Uberizing Discrimination

1. Introduction

  What does Uber-like work mean for the future of employment discrimination law?  Litigation over the status of Uber drivers – whether they should be considered employees or independent contractors for purposes of the wage and hour laws – has been closely chronicled in the press, and the issue has received some attention from employment law scholars. But almost no consideration has been given  to the question of discrimination in the hiring and firing of gig workers, and the effect of “customer preference” on the terms of their employment.  While this may not be a critical issue when it comes to taxi drivers, it has major ramifications as Uber-like platforms – for example, Task Rabbit, Woof, Upwork, Homejoy, BlueCrew, HourlyNerd, Wonolo, to name just a few – take over many service-related and professional employment opportunities, where employment discrimination indisputably has a significant impact. What protections can be afforded these workers, as compared to those available in traditional employment settings? What has been variously called Uberization, the sharing economy, the on-demand economy, the gig economy, and the permanent temp economy represents a sea change in how we think about work.   According to one study, by the year 2020, more than 40% of the United States workforce, some 7.6 million, will be so-called contingent workers — freelancers and temp workers, doubling the current total of 3.2 million.   It is estimated that almost 80% of those workers will be part-time. For some portion of these workers, wage and hour regulation will not be a significant concern because they perform white collar services at a level of compensation well beyond minimum wage.  But all of these workers face the possibility of discrimination on the basis of protected classifications: race, sex, national origin, religion, disability, age, and sexual orientation, at least in some jurisdictions.  In fact, Uber-like platforms may encourage discrimination, since they typically provide the ultimate customer/employer with a prospective worker’s name and photograph, and in some cases age, thereby allowing for the possibility of what may be either explicit or implicit bias. To understand how discrimination plays out in the gig economy, it is helpful to briefly describe the common features of Uber-like platforms.  These smartphone applications are designed to match persons or businesses who need a particular service [exclude goods like airbnb] with available workers having the requisite level of skills.    Uber itself, which calls itself a “technology company,” as opposed to a “transportation company,” provides  the simplest example.   Drivers must pass a screening process and background check, then are interviewed, and their own vehicles inspected.   If they are accepted, they sign a contract stating that they are independent contractors. They are matched through the Uber algorithm with those needing transportation.   Once the match is made, a photograph of the driver appears on the rider’s cellphone; the rider can cancel the ride without charge within a short time period. When the ride completed, the driver is credited with a proportion of the fare set by Uber, which retains the remainder of the cost.  The driver’s performance is monitored through customer ratings of one to five stars, and drivers can be discipline or terminated if their rating are not satisfactory. These ratings are a serious matter. As one journalist noted, “the penalty for falling below a certain rating and into the ‘death zone’ is something that’s appeared in virtually every news story about Uber …: deactivation . . . Uber can cut off access to this app at any moment for any number of reasons, reasons that haven’t been made entirely clear.”  The “death zone”  seems to be anything between a 4.4. and 4.7 rating.  Thus, what we typically think of as at-will discharge is a feature of the Uber platform. The possibility of discrimination in this model is obvious at several intersections, but perhaps not highly likely. At the initial stage of the driver/company relationship – what we think of as hiring – it is conceivable that Uber would be disinclined to accept drivers who are female, older, disabled, or even of certain ethnicities.  In the customer/driver relationship, explicit or implicit bias may find its way into how riders allocate their ratings, giving free rein to “customer preference,” which Title VII and its sister statutes prohibit as a justification for disparate treatment.  As noted above, the ending of the driver/company relationship seemingly can occur without explanation. But with Uber, the interaction between driver and rider is fleeting and relatively impersonal. It is unlikely that riders would actually go to the trouble of rejecting an available driver because of bias stemming from viewing his photograph. But let us consider the Uber model applied to services of a personal or professional nature.  According to its website, Zeel “delivers a top-quality massage to your home, hotel, workplace or event in as little as an hour, or at the time you want.”  Its massage therapists are “licensed, screened and fully vetted by the Zeel Team.”   For those interesting to working with Zeel, the website states that its therapists earn “on average” 75% of the total cost of the massage, and “you make your own schedule.”  Like Uber drivers, therapists sign into the app when they are available and log off when they are not.   A customer booking an appointment may request a male or female therapist [probably ok under Title VII privacy], and once the match is made, Zeel sends the customer “a text with the bio and photo of your therapist.” The customer may cancel the appointment within 10 minutes without charge, and rate the therapist, also using the 5-star system. HourlyNerd is a business consulting application that matches businesses with consultants and only accepts “professionals with MBAs from top 40 global programs,” with an average of 8+ years of work experience.”  Clients post a project, ranging from researching a market to starting, growing, managing, or selling a business.  HourlyNerd “experts,” who “generally work from $75 – $200 /hour” per hour, bid on the project, or they are matched by an algorithm and invited to bid.  Clients select a consultant after reviewing the resumes, including photographs, of those who have applied.  The fee is collected by HourlyNerd, but the funds are held until the project is complete and the client is satisfied.  Again, the “nerd” is rated by the client after completion of the project. The on-line service contract makes clear the HourlyNerd is not an employer, and the workers are independent contractors: The intent is that Contractors will be properly classified as independent contractors of Client and Client agrees (a) that Client does not in any way supervise, direct, or control Contractor’s work, (b) that Client does not, in any way, supervise, direct, or control Contractor’s work hours and location of work, and (c) Client does not provide Contractor with training or equipment needed for any Contract. Notwithstanding the foregoing, Client assumes all liability for proper classification of Contractors as independent contractors or employees based on applicable legal guidelines. Given the structure of these platforms, explicit bias is given free reign:  customers can reject potential workers on the basis of their gender, name or photograph. If a client doesn’t want to have a Black massage therapist, for example, he need only cancel and then reschedule the appointment.  But implicit bias – the unconscious mental processes which cause us to act upon negative stereotypes of stigmatized groups –  is even more problematic. The effect of implicit bias in the workplace and in society generally recently has been the subject of much popular attention, as well as theoretical and empirical research. It will not be explored in depth here, but a summary of a few studies will give a flavor of the potential for implicit bias to infect the gig economy.  In one experiment, two sets of identical resumes – one with “white-sounding” names and one with traditionally Black names were submitted in response to employment advertisements across a range of occupations and industries. Those with white names received 50% more callbacks for interviews.  Another study used a similar methodology to explore bias in the consideration of applicants for a lectureship at a British university, and found that white participants were ten times as likely to choose white applicants over Blacks or ethnic minorities with identical cv’s. Implicit bias is generally most pronounced when quick decisions are made. Therefore, Uber like platforms may be particularly susceptible to its effect.  For example, two Stanford economists found that online shoppers were significantly more likely to buy an unopened new model IPod when a photograph showed it being held by a white hand, as compared to a dark-skinned hand. Black sellers were also offered less money for the item. Perhaps the response to explicit or implicit bias in the gig economy should be, “so what’?  It might be argued that these platforms do nothing more than provide a technological assist to business arrangements that have always functioned on the independent contractor model.    Some massage therapists and some MBAs are salaried employees and are entitled to protections against discrimination. Those who choose to work for themselves are left to their own devices. Their clients select them at their complete discretion, and may exercise any biases or prejudices they may hold. But this view of employment relationships ignores important considerations. First, if statistical projections bear out, it will not be long before half or more of the US workforce is trying to make a living in the gig economy.  It seems contrary to the goals of our anti-discrimination statutes to leave so many workers without protection. Second, while they may fall within the legal definition, many gig workers are not truly independent contractors, in the sense that they pick and choose among clients and work at their own pace.   Despite the hype put forward by the “technology” companies, a good portion of these workers have as their goal to make a full-time salary with one company, and are in most respects indistinguishable from salaried employees. To include some or all gig workers within anti-discrimination protections may require statutory reinterpretation or amendment, or less drastically, a more generous reading of current doctrine,  issues that will be explored in this article. However, it is worth noting that not every problem necessarily requires a legal fix. In fact, as Susan Sturm has written, what she calls “second generation” discrimination may best be remedied with a problem solving approach at an organizational level. Here, a technological solution could guard against implicit bias at least at the customer/worker juncture. Photographs are ubiquitous in Uber-like apps — for no legitimate reason.  Why does a customer need to a picture, or a real name, of the TaskRabbit worker who is coming to fix his toilet?  Security concerns could easily be addressed by passwords or codes. In fact, “blind” hiring in traditional employment contexts is a burgeoning method of dealing with implicit bias, and was the subject of an extensive New York Times magazine article, showcasing software that scrubs ethnicity, age, and gender information from resumes before they are vetted for interviews. Scrubbing such identifying information from Uber-like platforms is a reasonable means of insulating workers from implicit bias.  Airbnb recently announced an analogous effort to prevent discrimination in rental accommodations. Nonetheless, litigation over discrimination in gig employment undoubtedly soon will arise, and this Article addresses the issues that will come to the fore.  Part I explores a bit of labor history to contextualize the gig economy’s relationship to regulation. In Part II, I consider the question of who is an employee under Title VII and its companion discrimination statutes.  Part III looks at the recent ongoing wage and hour litigation over the status of Uber drivers, and analyzes how those cases may impact discrimination claims. In Part IV, I examine the handful of discrimination cases brought by gig workers to date.  Part V discusses three theories under which gig employers might be brought under the antidiscrimination statutes: the relevance of Section 1981 as a strategy to protect some classes of workers and how its utilization may serve as a vehicle for redefining “employee” in other contexts; the possibility of considering these employers as “employment agencies” under Title VII; and the application of the Sibley doctrine, which bars discriminatory interference with access to jobs by non-employers.  Finally, I conclude by developing a taxonomy of worker/company/consumer relationships that can be used to determine who should be protected by anti-discrimination statutes.