employment law in separate silos. Consider the process by which New York State raised wages for fast food workers or Seattle raised its minimum wage. In the New York example, public protests underwritten by the Service Employees International Union led Governor Andrew Cuomo to convene a wage board, comprised of representatives of management, labor, and the public. The wage board then heard testimony, weighed the evidence, and announced a decision phasing in a fifteen-dollar minimum wage. Similarly, the process by which Seattle raised its minimum wage entailed creating a task force on which both labor and business were represented. The task force hammered out a compromise proposal, which was ultimately enacted by the City Council. In both instances (and many others, which Andrias documents), union representatives literally bargained with management representatives over conditions for large numbers of workers – though the final result then had to be ratified by government. Andrias calls this tripartite process “social bargaining,” and she argues that it could be replicated: “Under the emerging model, employment law is no longer just a collection of individual rights to be bestowed by the state. Instead, it is a collective project to be jointly determined and enforced by workers, in conjunction with employers and the public.” (P. 68.)
In addition to identifying the practice of social bargaining in the United States, Andrias argues that it could lead to two results: first, improving union prospects for traditional collective bargaining for working conditions above the minimum guaranteed by law, and second, creating new forms of labor organizations and funding mechanisms. The most concrete manifestation of the first can be seen in minimum employment standards that contain a carve-out for unionized employers – as Andrias notes, a practice that can generate controversy. But beyond that, Andrias argues that growing political support for better pay and benefits can soften employer resistance to offering higher wages at the bargaining table. As to union funding mechanisms, Andrias’s possibilities are less concrete, but they include possible fee-for-service models (which the NLRB could facilitate), as well as government funding of union programs that benefit non-union workers.
As Andrias acknowledges, finding a way for unions to pay for social bargaining is perhaps the greatest challenge to the model’s success. Unions’ loss of density and political power have gone hand-in-hand – they are a vicious cycle. Movements like Fight for $15 have shown that labor can still deliver the goods for low-wage workers, but they are expensive, and dues-paying members may not be willing to fund them indefinitely. Whether labor can solve the funding problem while fighting off new threats from the federal government will likely be the single largest determinant of whether social bargaining proliferates. At the same time, the lack of a long-term funding mechanism is not the only challenge to social bargaining – Andrias also points to legal threats, including federal or state preemption of local laws.
Over the course of the next four years, we are likely to see manifested two diametrically opposed visions for improving the lives of low-wage workers – a deregulatory one in red states and cities, and a social-bargaining-based one in their blue counterparts. Andrias’s article offers scholars and other commentators a framework for thinking about the latter, and in doing so, performs a great service.