The sharing economy inspires little consensus. Indeed, the very idea that there is a sharing economy, as well as the claim that it has anything to do with sharing, are both deeply contested.1 What most commentators do agree on is that the sharing economy is “disruptive.”2 For incumbent actors like taxi companies and hotels, the sharing economy disrupts longstanding licensing and employment models.3 For local governments, the sharing economy disrupts regulatory networks, infrastructure requirements, and tax income (although many cities have viewed these disruptions in a positive light).4 The sharing economy has even disrupted scholarship—not so much by offering a new way to learn (although it does this too5) but by making it difficult for academics and policy researchers to use traditional analytical categories in describing the size, shape, and impact of this new domain.6 And, of course, for the individuals who actually provide the services marketed by companies, the sharing economy disrupts existing notions of what it means to work.

But for all the many ways in which the sharing economy is disruptive, there is one very important way in which it is not: it does not fundamentally disrupt the legal infrastructure governing labor and employment in the United States. This is a much narrower claim than it may appear to be at first glance. Does the sharing economy “challenge our fundamental assumptions about employment types”?7 To be sure. Is it true that “we may need to construct platform-specific regulations”?8 Probably. It’s even fair to say that “employment in the sharing economy is just plain different,” insofar as it is characterized by a pace, level of flexibility, and entry requirements (among other things) that distinguish it from industrial wage labor. But it does not upend the basic structure of our labor and employment law.9

This is no great thing. It might have actually been better if the sharing economy had truly disrupted the building blocks of our work law, because then we would have had to reevaluate a system that has created great inequality right alongside great opportunity. That kind of disruption might have pushed us beyond the tiresome yet unavoidable issue of misclassification and into a world where there are more than two buckets into which we must pour all working individuals, or at least a world in which one’s bucket does not determine one’s access to the benefits and protections that are instrumental to a decent life.10

This type of disruption has not happened in large part because the labor and employment changes wrought by the sharing economy—important and intriguing though they may be—are of degree rather than of kind. Foremost among these changes is the extent to which sharing economy companies behave like employers while disclaiming the obligations attached to an employer-employee relationship—essentially, the way in which they function like invisible bosses.11 Yet this ostensibly unique feature of the sharing economy is simply the latest instance of workers being meaningfully controlled by someone who is not legally their employer: franchise workers and many independent contractors have long experienced a similar kind of invisible authority.

Because the idea that sharing economy companies operate as invisible bosses is central to many critiques of this new approach to labor exchange, Part I begins by explaining just what it is about their authority that makes it “invisible.” Part II extends this discussion to two earlier developments that, like the sharing economy, also significantly transformed the way Americans work: the franchise explosion of the 1950s and the spread of the independent contractor model in the late twentieth century. This article is the first to offer a detailed comparison of work practices used by sharing economy companies, franchises, and some independent contracting companies to discuss how these entities seem like invisible bosses.12 I conclude with a few thoughts on what the comparison tells us about regulating labor in the sharing economy.

  • 1. For terminology, see, e.g., Tom Slee, What’s Yours is Mine: Against the Sharing Economy 11–12 (2015); Arun Sundararajan, The Sharing Economy: The End of Employment and the Rise of Crowd-Based Capitalism 26–27 (2016); Sofia Ranchordas, Does Sharing Mean Caring? Regulating Innovation in the Sharing Economy, 16 Minn. J.L. Sci. & Tech. 413, 416 (2015); see also Deepa Das Acevedo, Regulating Employment Relationships in the Sharing Economy, 20 Emp. Rts & Emp. Pol’y 1, 3–14 (2016) (distinguishing between traditional but technologically-mediated “renters” like Rent the Runway and ZipCar, non-monetized forms of exchange that are facilitated by “swappers” like Craigslist, and true “platforms” like Uber, Airbnb, and Feastly).
  • 2. See, e.g., Orly Lobel, The Law of the Platform, 101 Minn. L. Rev. 87, 91 (2016) (arguing that platforms present “not only a paradigmatic shift for business, but also for legal theory”).
  • 3. These models have themselves experienced quite a bit of change in recent decades. V.B. Dubal, Wage Slave or Entrepreneur: Contesting the Dualism of Legal Worker Identities, 105 Cal. L. Rev. 65, 70 (2017) (observing that by the 1980s taxi drivers had gone from being largely unionized drivers to being independent contractors who leased medallions, and arguing that this shift in the taxi industry as well as the expansion of ridesharing companies “reflects a particular idealization of the ‘entrepreneur’”).
  • 4. See generally TransitCenter, Private Mobility, Public Interest: How Public Agencies Can Work With Emerging Mobility Providers 11 (Sept. 8, 2016),, (discussing both challenges and opportunities facing the public transit agencies that provide traditional fixed-route transportation in light of “emerging mobility providers in the private sector”); Lauren Hirshon et al., Nat’l League of Cities, Cities, The Sharing Economy, and What’s Next 1 (2015),, (arguing that “cities make the sharing economy work” but noting that “the unanticipated surge in sharing economy business models and the proliferation of companies that serve as catalysts for collaborative consumption has created a disruption of existing systems”). See also Nestor M. Davidson & John J. Infranca, The Sharing Economy as Urban Phenomenon, 34 Yale L. & Pol’y Rev. 215 (2016); Roberta A. Kaplan & Michael L. Nadler, Airbnb: A Study in Occupancy Regulation and Taxation, 82 U. Chi. L. Rev. Dialogue 103 (2015). On city governments’ positive attitudes toward the sharing economy, see the example of Baton Rouge, Louisiana, which “collectively pounced on a stray suggestion in a media report that the transportation networking company Uber was considering coming to town” and passed an ordinance legalizing ridesharing within a few weeks of that report. Uber Invasion–Baton Rouge Joins a Global Evolution in Ground Transportation, Business Report (Aug. 6, 2014),,
  • 5. See, e.g., the peer-to-peer learning platform Skillshare. Terms of Service,,,
  • 6. Josh Wright, Economic Implications: Sharing Means We’re Wealthier Than We Think but May Grow Slower Than We Want, in Bloomberg Brief: The Sharing Economy 5 (June 15, 2015) (observing that “the sharing economy has muddied existing measures of employment status, labor force participation and wages earned”).
  • 7. Das Acevedo, supra note 1, at 28.
  • 8. Id. at 29.
  • 9. Id.
  • 10. Katherine V. W. Stone, Legal Protections for Atypical Employees: Employment Law for Workers Without Workplaces and Employees Without Employers, 27 Berkeley J. Emp. & Lab. L. 251, 279 (2006) (“[I]n the United States there are only two categories—employee and independent contractor—where the former receives some employment law protections and the latter does not.”).
  • 11. See, e.g., Jeremias Prassl & Martin Risak, Uber, TaskRabbit, and Co.: Platforms as Employers? Rethinking the Legal Analysis of Crowdwork, 37 Comp. Lab. L. & Pol’y J. 619, 635 (2015–16) (advocating an “openly functional” way of identifying employers that emphasizes five types of functions and arguing that “some crowdwork platforms, such as Uber, exercise their full range”); Alex Rosenblat & Luke Stark, Algorithmic Labor and Information Asymmetries: A Case Study of Uber’s Drivers, 10 Int’l J. Communication 3758, 3759 (2016) (“examining how Uber drivers experience labor under a specific regime of automated and algorithmic management”).
  • 12. Like this paper, the following works also depart from the conventional focus on “employees”: Prassl & Risak, supra note 11; Jeremias Prassl, The Concept of the Employer (2015); Mitchell T. Rubinstein, Employees, Employers, and Quasi-Employers: An Analysis of Employees and Employers Who Operate in the Borderland Between an Employer-and-Employee Relationship, 14 U. Pa. J. Bus. L. 605 (2011–12); Katherine Hannan Wears & Sandra L. Fisher, Who is an Employer in the Triangular Employment Relationship? Sorting Through the Definitional Confusion, 24 Emp. Resp. & Rts. J. 159 (2012). Additionally, others have drawn comparisons between the sharing economy and franchising for the purposes of developing a general framework for regulating disruptive business models. See Eric Biber et al., Regulating Business Innovation as Policy Disruption: From the Model T to Airbnb, 70 Vanderbilt L. Rev. 1561, 1596–1602 (2017) (discussing the franchisor-franchisee relationship as a form of business innovation comparable to the rise of the sharing economy).

I.     Platforms as Invisible Bosses

Part of the difficulty in talking about something as new and rapidly changing as the sharing economy is the task of identifying and naming the object of discussion.13 The title and introduction of this paper used the term “sharing economy” because that is still the most widely recognized name for companies like Uber, Airbnb, and Feastly. Going forward, I will use the term “platform” to describe a subset of companies within the broader sharing economy that actually present work regulation issues.14 Platforms present work law concerns because they actively participate in the transactions they give rise to by shaping the behavior of providers and consumers alike. Moreover, they substitute themselves (albeit to varying degrees) for existing government safeguards.15 Because all of this is achieved via web or smartphone platforms, I use the word “platform” as a kind of proxy for the behaviors that make these companies unique within the sharing economy.

What features of platform technology allow companies like Uber and Handy to act like invisible bosses with respect to their providers? Just about all of them, as it turns out. The rest of this section considers four specific examples of invisible authority: rating systems, vetting and termination, real-time tracking, and brand management. Needless to say, this list is not exhaustive and it is sometimes hard to tell where one set of practices ends and the other begins.16

A.    Reputational Feedback Systems

Reputational feedback systems have emerged as a major point of contention in conversations about the platform economy, and with good reason: the systems effectively communicate the type of data-driven, responsive, community-based image that platforms market to consumers and providers alike. For example, Airbnb notes that “[g]enuine reviews are the cornerstone of our community,”17 Lyft says that its “two-way rating system helps ensure the safety and comfort of the Lyft community,”18 and Handy observes that since “none of this works without putting trust, safety, and security at the forefront of every decision” providers are “rated by other customers using Handy.”19

Reputational feedback systems have also been central to arguments that the platform economy is a naturally self-regulating ecosystem with clear historical analogies (the Maghribi traders of the eleventh century Mediterranean are a popular comparison).20 But as a growing body of analysis is beginning to suggest, the efficacy of platform feedback systems—and, by extension, the degree to which they can be counted on to realize the ends of regulation—is none too certain. The systems generate both artificially inflated ratings (when consumers hesitate to punish bad providers) and inaccurately low ratings (when consumers “spite grade” good providers because of factors unrelated to the provider’s service).21 Even more worryingly, consumer prejudices may drive feedback systems to “hardwire discrimination into the supervisory techniques” used by platforms.22

Perhaps a better way of thinking about all of these accuracy problems is that platform feedback systems are vulnerable to distortions caused by real-life power dynamics.23 This doesn’t make reputational feedback an unusually weak way of monitoring bad behavior; in fact, many of the problems facing reputational feedback likely also affect the “people analytics” used by large companies to predict and evaluate job performance in the conventional labor force.24 But supporters of platforms are deeply committed to the idea that reputational feedback is unusually effective because it relies on algorithms and self-policing.25 That simply isn’t the case. Maghribi traders notwithstanding, a regulatory system does not punch above its weight simply because it operates using technologically-mediated peer feedback.

What often takes a backseat to accuracy and fairness concerns is the degree to which reputational feedback systems mask active platform participation in consumer choice. Feedback systems don’t consist of raw information passed whole cloth on to potential riders or diners. To begin with, scores are averaged and rounded in order to present consumers with an easily digestible four- or five-star rating. This seemingly insignificant “packaging effect” is actually quite powerful: a 2012 study of Yelp reviews showed that a half-point difference on a five-point scale (due to the website’s algorithm rounding up or down from the actual score) made restaurants nineteen to twenty-one percent more likely to sell out prime time tables, even though the original un-rounded difference in scores was a mere 0.02 out of 5.26

Average scores are also not always just average scores: they are often average scores of a subset of information that platforms classify as usable inputs. Airbnb, for instance, calculates response rates (the number of inquiries a provider replies to) and response times (the time taken to send each reply) using all inputs from the previous thirty days except when a host has received fewer than ten inquiries during that thirty-day period.27 Similarly, Uber reassures drivers in some markets that it will not rely on reviews earned when surge pricing is in effect.28 The fact that platforms average consumer ratings becomes even more important in cases where they also rank providers by averaged score or, as with TaskRabbit’s revised format, when they suggest a limited set of providers based on those ranked averages combined with the consumer’s criteria.29

Averaging, input determination, and ranking are design features rather than design flaws, and from the perspectives of both consumers and platforms they are vital to the success of platform-based transactions. Consumers get up-to-date, peer-sourced, easily digestible information, while platforms—besides reassuring consumers—add to the growing mountain of data that helps them refine their business models. What’s more, none of these features necessarily work to the disadvantage of providers, although it’s not hard to imagine situations in which they might. What these practices do demonstrate, however, is that platforms are deeply involved in shaping and managing the very reputational systems that are supposed to offer unvarnished peer-to-peer feedback in a marketplace driven only by consumer demand.

B.    Vetting & Termination

Platforms, like conventional employers, vet aspiring workers. Admittedly, their application processes are often minimally selective: journalists report that “[Uber will] pretty much take anyone”30 or that “none of these food-sharing apps asks for training or experience.”31 Critiquing admission standards is sometimes useful as a counter-narrative to specific marketing claims, but it distracts from the fact that most of the real vetting occurs after providers have joined a platform. That is to say, platform vetting is real, but it consists of performance cutoffs rather than admission cutoffs.

One common performance cutoff is a star requirement applicable to all providers; for example, Lyft tells drivers that “If your rating drops below 4.8, you might want to start thinking about what you can do to improve it, since consistently low ratings can put you at risk of deactivation.”32 Because star ratings systems are highly inflated and often directly tied to termination, they can trigger anxiety in all parties involved: providers worry that the ratings systems are opaque, while consumers don’t want to be responsible for initiating a termination.33

Another popular performance cutoff is a “ratings over time” metric that allows some providers to market themselves as elite workers: Airbnb has “SuperHosts;”34 TaskRabbit has “Elite Taskers;”35 and Fiverr has both automatically granted elite statuses (“Level 1” and “Level 2”) as well as a discretionary elite status (“Top-Rated Sellers”).36 These elite statuses not only attract more consumers but also often come with special privileges like the ability to offer tailored services or charge higher fees, thereby compounding the effect they have on consumer demand and providers’ earning capacity.

As with reputational feedback systems, vetting and termination procedures are neither inherently nor unalterably bad for providers. And, also like feedback systems, performance cutoffs matter because they constitute forms of platform authority that operate relatively unseen: they are neither arbitrary nor inevitable reflections of a marketplace reality. Rather, they reflect “managerial decisions” on the part of platforms that feel the need “to manage their communities” by vetting and sometimes terminating those who seek to belong.37

C.    Real-Time Tracking

Many platforms actively monitor, analyze, and try to improve the means by which providers meet performance cutoffs using real-time tracking.38 Consider the “time to respond” metric: most platforms record, weigh (for elite status purposes), and report to the provider how long it took her to answer a consumer request for services.39 A more striking example of tracking comes from Uber, which in 2016 announced plans to monitor the braking and acceleration patterns of its drivers in order to analyze and improve their driving style.40 In the same year, the pet care platform Rover introduced the “Rover Card”—a detailed, real-time electronic report that workers can send to dog owners with descriptions of the dog’s behavior, a visual map of the actual walk route, exact start and stop timings for the worker’s visit and, of course, photos.41 Rover Cards are currently optional and it is not clear how they will affect a provider’s earning ability, but the company has indicated that the cards will soon become mandatory.42

Real-time tracking scores are usually reported to the provider along with tips, warnings, or reprimands.43 While the tracking practices that produce these scores may not involve over-the-shoulder supervision by a human being, they are no less meant to measure and shape the way a provider does her work. Most importantly, tracking metrics depend on input categories that platforms themselves construct and thus reflect platforms’ efforts to fulfill the quintessential managerial function of encouraging better work styles.

D.    Brand Management

Branding is surprisingly absent in conversations about the platform economy.44 No doubt this is partly because employers outside the platform economy have long maintained—and courts have long accepted—that brand management strategies do not speak to the existence of employment relationships even when those strategies place significant constraints on working conditions.45 It is also because platforms themselves have been at pains to emphasize that they are not selling anything beyond their technology, and that branding is simply something that each “partner” undertakes on her own by providing excellent service. Feastly, for instance, argues that its providers are “creatives and artists” and that the Feastly platform is simply about “appreciating what they’re making, and allowing them to share their work in an easier way.”46 Indeed, many providers view themselves as creating a personal brand for their platform work. Internet chat forums for Airbnb hosts are full of advice—“[g]uests love to see the towels with a chocolate on it”—on how to curate a special experience for consumers that will help consolidate a host’s personal brand.47 What’s more, some providers use their platform work to advance their other business ventures.48

Finally, branding is probably under-discussed because the sheer range of options consumers encounter makes it difficult to see platforms as constructing the kind of uniform experience or product we expect of a single brand. Not only might a consumer on Airbnb choose between a tree house or the equivalent of a five-star suite, she might also decide between a simple set of clean sheets and towels or a twenty-four-hour check-in, hot breakfast, and tiny organic toiletries.49 The combinations and permutations are often vast, as platforms quite rightly point out, and this heterogeneity is largely due to the fact that many of the most prominent platforms sell services rather than goods.50

Nevertheless, providers are still crucially involved in building the platform’s brand, and suggesting otherwise flies in the face of how marketing experts, especially within the sub-field of services marketing, have come to think of the way branding works. Services, unlike goods, are inherently not the sole creation of a firm because they are characterized by intangibility, perishability, inseparability (between the services on the one hand and the provider and consumer on the other hand), and heterogeneity (due to human involvement on both sides of the transaction).51 This does not mean that firms cannot or do not try to construct brand identities for the services they offer. Rather, brands are now “dynamically constructed through social interactions” rather than being embedded in something a company produces.52

The fact that there is no standard Uber car ride or Airbnb homestay does not mean that Uber and Airbnb lack either recognizable brands or strategies to curate brand value. On the contrary, curating brand value is exactly what metrics, guidelines, and cutoffs are meant to do. Likewise, the participation of individual workers in creating brand value does not negate the firm’s role in constructing the brand. Feastly may not cook a meal for you and stick a label on it, but Feastly, its providers, and its consumers are constantly engaged in the process of defining what it means to have a “Feastly meal.” In an era of services-dominant marketing, workers “shape and represent the brand promises made to external customers” by virtue of their power over consumer experience.53

Indeed, platforms—far from being simple matchmakers—are actually services marketers extraordinaire. One of the hallmarks of services marketing is that worker satisfaction or perceived satisfaction is integral to consumer appeal. Services marketers view their workers as “internal consumers” of the brand, a term that jibes oddly but strikingly well with Uber’s documented views of its drivers.54 Advertising campaigns by services marketers usually feature consumers and workers jointly experiencing the magic of the brand instead of exclusively focusing on consumer satisfaction, and this is absolutely true of platforms as well.55

With respect to branding (as with so much else) platforms reveal tensions or weaknesses in our work law—here, the illogicality of excluding brand management practices from worker classification analysis—that other business models also made apparent, albeit in less stark a fashion. Virtually all of the mechanisms described earlier in this section—reputational feedback, real-time tracking, and vetting and termination—help build brand identity, either internally among providers or externally among consumers. And in doing all of this while disclaiming an employment relationship, platforms merely follow in the footsteps of two earlier approaches to exercising invisible authority in the workplace.

II.    Invisible Authority outside the Platform Economy

The easiest way to see that platforms do not profoundly disrupt American work law (and should not be treated as if they do) is by comparing them with earlier corporate forms that also operate as invisible bosses. This section considers two such forms: franchises and some businesses constructed around the independent contractor model. Like platforms, these types of companies have attracted media and regulatory attention in recent years because of the ways in which they structure and characterize relationships between parent companies and individual workers.56

A.    Franchising

Although franchising is considerably older than the platform economy, it involves many similar and legally invisible forms of behavioral control over workers. Business-format franchising has been around since at least the 1920s, when companies began to pre-package their corporate functions and processes for sale as complete “business formats.”57 This turnkey approach to business formation took off in the post-war era as veterans returned in search of jobs and as attitudes toward mass consumerism began to shift.58 The United States went from having 50,000 franchisees grossing $2.5 billion in 1955 to 670,000 franchisees grossing $90 billion in 1970 and around 782,000 franchisees grossing $523 billion in 2015.59

Franchise lawyers sometimes complain that, despite this fairly long and strong history, courts only began to understand the unique characteristics and needs of the business-format enterprise in the 1990s.60 Instead of realizing that business-format franchising is “a multitier marketing device” and accepting that “a franchisor should be permitted to retain as much control as is necessary to protect and maintain its trademark,” franchise lawyers argue that courts and regulators “have not always understood and, despite the passage of time, still do not universally appreciate the control-like features inherent in franchising.”61 Indeed, control is so important to franchising that the franchise system has been called—appreciatively—“a business form that borders on a dictatorship.”62

High among the list of franchisor grievances is the sense that franchisors are caught between a rock and a hard place because they must simultaneously exert control over distant workers and avoid the employment obligations that usually attach to control in order to derive value from licensed trademarks.63 After all, business-format franchising is built on the premise that non-unitary corporate actors can present a unitary appearance and experience to consumers, and consequently, a franchisor’s need to preserve uniformity extends to virtually anything a consumer might associate with the trademark.64 Precisely because of this, protecting a trademark like the McDonald’s Golden Arches can never just be about monitoring how the bare symbol itself is used: it must also be about brand management.65

McDonald’s efforts to protect its trademarks and brand value are a good example of how franchisors have come to exercise considerable yet legally invisible authority over the working conditions of their franchisees’ direct employees. Like many franchisors, McDonald’s often facilitates the hiring process for franchisees by listing, describing, and processing employment positions via a centrally-managed website.66 Once they are hired, McDonald’s software schedules and occasionally tracks workers in real-time.67 Like the vetting standards and real-time tracking procedures used by many platforms, these actions allow McDonald’s to invisibly encourage uniform behaviors (although platforms, who always facilitate the “hiring” process between consumers and providers and whose real-time tracking involves fine-grained GPS monitoring, take matters a sizable step further).

Admittedly, the publicly available requirements flowing from McDonald’s to the franchisee and then to the worker are often articulated at a high level of generality—say, “a neat and clean appearance” or “competent and courteous service.”68 Consequently McDonald’s can argue that any specific requirements placed on in-store workers are due solely to the discretion of individual franchisees. But even if McDonald’s does not impose detailed requirements via the confidential Operations & Training Manual that is incorporated into every franchise agreement,69 its field inspectors—“Business Consultants” who conduct three different levels of review and “Mystery Shoppers” who engage in test transactions70—evaluate franchisees at a much higher level of granularity than can be found in the requirements publicly acknowledged by McDonald’s and analyzed by courts.71 Platform consumers and the behavioral suggestions that their reviews help generate serve much the same role within the sharing economy.72

For instance, McDonald’s training flyers, as well as the manuals that individual franchisees develop to meet the company’s requirements, are full of service speed recommendations: assembling and wrapping a burger should take about forty-five seconds, while a guest’s total experience time should not exceed three minutes and thirty seconds.73 Front-counter workers follow a precise code of behavior: the McDonald’s customer who orders, say, a Filet-o-Fish gets asked, “Will there be any fries today?” and experiences, theoretically, a worker who will “smile, establish eye contact and greet the guest using a pleasant tone of voice while being friendly and enthusiastic.”74 And because the look and taste of food items are precisely fixed by McDonald’s, the most minute details of product assembly inevitably become uniform across franchisees as well: buns are toasted for eleven seconds, coffee with sugar is stirred three to four times, and hamburgers get exactly one squirt of room temperature ketchup.75 When put together, and even when compared against the not-insignificant behavioral oversight exercised by many platforms, McDonald’s effective authority over its franchisees’ employees is mind boggling.76

Despite the fact that McDonald’s imposes a wealth of oversight via its inspections, its manuals and flyers, and nearly two thousand hours of uncompensated pre-approval training for aspiring franchisees, the company is mostly not vicariously liable as a joint employer of in-store workers.77 Courts often simply hold that McDonald’s “did not control the day-to-day operation of the franchise, it did not have the authority to hire or fire employees, and it did not own or operate the franchise at issue.”78 Sometimes courts work a little harder to reconcile a franchisor’s high need for operational control with its equally high need for legal distance. The “instrumentality” test used by several state supreme courts and at least one federal district court narrows the relevant type of control to authority over “the daily operation of the specific aspect of the franchisee’s business that is alleged to have caused the harm.”79 Similarly in Patterson v. Domino’s Pizza,80 the California Supreme Court held that joint employer status only applies where the franchisor retains or assumes a general right of control over human resources practices.81

Two recent developments have prompted widespread speculation that franchisor authority may incur legal recognition as employer authority. First, in Browning-Ferris82 the NLRB held that it would only require reserved authority over the terms and conditions of employment and that such authority, where exercised, could be exercised via an intermediary.83 While plaintiffs seeking to categorize franchisors as joint employers will still need to establish a common law employment relationship between a franchisor and in-store workers, the Board’s move away from actual and direct control is important enough to have generated every kind of reaction save a moderate one.84 Second, and virtually alongside Browning-Ferris, the Board’s General Counsel has pursued a series of unfair labor practice charges against McDonald’s and several of its franchisees on the grounds that McDonald’s is a joint employer of its in-store workers. Both of these developments are likely to be subject to change under the current Republican-majority Board.85

The franchise community’s reactions to Browning-Ferris and the consolidated McDonald’s cases are strikingly similar to the way most platforms have responded to calls for greater regulation and “employee” classification. Franchise advocates, like their platform counterparts, have stressed the unique nature of their business model,86 its facilitation of a quintessentially American vision of the good life,87 its status as an “evolutionary business response,”88 and the guaranteed destruction of that model if operational control is held to trigger employment obligations.89 Also like platforms, franchisors have stressed that they are not in the business of selling end products or services and consequently are uninvolved in managing the workers who directly facilitate transactions. In other words, franchisors join both platforms and some independent contractor companies in arguing that they are merely intermediaries connecting two autonomous economic actors.

B.    Independent Contracting

This section roughly traces the arc of independent contracting in the United States and uses examples from several industries and time periods to show how some companies exercise invisible authority over independent contractors. Not all companies that rely on independent contractors fit into this narrative, but the several that do raise concerns that are consistent with one another, with franchises, and with platforms.

Labor and employment scholars generally agree that independent contracting became a separate type of work relationship over the course of the nineteenth century.90 As industrialization shifted the bulk of work from within the home to increasingly centralized locations and as free labor ideology became more prominent, courts and legislatures struggled to reframe the rights and responsibilities triggered by work relationships.91 Parties requesting tasks were no longer masters exercising property rights and “domestic rule” over servants so they could no longer be indiscriminately held to account for the injuries suffered or caused by the individuals who worked for them.92 By the 1850s, tort law—specifically, workers’ compensation and vicarious liability—had begun to construct the new category of “employee” out of the old concept of “servant” as a way of assigning these obligations.93 And by the end of the century, courts had begun to distinguish heavily supervised “employees” from “independent contractors” largely, though not exclusively, based on the degree of control they were subjected to while at work.94

The distinction between independent contractors and employees gained more significance when New Deal legislation created entitlements and protections that were specifically tied to “employee” status.95 The same distinction acquired even more importance when Taft-Hartley underscored the exclusion of independent contractors and subsequent legislation like the Equal Pay Act and Title VII augmented the legal protections available, almost exclusively, to employees.96 By the middle of the twentieth century, work-related safeguards vastly outstripped what they had been when independent contracting first began to emerge as a cognizable legal category, but independent contractors were set firmly outside the scope of most of these safeguards.97 Around the same time—and especially after the “contingency explosion” of the 1970s onwards—independent contractors began to occupy a growing percentage of the American labor force.98

Many of these independent contractors were employees who had recently been reclassified by their employers. Taxi and trucking companies were among the first to act; many of them began to convert their drivers from employees into independent contractors in the 1970s, and taxis in particular moved quickly.99 Yellow Cab Company and Checker Taxi Company, which were owned by the same family and held nearly eighty percent of the medallions in Chicago, went from having no independent contractors in January of 1975 to leasing seventy-two percent of their medallions to drivers classified as independent contractors by November of 1977.100 Both companies fought unionization rights for the new lessee-drivers on the grounds that they were independent contractors operating outside the scope of the Wagner Act, but the NLRB sided with the drivers because, among other things:

[T]he cabs display the companies’ insignia, [hence] all the goodwill inures to the companies; the lessee drivers’ work is an essential part of the companies’ normal operations; the lease is short and renewable only at the companies’ discretion; sub-leasing is prohibited; like the regular employees, the lessees are subject to reference checks when they apply for a lease; the companies unilaterally determine whether drivers are at fault for accidents; the companies impose [a] 250-mile limitation; [and] the companies mandate dress restrictions . . . .101

The D.C. Circuit disagreed. The court held that since companies did not exert direct physical control over lessee-drivers, “the drivers essentially work for themselves and merely pay the companies for the service of providing the use of a cab and medallion.”102 Current debates over Uber and Lyft are little more than a technologically compelled re-visitation of the forty year old debate over taxi drivers’ classification, from the specific observations of the NLRB regarding driver restrictions right down to the exact reasoning of the D.C. Circuit.

Transportation may have been especially suited to reclassifying employees as independent contractors because it involves neither centralized performance nor over-the-shoulder supervision, but it has hardly been alone. During the 1980s, “construction workers, nurses and allied health workers, janitors, carpet layers, casual and temporary retail employees, industrial homeworkers, [and] forestry workers” were among the low-income workers who increasingly found themselves re-categorized as independent contractors.103 And in the 1990s, reclassification and the growing use of independent contractors spread to two industries at the heart of the contemporary economy: computer programming and rapid delivery services.

The late 1990s case Vizcaino v. Microsoft104 involved independent contractors who were deeply integrated into Microsoft’s daily operations.105 After the IRS determined that the contractors were in fact common law employees, Microsoft undertook a reverse conversion of sorts by transforming some workers into its own employees and others into employees of a temporary agency that it relied on for staffing.106 The Ninth Circuit’s three Vizcaino decisions “sent a lightning bolt through corporate America” because they threatened the usefulness of categorizing workers as independent contractors for the purposes of limiting employment related obligations.107 Speaking before the ABA’s committee on Employee Rights and Responsibilities a few years after Vizcaino I was decided, one attorney proclaimed that the ruling “raised fears that thousands of ‘contingency’ workers, heretofore excluded from benefit plans, might now be deemed covered by judicial fiat.”108

Another line of cases that began around the same time as Vizcaino but continues to be litigated some nineteen years later involves delivery drivers who work for various divisions of FedEx and one of its predecessors, Roadway Package System.109 In 1998, drivers working for Roadway in California won their bid to have the NLRB recognize them as employees with rights to unionize.110 However, similar claims filed over the next few years did not end as well for the drivers, particularly when the D.C. Circuit categorized drivers as independent contractors on the grounds that the drivers had significant “entrepreneurial opportunity.”111

Between 2003 and 2009, FedEx found itself fighting misclassification lawsuits in federal courts around the country that were eventually consolidated and heard before a multi-district litigation court in Indiana.112 All the multi-district plaintiffs moved for summary judgment, and most were denied.113 However, some claims were remanded to the Northern District of California and eventually reheard by the Ninth Circuit.114 In 2014, the Ninth Circuit rejected “entrepreneurial opportunity” as the key distinction between employees and independent contractors and, applying California’s “right to control” test, found that FedEx drivers were employees as a matter of law.115

The Ninth Circuit held that the written agreement between FedEx and the drivers as well as FedEx’s policies and procedures “unambiguously allow FedEx to exercise a great deal of control over the manner in which its drivers do their jobs.”116 Among other things, the court determined that FedEx “controls its drivers’ clothing from their hats down to their shoes and socks . . . [and] requires drivers to paint their vehicles a specific shade of white, [as well as] mark them with the distinctive FedEx logo.”117 Similarly, it decided that FedEx “can and does control the times its drivers can work [as well as] aspects of how and when drivers deliver their packages.”118 finally, the court specifically rejected FedEx’s argument that despite these “exquisite” forms of oversight, FedEx merely controls drivers with respect to contracted-for results rather than the means of performance.119 To be sure, these forms of oversight exceed the control that most platforms—even Uber—exercise over their workers.120 Nevertheless, the long list of specific vehicle models that Uber drivers may use and the specific verbiage they are instructed to employ or avoid similarly speak to the authority that platforms intentionally and invisibly exercise over their workers.121

The Ninth Circuit’s opinion is undoubtedly not the end of the FedEx litigation, much less the end of debates over worker misclassification. Several state legislatures have introduced, enhanced, or enforced anti-misclassification laws, and the Obama Department of Labor signaled its growing interest in the issue of worker classification via a series of white papers and collaborations with state agencies.122 The Internal Revenue Service and the NLRB have done likewise (although, perhaps predictably, all three entities advocate different classification tests).123 Even Congress has signaled an interest in classification, particularly as it affects the sharing economy.124 All of this regulatory activity underscores the extent to which misclassification triggered by “invisible authority” has raised concerns before and beyond the platform economy.


The comparison undertaken here between platforms, franchises, and some independent contractor companies has two implications for regulators seeking to determine the best way forward. First, and most importantly, it demonstrates that the platform economy is not fundamentally disruptive vis-à-vis our labor and employment infrastructure. On the contrary, it is simply the latest in a long line of business models where workers are meaningfully controlled by someone who is not legally their employer. In some instances, such as when they encourage preferred behaviors via algorithms instead of via mystery shoppers, platforms have simply found new, technology-enabled ways to replicate the kind of invisible authority exercised by earlier business models. In other cases, like when they use GPS tracking to monitor worker patterns instead of relying on a shift manager, platforms have arguably augmented or refined earlier approaches to invisibly exercising authority. Regardless, platforms have not disrupted our work law so much as they, like their predecessors, have learned to operate too well within it.

Second, by comparing platforms with both franchises and some independent contractor companies, this article has shown that the relationship between platforms and their providers has multiple antecedents across business models, industries, and time periods.125 It is hardly clear that any one industry or business model constitutes an ideal comparison for platform work; moreover, even if such a comparison exists—say, franchising—it very likely triggered and left unresolved the same sense of invisible authority that makes platform labor seem troubling today. Neither of the contexts considered here present success stories from which we can draw satisfying lessons.

When we look for analogies to help guide our way forward we should remember to ask two questions instead of the one we have been focusing on so far. It is certainly worth asking what feature of our work law—missing, misunderstood, or misappropriated as it may be—may have facilitated this latest way of invisibly exercising authority over workers. But a second question should be close behind, namely, whether the legal infrastructure governing work—problematic though it is—also speaks to cultural conceptions about the nature of work and the dynamics of control and freedom between employers and workers. That kind of inquiry demands that we look beyond the structure of work law and into the social processes that it both shapes and reflects.

  • 13. Adam Chandler, What Should the “Sharing Economy” Really Be Called?, Atlantic (May 27, 2015),, (“[I]t matters which term journalists, economists, and academics settle on, because it clearly can shape consumers’ perceptions of these businesses as they take shape.”); see also Kenneth Olmstead, How Americans Define the Sharing Economy, Pew Research Center (May 20, 2016),, (drawing on the results of a national survey to argue that “the public as a whole has a very tenuous grasp of what the sharing economy stands for” and that “[t]he most common description of the sharing economy emphasizes the ‘sharing’ component of the phrase while ignoring the ‘economy’ aspect.”).
  • 14. On the distinction between platforms and other types of sharing economy companies like “renters” or “swappers,” see Das Acevedo, supra note 1, at 9–12. Some other labor and employment scholars have also settled on the term “platform economy.” See, e.g., Lobel, supra note 2, at 4; Brishen Rogers, Employment Rights in the Platform Economy, 10 Harv. L. & Pol’y Rev. 479 (2016); Juliet B. Schor, Does the Sharing Economy Increase Inequality Within the Eighty Percent?: Findings from a Qualitative Study of Platform Providers, 10 Cambridge J. Regions, Econ. & Soc’y 263, 267 (2017) (using the term platform “to denote for-profit companies that use platforms and apps, crowdsource ratings and reputational data, and use digital technology to organize exchanges”).
  • 15. Das Acevedo, supra note 1, at 9–12.
  • 16. This is perhaps especially the case as between “vetting and termination” and “real-time tracking.” See infra Sections I.B, I.C.
  • 17. Airbnb used this language until recently. See, e.g., Irene, Comment to Guests Reviews, Community With (July 15, 2016, 4:55pm),, (quoting this language from Airbnb’s Privacy Policy); Gilles, Genuine Review, Community With (Feb. 19, 2016, 3:04 PM),, (ascribing this language to Airbnb’s “rules and regulations”). However, it now states that “Our community relies on honest, transparent reviews.” How Do Reviews Work?, Airbnb,,
  • 18. Driver and Passenger Ratings, Lyft,,
  • 19. Trust & Support, Handy,,
  • 20. Adam Thierer et al., How the Internet, the Sharing Economy, and Reputational Feedback Mechanisms Solve the “Lemons Problem” 12–13 (May 2015),,; Sundararajan, supra note 1, at 142–46. The original view of the Maghribi traders as a community that regulated itself using a “reputation mechanism” was put forward by Avner Greif. See, e.g., Avner Greif, Reputation and Coalitions in Medieval Trade: Evidence on the Maghribi Traders, 49 J. Econ. Hist. 857, 858 (1989).
  • 21. Erica Ho, Why You Should Think Twice Before Trusting Airbnb Reviews, Mashable (May 18, 2015),, (discussing grade inflation); Das Acevedo, supra note 1, at 19 (discussing “spite grading”).
  • 22. Noah Zatz, Beyond Misclassification: Gig Economy Discrimination Outside Employment Law, On Lab. (Jan. 19, 2016),,; see also Nancy Leong, New Economy, Old Biases, 100 Minn. L. Rev. 2153, 2154 (2015–2016) (“certain features specific to the sharing economy actually increase the potential racial discrimination”); Nancy Leong & Aaron Belzer, The New Public Accommodations: Race Discrimination in the Platform Economy, 105 Geo. L.J. 1271 (2017); Benjamin Edelman & Michael Luca, Digital Discrimination: The Case of (Harv. Bus. School, Working Paper No. 14-054, 2014). But see TED, How Airbnb Designs for Trust, YouTube (Apr. 5, 2016), (containing presentation by the founder of Airbnb arguing that 10+ reviews significantly alter platform users’ willingness to trust people unlike themselves).
  • 23. Bogdan State et al., Power Imbalance and Rating Systems, 2016 Proc. 10th Int’l Assoc. Advancement Artificial Intelligence Conf. Web & Soc. Media 368, 368.
  • 24. “People analytics,” or the “process or method of human resources management based on the use of ‘big data,’” operates on the idea that “unstructured subjective judgment is not rigorous or trustworthy as a way to assess talent or create human resources policies.” Matthew T. Bodie et al., The Law and Policy of People Analytics 3 (St. Louis U. Legal Studies Research Paper Series, Paper No. 2016-6, 2016). Both reputational feedback and people analytics run the risk of reproducing exclusive social patterns and both “could make masking intentional discrimination easier.” Id. at 51, 67. At the same time, the thinking behind reputational feedback is almost the opposite of people analytics: rather than assuming that “unstructured subjective judgment” is a poor basis for managerial decision-making, reputational feedback assumes that in sufficient quantities those same judgements can provide trustworthy ways to assess talent.
  • 25. See, e.g., Christopher Koopman et al., The Sharing Economy and Consumer Protection Regulation: The Case for Policy Change 2 (Mercatus Working Paper, May 2015) (on file with author) (“[T]he Internet, and the rapid growth of the sharing economy, alleviates the need for much of this top-down regulation, with these recent innovations likely doing a much better job of serving consumer needs.”).
  • 26. Michael Anderson & Jeremy Magruder, Learning from the Crowd: Regression Discontinuity Estimates of the Effects of an Online Review Database, 122 Econ. J. 957, 966 (2012).
  • 27. How are my Response Rate and Response Time Calculated?, Airbnb,,
  • 28. Alex Rosenblat et al., Discriminating Tastes: Customer Ratings as Vehicles for Bias, Data & Society 7 (Oct. 2016).
  • 29. Other platforms use this format too, and providers clearly understand the importance of being ranked highly on their platforms’ homepages. See, e.g., Nick Loper, How I Got on the Homepage of Fiverr and Earned $920 in 10 Days, Side Hustle Nation (Apr. 21, 2014),,
  • 30. David Fagin, Life as an Uber Driver: It’s Just Not Fare, HuffPost (Feb. 3, 2014),, (“Becoming an Uber driver is actually a piece of cake . . . as far as I can tell there’s no license check . . . get your vehicle’s paperwork and head to a nearby Holiday Inn . . . the cute 12-year-old girl running the show gives you your Uber’d-out iPhone 5 . . . and then says, ‘Next?!’”); see also Emily Guendelsberger, I Was an Undercover Uber Driver, Phila. City Paper (May 7, 2015),, (“[T]he application was just uploading my car’s information, banking details and my Social Security number for a background check.”).
  • 31. Samantha Melamed, Philly’s Newest Takeout Option: Food from a Stranger’s Kitchen, (Apr. 29, 2016),, (referring both to dine-out platforms like EatWith, Feastly, and Meal Sharing, as well as to take-out platforms like Homemade and new Philadelphia service Local Stove). Similarly, one journalist remarked in reference to the TaskRabbit introductory video that “all this” could be hers “with just a quick background check and the completion of a quiz with true or false questions.” Sarah Kessler, Pixel and Dimed: On (Not) Getting by in the Sharing Economy, Fast Company (Mar. 18, 2014),,
  • 32. Driver and Passenger Ratings, Lyft,,
  • 33. On the anxiety caused by ranking systems, see Erica Ho, supra note 21 (“Courtesy now dictates behavior and guests/hosts will often refrain from leaving a critiqued review unless it was just truly, truly an awful experience” because “bad reviews can influence future earnings or the ability to save some money.”); Kat Kane, The Big Hidden Problem with Uber? Insincere 5-Star Ratings, Wired (Mar. 19, 2015) (explaining that the author gave a five-star rating despite experiencing a “white-knuckle ride” because she felt it was inappropriate to “nitpick” as she “had been chauffeured door to door without incident—for half the cost of a cab and infinitely less hassle than the bus”—and because her driver now knew her home address); Alex Rosenblat & Luke Stark, Uber’s Drivers: Information Asymmetries and Control in Dynamic Work 12 (Oct. 15, 2015) (unpublished manuscript) (on file with author) (“Many drivers discuss feeling anxious about what they did wrong or in some instances, which passenger wronged them.”); SNL, Five Stars, NBC (Jan. 1, 2017),, (depicting the troubles of a fictional driver and passenger who are both trying to earn a 5 star rating).
  • 34. Superhost, Airbnb,, (describing the privileges according to Superhosts, including priority phone support, anniversary credits to use toward their own travel, and invitations to product and event exclusives).
  • 35. How Do I Become an Elite Tasker?, TaskRabbit,, (stating that elite taskers command higher rates and noting that platinum and diamond-level elite Taskers are charged reduced service fees by the platform).
  • 36. Fiverr’s Levels, Fiverr,, (Workers can offer more “add-on” services and repeated services as they progress up the levels; the final level involves “a manual selection process performed by the Fiverr Editorial team, based on a number of criteria.”).
  • 37. Eric Goldman, Online User Account Termination and 47 U.S.C. § 230(c)(2), 2 U.C. Irvine L. Rev. 659, 673 (2012) (defending the unilateral account termination rights of online actors like platforms because “[o]nline providers need the discretion to manage their communities”).
  • 38. The haziness between the “means” and “ends” of work is beyond the scope of this paper, but for an in-depth analysis, see Julia Tomasetti, The Contracting/Producing Ambiguity and the Collapse of the Means/Ends Distinction in Employment, 66 S.C. L. Rev. 315 (2014–2015).
  • 39. See, e.g., What Factors Determine How My Listing Appears in Search Results?, Airbnb (Aug. 12, 2016),, (“The quicker and more consistently you respond to guests, the better your listing can do in search . . . failing to respond to inquiries will affect your placement and visibility in search.”).
  • 40. Amy Edelen, Uber Plans to Track Drivers Who Speed, Slam the Brakes or Cut Corners, LA Times (June 29, 2016),,
  • 41. New! Get Detailed Rover Cards from Your Sitter or Dog Walker, (Aug. 22, 2016),,
  • 42. Email, Next Step: Send a Rover Card, from to (Jan. 20, 2017) (on file with author).
  • 43. Rosenblat & Stark, supra note 11, at 3772–77.
  • 44. Litigation materials are an exception. See, e.g., Plaintiffs’ Brief at 37–38, New York Taxi Workers Alliance v. Uber Technologies Inc., No. 1:16-cv-04098 (S.D.N.Y. June 2, 2016) (stating that most drivers “place a small sign in a front window that reads ‘Uber’ or a black square with a large ‘U’ for Uber in their front windshields” and that “drivers are encouraged to give cards to passengers in a group who may not have used Uber yet . . . [and] are barred from promoting other businesses to passengers”).
  • 45. As Section II.A shows, this has also been a recurring theme in franchise litigation.
  • 46. Annie Melton, Feastly Co-founder: “We’re Looking to Become the Largest Dining Establishment in the World”, (Jan. 19, 2016),, htt
  • 47. Response of Sylvainbg, Experienced Hosts: Best Airbnb Hacks?, (Aug. 9, 2016),,
  • 48. Jon Youshaei, The Uberpreneur: How an Uber Driver Makes $252,000 a Year, Forbes (Feb. 4, 2015),,
  • 49. How Do I Use Search Filters?, Airbnb,, (advising prospective travelers to “[s]elect the amenities you want for your stay, for example breakfast or a hot tub”).
  • 50. See, e.g., Airbnb, Airbnb Summer Travel Report 6 (2015),, (“[O]ver 10,000 guests stayed in tree houses on Airbnb this summer, over 12,000 guests stayed in yurts, and nearly 13,000 guests stayed in castles.”).
  • 51. Kathleen Mortimer, Integrating Advertising Theories with Conceptual Models of Services Advertising, 16 J. Services Mktg 460, 461 (2002) (describing services as having intangibility, inseparability, heterogeneity, perishability, and ownership); Valarie A. Zeithaml, A. Parasuraman & Leonard L. Berry, Delivering Quality Service: Balancing Customer Perceptions and Expectations 15–16 (1990) (describing services as intangible, inseparable, and heterogeneous).
  • 52. Merz et al., The Evolving Brand Logic: A Service-Dominant Logic Perspective, 37 J. Acad. Marketing Sci. 328, 340 (2009); see also Stephen L. Vargo & Robert F. Lusch, Evolving to a New Dominant Logic for Marketing, 68 J. Marketing 1, 6 (2004) (The service-centered view of marketing is customer-centric and market driven. This means more than simply being consumer oriented; it means collaborating with and learning from customers and being adaptive to their individual and dynamic needs. A service-centered dominant logic implies that value is defined by and cocreated with the consumer rather than embedded in output.).
  • 53. Merz et al., supra note 52, at 336.
  • 54. Reply of Defendant Uber Techs., Inc. in Support of Its Motion for Summary Judgment at 2, O’Connor v. Uber Techs., 82 F. Supp. 3d 1133 (N.D. Cal. 2015) (No. C-13-3826 EMC) (“Uber provides drivers with a service and not the reverse.”); see also Marion Crain, Managing Identity: Buying into the Brand at Work, 95 Iowa L. Rev. 1179 (2010) (discussing the absence of regulation regarding internal branding).
  • 55. See, e.g., Celebrating 100 Cities, YouTube (Feb. 3, 2015),; Airbnb Views, YouTube (May 9, 2014), On the importance of worker happiness to consumer satisfaction, see Spiros Gounaris, The Notion of Internal Market Orientation and Employee Job Satisfaction: Some Preliminary Evidence, 22 J. Servs. Marketing 68 (2008). On the advertising strategies of services marketers generally, see Avery M. Abernethy & Daniel D. Butler, Promoting Customer Contact People: A Key Difference in Service and Product Advertising, 7 J. Servs. Marketing 4, 5 (1993) (arguing that “it may be more important for service marketers to feature illustrations or descriptions of service providers and contact people in advertising than product marketers . . . [because] [e]mployee experience and competence are key indicators of service quality for many purchasers”); George M. Zinkhan et al., Differences Between Product and Services Television Commercials, 6 J. Servs. Marketing 59, 65 (1992) (arguing that “[t]he prevalence of transformational advertising in service ads”—meaning ads that link the brand to positive experiences or states of being—“is not surprising given the intangibility and heterogeneity of services”).
  • 56. There is one important set of work relationships that this section does not discuss: domestic and agricultural workers. Their omission may seem odd given that these workers also are excluded from many labor and employment protections simply by virtue of how their work is labeled rather than based on the conditions under which they work. But the unprotected status of domestic and agricultural workers is not due to the fact that they are controlled by someone not legally recognized as their employer: it arises out of our decision to deny them protections regardless of who exercises authority—invisibly or otherwise—over their work. 29 U.S.C. §§ 213 (a)(6) (exempting agricultural workers from minimum wage and overtime laws) and (a)(15) (exempting companionship workers and casual domestic workers). Of course, some states have sought to partially fill this gap by enacting bills of rights for domestic workers. See, e.g, Domestic Workers Bill of Rights (A1470B/S2311E) (New York) (passed July 1, 2010); California Bill of Rights (AB 241) (Sept. 26, 2013); Act 248 (July 1, 2013) (Hawaii) (amending Rev. Statutes §§ 378-1 & 378-2); House Bill 1288 (“Domestic Workers Bill of Rights”) (Illinois) (Aug. 12, 2016); Sean Farhang & Ira Katznelson, The Southern Imposition: Congress and Labor in the New Deal and Fair Deal, 19 Stud. Am. Pol. Dev. 1 (2005) (describing Southern efforts to shield agricultural labor from New Deal legislation so as to preserve antebellum work structures).
  • 57. Francine Lafontaine & Roger D. Blair, The Evolution of Franchising and Franchise Contracts: Evidence from the United States, 3 Entrepreneurial Bus. L.J. 381, 385–86 (2008–2009).
  • 58. Id. at 386. Consumption, which was perhaps already an unusually visible, status-maximizing activity in the American context, arguably became even more socially important after the war. It was recast as doubly virtuous, being both democratizing (since more people could now access the same quality products) and patriotic (because spending contributed to economic recovery). Lizabeth Cohen, A Consumer’s Republic: The Politics of Mass Consumption in Postwar America, 31 J. Consumer Res. 236, 236–37 (2004).
  • 59. William L. Killion, The Modern Myth of the Vulnerable Franchisee: The Case for a More Balanced View of the Franchisor-Franchisee Relationship, 28 Franchise L.J. 23, 25 (2008) (listing statistics for 1955 and 1970); International Franchise Association Educational Foundation, Franchise Business Economic Outlook for 2016 8, 14–18 (Jan. 2016) (on file with author) (listing statistics for 2015).
  • 60. Jeffrey H. Wolf & Aaron C. Schepler, Caught Between Scylla and Charybdis: Are Franchisors Still Stuck the Rock of Non-Uniformity and the Hard Place of Vicarious Liability?, 33 Franchise L.J. 195, 196 (2013–2014).
  • 61. Dean T. Fournaris, The Inadvertent Employer: Legal and Business Risks of Employment Determinations to Franchise Systems, 27 Franchise L.J. 224, 224–27 (2008) (emphasis added).
  • 62. William L. Killion, Putting Critical Decision Making Where It Belongs: Scouring the Franchise Agreement of the “D” Word, 24 Franchise L.J. 228, 228 (2005).
  • 63. Wolf & Schepler, supra note 60, at 198; see also David J. Kaufmann et al., A Franchisor is Not the Employer of its Franchisees or Their Employees, 34 Franchise L.J. 439, 441 (2015) (“[T]he integral importance of the Lanham Act and its requirement that franchisors exert the very types of controls on their franchisees which the NLRB General Counsel Complaints wrongly characterize as indicia of an employment relationship.”).
  • 64. Trademarks are like symbols, which by social and marketing convention convey a huge range of ideas, experiences, and objects; they are not things that communicate information directly like photos to faces or heavy clouds to rain. The latter two types of signifiers are usually called “icons” and “indices” rather than “symbols” in the system developed by C.S. Pierce. Barton Beebe, The Semiotic Analysis of Trademark Law, 51 UCLA L. Rev. 621, 637 (2004) (citing Charles Sanders Peirce, Collected Papers of Charles Sanders Peirce 247–49 (Charles Hartshorne & Paul Weiss eds., 1934)). Note that the actual trademark or signifier (e.g., the Golden Arches) is only one part of a broader relation collectively called a “sign”; another part of the sign is the “signified,” which in the case of trademarks is the source or goodwill indicated by the signifier. Beebe argues that “the failure to recognize that the trademark is not merely the signifier, but is rather a full-blown sign is often the cause of judicial error.” Beebe, Semiotic Analysis, at 650.
  • 65. On the Golden Arches generally, see Alan Hess, The Origins of the McDonald’s Golden Arches, 45 J. Soc’y Architectural Historians 60 (1986).
  • 66. Thomas J. Walsh III, Comment, Supersizing the Definition of Employer Under the National Labor Relations Act—Broadening the Joint-Employer Standard to Include Franchisors and Franchisees, 47 U. Tol. L. Rev. 589, 606–07 (2016); see also Salazar v. McDonald’s Corp., 2016 U.S. Dist. LEXIS 108764, 25–26 (N.D. Cal. Aug. 16, 2016) (observing that “individuals can apply for employment” at one of the restaurants owned by the defendants via a website called “Hiring to Win” where “the screening questions were written by McDonalds” but concluding that this did not point to a joint employer relationship since use of website was optional); Ochoa v. McDonald’s Corp., 133 F. Supp. 3d 1228, 1240 (N.D. Cal. Sept. 24, 2015) (noting that all but one of the plaintiffs “applied for a job through McDonald’s website” in the course of finding that McDonald’s might be a joint employer under a theory of ostensible agency).
  • 67. Walsh, supra note 66, at 606–07.
  • 68. See, e.g., Gray v. McDonald’s USA, LLC, 874 F. Supp. 2d 743, 750 (2012).
  • 69. Complaint at 32, Pullen v. McDonald’s Corp., No. 5:14Cv11081, 2014 WL 978792 (E.D. Mich., Mar. 13, 2014) (“[T]he O&T Manual establishes standards with which McDonald’s Corporate expects franchisees, including ECS McDonald’s, to comply in regards to uniforms and grooming.”). The same complaint also details various other aspects of workers’ experiences that are substantially affected or evaluated according to standards set out in the O&T Manual. Id. at 31–32.
  • 70. Complaint at 13–16, Betts v. McDonald’s Corp., No. 4:15-cv-00002-JLK (W.D. Va.) (filed Jan. 22, 2015) (noting that Business Consultants conduct “Full Operations Reviews,” “Short Operations Reviews,” and “Support Visits” and that Mystery Shoppers evaluate “tone of voice and eye contact with customers, facial expressions, exact words used, and assembly of food items”).
  • 71. See also Alex Felstead, The Social Organization of the Franchise: A Case of “Controlled Self-Employment”, 5 Work, Emp. & Soc’y 37, 45 (1991) (stating that in the late 1980s and early ‘90s McDonald’s “employs nearly 300 field service consultants, each of whom visits and evaluated about 18 stores four times a year” each time assessing a store “on more than 500 items,” and adding that “Kentucky Fried Chicken makes similar detailed inspections”).
  • 72. Alex Rosenblat, The Truth About How Uber’s App Manages Drivers, Harv. Bus. Rev. (Apr. 6, 2016),, https://perma.
    cc/HG9Z-FT5U (discussing the “redistribution of managerial oversight and power away from formalized management and toward a triadic relationship between employers-workers-consumers”).
  • 73. Global Front Counter Service: Greeting, Order Taking, Receiving Payment (July 2014),, (describing total experience time); Ana Rodriguez, McDonald’s New Employee Training Manual 6, Restaurant #16866 (May 2, 2014),, (showing a diagram for hamburger assembly time).
  • 74. Global Front Counter Service, supra note 73, at 1 (some emphasis removed from original); Conrad P. Kottak, Rituals at McDonald’s, 1 J. Am. Culture 370, 373 (1978).
  • 75. Coffee: Preparation, Brewing and Serving 2 (Aug. 2012),, (describing coffee preparation); Rodriguez, supra note 73, at 6 (describing bun toasting and room temperature condiments); Hola, Comment to McDonald’s Burger Assembly Chart?, Yahoo!Answers (May 14, 2012, 3:17pm),, (listing, in the form of equations, the ingredients for various sandwiches); see also Robin Leidner, Fast Food, Fast Talk: Service Work and the Routinization of Everyday Life 49 (1993) (noting that McDonald’s managers call the O&T Manual “the Bible” and that the Manual’s “600 pages include, for instance, full-color photographs illustrating the proper placement of ketchup, mustard, and pickle slices on each type of hamburger”).
  • 76. At the same time, it is quite familiar. Compare the McDonald’s specifications on worker behavior and speed with the following allegations regarding the home cleaning service Handy: “The cleaning personnel . . . must wear Handybook uniforms . . . adhere to detailed requirements governing how they clean a customer’s home, how they behave on the job, how they greet customers, communicate with customers, and the steps to follow in cleaning their residence, including how long the job is to take.” Complaint at 2, Malone v. Handybook Inc., No. BC555367 (Sup. Ct. Cal., Aug. 21, 2014).
  • 77. On pre-approval training, see D.L. Noren, The Economics of the Golden Arches: A Case Study of the McDonald’s System, 34 Am. Economist 60, 60 (1990); Stephanie Sullivant, Comment, Restoring the Uniformity: An Examination of Possible Systems to Classify Franchisees for Workers’ Compensation Purposes, 81 UKMC L. Rev. 993, 998 (2013) (“Most franchise companies offer a 1–4 week training program that is usually held at their corporate offices or at an actual franchise location.”); Gray v. McDonald’s USA, LLC, 874 F. Supp. 2d 743, 747 (E.D. Tenn. 2012). On the absence of vicarious liability, see Parmenter v. J&B Enters., 99 So. 3d 207, 215 (Miss. Ct. App. 2012); Catalano v. GWD Mgmt. Corp., 2005 U.S. Dist. LEXIS 45255, *30–42 (S.D. Ga. Mar. 30, 2005); Rainey v. Langen, 998 A.2d 342, 350 (Me. 2010); Hoffnagle v. McDonald’s Corp., 522 N.W.2d 808, 814–15 (Iowa 1994).
  • 78. Parmenter, 99 So. 3d at 214.
  • 79. Kerl v. Rasmussen, 273 Wis. 2d 106, 131–33 (S.C. Wis. 2004); see also Papa John’s Int’l, Inc. v. McCoy, 244 S.W.3d 44, 47 (Ky. 2008); Vandemark v. McDonald’s Corp., 904 A.2d 627, 636 (N.H. 2006); Wendy Hong Wu v. Dunkin’ Donuts, Inc., 105 F. Supp. 2d 83, 88–90 (E.D.N.Y. 2000).
  • 80. 60 Cal. 4th 474, 497–98 (Cal. 2014).
  • 81. Id. at 478 (holding that liability only accrues where the franchisor has “retained or assumed a general right of control over factors such as hiring, direction, supervision, discipline, discharge, and relevant day-to-day aspects of the workplace behavior of the franchisee’s employees”).
  • 82. Browning-Ferris Indus. of Calif., Inc., 362 N.L.R.B. 186 (2015).
  • 83. Id. at 2.
  • 84. See, e.g., Kaufmann et al., supra note 63, 448–52; Erin Conway & Caroline Fichter, Surviving the Tempest: Franchisees in the Brave New World of Joint Employers and $15 Now, 35 Franchise L.J. 509, 539 (2016) (observing that “[i]n the commentary surrounding both the minimum wage debate and the joint employer standard, the franchise industry’s refrain has been: if X happens, franchising will not survive”); John T. Bender, Barking Up the Wrong Tree: The NLRB’s Joint-Employer Standard and the Case for Preserving the Formalities of Business Format Franchising, 35 Franchise L.J. 209, 235 (2015–2016) (warning that “[t]aken to its logical end, franchisors in every sector of the economy will be forced to either eliminate many of the systems and standards that secure their position in highly competitive markets or even abandon franchising entirely”).
  • 85. Following the confirmation of President Trump’s two appointees, William J. Emanuel and Marvin Kaplan, the Board is split 3–2 in favor of Republicans (although the third Republican, Chairman Miscimarra, was appointed by President Obama). Noam Scheiber, Trump Takes Steps to Undo Obama Legacy on Labor, N.Y. Times (June 20, 2017),,; see also Who We Are—The Board, Nat’l Lab. Rel. Board, [
  • 86. Fournaris, supra note 61, at 230 (“Franchising is a unique type of business arrangement. By its nature, it cannot be accommodated easily by the control criteria used by courts and regulators to identify employment.”); Lobel, supra note 2.
  • 87. Bender, supra note 84, at 224 (“The popularity of franchising stemmed in large part from the fact that it was viewed as striking the right balance between ‘the economic efficiency of big business with the personal satisfaction and social advantages of small business ownership.’”); Kessler, supra note 31 (describing the appeal of the sharing economy as follows: “Instead of selling your soul to the Man, it goes, you are empowered to work for yourself on a project-by-project basis . . . . The best part? The work will come to you, via apps on your smartphone, making the process of finding work as easy as checking your Twitter feed.”).
  • 88. Kaufmann et al., supra note 63, at 454; see also, Koopman et al, supra note 25, at 2 (stating that “[t]he key contribution of the sharing economy . . . is that it has overcome market imperfections without recourse to traditional forms of regulation”).
  • 89. Compare Fournaris, supra note 61, at 230; with Harris & Kreuger, supra note 87, at 8 (“Forcing these new forms of work into a traditional employment relationship could be an existential threat to the emergence of online-intermediated work, with adverse consequences for workers, consumers, businesses, and the economy.”).
  • 90. Richard R. Carlson, Why the Law Still Can’t Tell an Employee When It Sees One and How It Ought to Stop Trying, 22 Berkeley J. Emp. & Lab. L. 295, 302–06 (2001); Roscoe T. Steffen, Independent Contractor and the Good Life, 2 U. Chi. L. Rev. 501, 501 (1935).
  • 91. In pointing out that industrialization moved the center of work from the home to the factory I do not mean to say that the older system of “putting-out” work to be done in the home was merely transitional or that it no longer exists—if anything, the platform economy demonstrates that neither of these is the case. Matthew W. Finkin, Beclouded Work, Beclouded Workers in Historical Perspective, 37 Comp. Lab. L. & Pol’y 603, 604–08 (2015–16). On the changing connection between labor and autonomy in the nineteenth century, see, e.g., Nancy Fraser & Linda Gordon, A Genealogy of Dependency: Tracing a Keyword of the U.S. Welfare State, 19 SIGNS 309, 315–16 (1994); Robert J. Steinfeld, The Invention of Free Labor: The Employment Relation in English and American Law and Culture, 1350–1870, 185–87 (1991).
  • 92. Steinfeld, supra note 91, at 136–37 (noting that by the early nineteenth century, “aside from indentured servitude, jurisdiction too was . . . a feature of the traditional relationship limited to minors” and masters might discipline servants “only when it served to prepare the minor for a self-sufficient and independent adulthood”).
  • 93. Gerald M. Stevens, The Test of the Employment Relation, 38 Mich. L. Rev. 188, 189, 194 (1939).
  • 94. See, e.g., John Bruntz, The Employee/Independent Contractor Dichotomy: A Rose is Not Always a Rose, 8 Hofstra Lab. L.J. 337, 338–40 (1990–1991); Steffen, supra note 90, at 511 (stating that the independent contractor did not become a prominent judicial category “before the court nineteenth century was well advanced”).
  • 95. Carlson, supra note 90, at 315.
  • 96. Dubal, supra note 3, at 86–87 (discussing Taft-Hartley); see also Equal Pay Act of 1963, 29 U.S.C. §206d(1) (“No employer having employees subject to any provisions of this section shall discriminate, within any establishment in which such employees are employed, between employees on the basis of sex.”); Title VII of the Civil Rights Act of 1964, 42 U.S.C. 2000e (defining employee as “an individual employed by an employer” and employer as “a person engaged in an industry affecting commerce who has fifteen or more employees for each working day in each of twenty or more calendar weeks in the current or preceding calendar year, and any agent of such a person”).
  • 97. The Civil Rights Act of 1866, 42 U.S.C. § 1981 is an important exception to this state of affairs, since it protects “[a]ll persons within the jurisdiction of the United States” against racial discrimination in the making and enforcement of contracts.
  • 98. See, e.g., Stephen F. Befort, Labor and Employment Law at the Millennium: A Historical Review and Critical Assessment, 43 B.C. L. Rev. 351, 368 (2002) (describing various factors for the “contingency explosion”); Richard S. Belous, Contingent Economy: The Growth of the Temporary, Part-Time and Subcontracted Workforce 12–17 (1989) (locating the beginning of the shift towards contingent work in the 1980s).
  • 99. Dubal, supra note 3, at 81–82; Marc Linder, Towards Universal Worker Coverage Under the National Labor Relations Act: Making Room for Uncontrolled Employees, Dependent Contractors, and Employee-Like Persons, 66 U. Det. L. Rev. 555, 576–77 (1988–89).
  • 100. Linder, supra note 99, at 578–79.
  • 101. Id. at 582 (citing Yellow Cab Co. and Local 777, Democratic Union Org. Comm., 229 N.L.R.B. 190, 1977–78 N.L.R.B. Dec. (CCH) ¶ 18,262 at 30,322).
  • 102. Linder, supra note 99, at 586 (citing Local 777, Democratic Union Org. Comm. v. NLRB, 603 F.2d 862, 877–78 (D.C. Cir. 1978)).
  • 103. Marc Linder, The Involuntary Conversion of Employees into Self-Employed: The Internal Revenue Service and Section 530, 22 Clearinghouse Rev. 14, 14 (1988–1989).
  • 104. Vizcaino v. Microsoft Corp. (Vizcaino I), 97 F.3d 1187 (9th Cir. 1996).
  • 105. Id. at 1190 (observing that “Microsoft fully integrated the plaintiffs into its workforce . . . [T]hey received admittance card keys, office equipment and supplies from the company . . .” but adding that the same workers “wore badges of a different color, had different electronic-mail addresses, and attended a less formal orientation than that provided to regular employees . . . were not permitted to assign their work to others, invited to official company functions, or paid overtime wages . . . [and] were not paid through Microsoft’s payroll department”).
  • 106. Id. at 1190–91.
  • 107. Steven J. Arsenault et al., An Employee by Any Other Name Does Not Smell as Sweet: A Continuing Drama, 16 Lab. L. 285, 285 (2000).
  • 108. Paul H. Tobias, The Microsoft Decision and Its Side Effects: What Are Employers’ Obligations Regarding Benefits and Employees in the Contingent Workforce? (July 11, 2000),,
  • 109. Compare Vizcaino I, 97 F.3d at 1190 (noting that suit was originally filed in 1993) with Roadway Package System, 326 N.L.R.B. 842, 842 (1998) (noting that a Regional Director issued a decision finding the drivers to be employees in 1995).
  • 110. Roadway, 326 N.L.R.B. at 842.
  • 111. FedEx Home Delivery v. NLRB, 563 F.3d 492, 504 (D.C. Cir. 2009). Several commentators have critiqued the D.C. Circuit’s analysis both for its emphasis of one particular factor in the usual common law analysis—entrepreneurship—and for its focus on entrepreneurial opportunity rather than evidence of actual entrepreneurial behavior. See, e.g., Griffin Toronjo Pivateau, Rethinking the Worker Classification Test: Employees, Entrepreneurship, and Empowerment, 34 N. Ill. U. L. Rev. 67, 95 (2013) (proposing a new classification test based on “genuine entrepreneurial opportunity” and noting that the D.C. Circuit “focused on theoretical opportunity and not actual opportunity”); Jeffrey M. Hirsch, Employee or Entrepreneur?, 68 Wash. & Lee L. Rev. 353, 354–55 (2011) (arguing that the D.C. Circuit’s “approach flies in the face of a long history of common law and undermines the aims of the NLRA and other labor and employment laws” and also raises troubling questions “for the court’s deference to the NLRB”); Jeffrey E. Dilger, Pay No Attention to the Man Behind the Curtain: Control as a Nonfactor in Employee Status Determinations Under FedEx Home Delivery v. NLRB, 26 ABA J. Lab. & Emp. L. 123, 141 (2010) (stating that the D.C. Circuit’s approach involves “no inquiry into the quality or limitations on the potential entrepreneurialism,” “upset[s] over fifty years of court decisions,” and “entirely discounts certain Restatement factors”).
  • 112. Alexander v. FedEx Ground Package Sys., 765 F.3d 981, 987 (9th Cir. 2014) (“Between 2003 and 2009, similar cases were filed against FedEx in approximately forty states.”).
  • 113. Id.
  • 114. Id. (“The MDL Court remanded this case to the district court to resolve the drivers’ claims under the FMLA.”).
  • 115. Id. at 988.
  • 116. Id. at 989–90.
  • 117. Id. at 989.
  • 118. Id. at 990.
  • 119. Id. (citing Ruiz v. Affinity Logistics Corp., 754 F.3d 1093, 1102 & n.5 (9th Cir. 2014) and stating that “no reasonable jury could find that the “results” sought by FedEx includes detailed specifications as to the delivery driver’s fashion choices and grooming . . . [or] include having all of its vehicles containing shelves built to exactly the same specifications”).
  • 120. Cotter v. Lyft, 60 F. Supp. 3d 1067, 1081 (N.D. Cal. Mar. 11, 2015) (stating that “[t]he experience of the Lyft driver is much different from the experience of the FedEx driver”).
  • 121. See, e.g., Vehicle Options in Los Angeles, (listing vehicle requirements for each type of Uber service within the Los Angeles market and specific models that satisfy those requirements); Plaintiffs’ Brief, New York Taxi Workers Alliance v. Uber Technologies Inc., supra note 44, at 29 (stating that drivers were instructed on specific terms to avoid with disabled passengers) and 29–30 (stating that drivers were instructed on how to respond to late passengers, including the number of times to contact such passengers).
  • 122. On state efforts, see Mass. Workers’ Compensation Advisory Council, Fiscal Year 2014 Annual Report (Apr. 2015),, (stating that Massachusetts has set up a task force to investigate worker misclassification). On the Department of Labor’s collaborations with state governments, see Janet Marsky et al., October 2014 Independent Contractor Compliance and Misclassification Update, (Nov. 5, 2014),, (observing that Alabama became the 16th state to enter into a federal-state misclassification pact and describing joint anti-misclassification efforts by Illinois and the Department of Labor). See also Admin. Op. No. 2015–1, The Application of the Fair Labor Standards Act’s “Suffer or Permit” Standard in the Identification of Employees Who Are Misclassified as Independent Contractors (July 15, 2015),, (advocating for a six-factor economic realities test).
  • 123. Less than two months after the DOL issued its administrative opinion, the Internal Revenue Service issued a “reminder” and the NLRB issued an opinion regarding the proper way to determine worker classification. See Admin. Op. No. 2015–1, supra note 121; Payments to Independent Contractors, US Dep’t of Treasury, IRS (Aug. 2015),, (advocating a three-factor common law test); Sisters’ Camelot, 363 N.L.R.B. 13 (Sept. 25, 2015) (using an eleven-factor common law test guided by the Restatement (Second) of Agency).
  • 124. Tyrone Richardson, Gig Workers Need Updated Wage and Hour Law, House Panel Chair Says, Bloomberg BNA (May 4, 2017),,
  • 125. For other comparative efforts, see, e.g., Ryan Calo & Alex Rosenblat, The Taking Economy: Uber, Information, and Power, 117 Colum. L. Rev. (2017) (forthcoming) (comparing platforms with the multilevel marketing company Amway); Sanjukta M. Paul, Uber as For-Profit Hiring Hall: A Price-Fixing Paradox and its Implications, 38 Berkely J. Lab. & Emp. L. 233 (2016) (comparing Uber with union hiring halls); Biber et al., supra note 12 (comparing platforms with product and business format franchisors and the early electricity industry, although not explicitly focusing on labor and employment concerns).