Introduction

In April 2019, the Office of the General Counsel (GC) of the National Labor Relations Board (NLRB) issued an “advice memo” addressing the issue of independent contractor versus employee classification for internet-based service firms (IBSFs) (NLRB GC 2019). Under the National Labor Relations Act (NLRA), only workers who are classified as employees have the right to form or join unions or to engage in collective bargaining or other efforts to improve their working conditions. Independent contractors are not covered by the NLRA and therefore do not have the rights and protections it grants employees. The NLRB is the federal agency that enforces the NLRA. Employees whose rights under the NLRA are violated must rely on the NLRB’s Office of the General Counsel to represent them (they can’t bring their own case to court, i.e., they have no “private right of action”).

That is why the GC’s advice memo is so significant. The GC’s advice memo—which is not reviewable—sets forth the General Counsel’s position that workers who drive for the on-demand transportation service company Uber are independent contractors, not employees. This determination denies Uber drivers coverage under the NLRA, effectively robbing them of the right to engage in collective action, such as organizing a union or collectively bargaining, under the current general counsel. And, although the memo specifically covers Uber and UberX drivers, GC advice memos are released to show how the GC analyzes certain facts, and thus in effect serve as guidance on the Board’s treatment of similarly situated workers and employers.

In order to deny Uber drivers NLRA protections, the general counsel’s advice memo misstates and manipulates the facts and bends the law. Namely, the GC’s advice memo overstates the realities of Uber drivers’ “entrepreneurial freedom,” understates the degree of control Uber has over its drivers, and discounts relevant analysis related to how Uber drivers are paid. It is only through cherry-picking facts and ascribing outsized weight to some factors that the office of the GC reaches its erroneous conclusion.

In this report, we first review the law surrounding questions of employee status under the NLRA. We then examine the realities of digital platform driving—specifically for Uber drivers. Finally, we demonstrate that applying relevant facts to the law governing classification demands that drivers providing personal transportation services using an employer’s app-based program be covered employees for purposes of the NLRA.

Rights guaranteed under the NLRA

Congress enacted the NLRA in 1935. Under the NLRA, employees have the right to join a union and bargain collectively. But the act provides protections beyond this as well. Workers covered by the NLRA may engage in “concerted activities for the purpose of…mutual aid or protection,” regardless of whether they are union-related.1 This means that workers’ efforts to improve wages and working conditions can be protected even in nonunion workplaces. Recently, there have been a number of examples that demonstrate the power of workers’ collective action. Google workers walked off the job in protest of the company’s handling of sexual harassment, and the company changed its policy of forced arbitration of these cases (Campbell 2019; Rhinehart 2019). Fast-food workers have also participated in strikes to fight for $15 per hour and for a union (NELP 2018). The NLRA protects such actions, prohibiting employers from firing workers in retaliation for exercising their rights under the act.

Most private-sector workers are covered by the NLRA. However, there are some notable exceptions. For example, it does not cover certain agricultural and domestic workers. The Taft-Hartley Act of 1947 amended Section 2(3) of the NLRA to exclude independent contractors from coverage.2

How the NLRB and the courts have determined employee status

Whether a worker is an employee or an independent contractor is a threshold question in labor and employment law.3 The NLRB has considered the issue in countless cases. The Supreme Court has long held that the determination of whether a worker is a covered “employee” or an exempt “independent contractor” under the NLRA is governed by “common-law agency” principles.4 In a seminal 1968 opinion, the Court stated that “there is no shorthand formula or magic phrase that can be applied to find the answer.”5 Rather, “all of the incidents of the relationship must be assessed and weighed with no one factor being decisive.6 What is important,” the Supreme Court explained, “is that the total factual context is assessed in light of the pertinent common-law agency principles.” The Supreme Court has been guided in its analysis by the multifactor test articulated in the Restatement (Second) of Agency § 220 (1950).7

In recent years, the NLRB has examined the issue of employment status in a series of cases involving FedEx drivers. In several cases, the Obama Board followed the long-settled law and applied the common-law factors to reach the conclusion that these drivers were employees for purposes of the NLRA. However, the District of Columbia Circuit found otherwise.

The D.C. Circuit reached its conclusion that FedEx drivers are not employees under the NLRA by moving from the traditional discussion of an employer’s right to control (the right to direct how, where, and when a worker does her job) to a “more accurate proxy: whether the putative independent contractors have ‘significant entrepreneurial opportunity for gain or loss.’”8 While the D.C. Circuit explicitly stated that all common-law considerations remain in play, it held that an “animating principle” by which to evaluate those factors is “whether the position presents the opportunities and risks inherent in entrepreneurialism.”9

In its 2014 decision in FedEx, the Obama Board reaffirmed its commitment to the “nonexhaustive common-law factors enumerated in the Restatement (Second) of Agency.” The Board specifically declined to adopt the entrepreneurial-opportunity animating-principle inquiry the D.C. Circuit had advanced, reasoning that to do so “would mean a broader exclusion from statutory coverage than Congress appears to have intended.”10

However, in 2017, the D.C. Circuit again embraced the entrepreneurial opportunity analysis. And, in January 2019, the Trump NLRB overruled the 2014 FedEx decision in SuperShuttle DFW, Inc. Specifically, the Trump Board expressly adopted the D.C. Circuit’s decision approach and held that an “important animating principle by which to evaluate those [multiple common-law] factors…is whether the position presents opportunities and risks inherent in entrepreneurialism.”11

Board member Lauren McFerran’s dissent offers a comprehensive critique of the issues with the majority’s approach in SuperShuttle, namely that it cannot be reconciled with Supreme Court and Board precedent since entrepreneurial opportunity is not at the core of the common-law agency test of employee status. Nevertheless, in his April 2019 advice memo, the NLRB’s GC uses the SuperShuttle analysis to deny NLRA protections to drivers for the internet-based service firm Uber. In essence, the Trump NLRB and its GC have advanced a new analysis emphasizing entrepreneurial opportunity while continuing to erroneously label the analysis the traditional common-law test.

Regardless of whether the SuperShuttle decision employs the appropriate analysis for evaluating employee vs. independent contractor status, the GC’s advice memo is wrong in its application of the facts to conclude that Uber drivers have entrepreneurial opportunity.

The realities of Uber driving: A quick overview

Uber advertises to drivers that they will work for themselves. Driving for Uber is flexible, with the driver in control, according to the company’s website. Interested drivers just download the driving app and complete a “sign-up” process that requires only that drivers have a valid driver’s license and insurance and “complete a background screening.” The company states that drivers set their own hours and may “cash out” after each trip (up to five times per day on the app). Uber brands itself as merely a technology platform that allows drivers to find earning opportunities for their own entrepreneurial endeavors (Uber 2019a).

However, in reality, Uber drivers’ work experience is quite different. Drivers have no say on setting fares, on what they are paid, or on the commissions the company takes. Drivers are not shown the passenger’s destination or how much they could earn on a fare before being asked to accept a ride request, and they have limited say on whom they choose to have as customers (Rosenblat 2018a).

Drivers are not even able to choose the route to take—Uber reserves the right to retroactively adjust the fare if it decides that an inefficient route was taken (Rosenblat and Stark 2016; U.K. Judiciary 2016). And Uber also exerts control over drivers through an automated rating system based on passenger ratings. Tools like the fare and rating systems serve as “algorithmic managers,” nudging drivers to act in certain ways and penalizing them when they don’t. For example, in certain services on Uber’s platform, if drivers fall below 4.6 stars on a 5-star rating system, they may be “deactivated”—never “fired.” So some drivers tolerate bad passenger behavior rather than risk losing their livelihoods because of retaliatory reviews (Rosenblat 2018b). Given that a driver’s low rating (as unilaterally defined by Uber) may lead Uber to deactivate them from the app (i.e., fire them), drivers do not have the independent control typically associated with a small business owner.

In fact, a full rendering of the realities of driving for Uber (or the competing on-demand transportation service company Lyft) eliminates any commonsense understanding of such work as self-employment or as running an independent business. Uber drivers’ control over earnings opportunities primarily boils down to deciding how many hours they drive.12 Beyond that, drivers lack entrepreneurial opportunity—the basic ability to improve their earnings by expanding their business and their earnings—and control over their work.

The GC’s analysis of the realities of Uber driving gets the facts wrong

As noted earlier, the NLRB general counsel relies on the legal analysis in the Board’s recent decision in SuperShuttle to deny NLRA coverage to Uber drivers. In SuperShuttle the Board concluded that an “important animating principle by which to evaluate those [multiple common-law factors]…is whether the position presents the opportunities and risks inherent in entrepreneurialism.” Adopting this reasoning, the GC finds that Uber drivers have significant entrepreneurial opportunity by virtue of their “near complete control of their cars and work schedules” and their freedom to “choose log-in locations and to work for competitors.” The GC finds particularly compelling the notion that drivers have “unlimited freedom to look elsewhere for better earnings,” arguing that this consideration “minimized the impact that certain other features of the Uber system would otherwise have on their entrepreneurial opportunity.” Essentially, the GC finds that the fact that drivers can search for better earnings from driving for other transportation network companies (TNCs), or can seek another venture, outweighs Uber’s control of fare-setting and prohibition on subcontracting their work—constraints that clearly limit drivers’ freedom. This logic is confounding since all employees not bound by noncompete agreements or other contractual arrangements may look elsewhere for better earnings or to start a new venture. Setting aside whether the Board majority’s approach in SuperShuttle is correct under the law, the GC memo is wrong on the facts.

Uber drivers have no control over prices

The GC memo claims that Uber drivers have entrepreneurial opportunity. Common understanding of entrepreneurial driving involves the opportunity to expand revenues. However, Uber drivers are blocked from employing all the traditional tools one uses to expand revenues, starting with prices. Uber drivers have no say on prices because Uber calls all the shots. As the GC memo notes, “Uber sets baseline fares (subject to a driver’s contractual right to negotiate a lower fare).” Uber unilaterally sets rates and has the power to change the fares, and it repeatedly alters the fares passengers pay, what drivers are paid, and the commissions it takes (Rosenblat 2016).

In 2016 Uber even altered the whole pricing regime from a time and distance standard to “upfront pricing” that now severs the connection between passenger fares and driver earnings. Driver earnings used to be a share of the fares: Uber took a flat fee plus a percentage commission and the remainder went to the driver. Now the fare is set by Uber up front and the driver is paid based on time and distance driven (this is explained in Rosenblat 2018a, 121–125).

Uber drivers cannot increase revenues by expanding their customer base

In addition to having no say over prices, Uber drivers are prohibited from expanding their own revenue by building a customer base. Uber prohibits drivers from collecting rider logistical information (e.g., names, phone numbers, addresses, etc.) or contacting riders, eliminating the possibility of marketing for repeat rides.13 Even if drivers were somehow able to market themselves to obtain more customers, such efforts would have no impact on earnings as potential passengers cannot select a particular driver through the Uber app. The company completely controls the matching of passengers to drivers.

The right to subcontract has traditionally been a significant factor in determining a worker’s classification as an independent contractor. However, Uber prohibits subcontracting, as the GC memo acknowledges: “Drivers could not subcontract their work” (NLRB GC 2019, 7). This limitation obviously constrains a driver’s ability to expand revenue as the driver cannot provide services to a potentially enlarged customer base. Even if a driver could develop a larger customer base through marketing (which is a fiction), it is not possible to expand the driver’s “business” capacity via a fleet of cars serving an enlarged market since a driver cannot allow another driver to use his or her Uber account.

In fact, aside from putting in more hours, the only plausible ways drivers can improve revenues is through customer tipping, picking when and where to drive, and other driving strategies that in reality have little impact on overall earnings.

Uber drivers can’t expand revenues through enhanced services leading to tipping

A driver’s ability to enhance his or her own revenue by providing additional services (e.g., a phone charger, bottled water, etc.) was limited by the constraints on tipping initially imposed by Uber. Until 2017, Uber prohibited tipping. Now, Uber provides for tipping in the app. However, the company limits the amount a rider can tip a driver—capping tips at 200% the cost of the ride (K. Wong 2018). It is difficult to understand how these controls on a driver’s earnings ability support a finding of independent contractor status.

Uber drivers cannot expand earnings through experience and strategic driving

Another indicator that Uber drivers lack entrepreneurial opportunity is their demonstrated limited ability to expand earnings through better strategies and experience. This is shown in an academic paper on gender equity of Uber driver earnings, co-authored by Uber’s chief economist. The paper quantifies the value of experience and argues that earnings rise through experience. As the authors note:

Indeed, there is much to learn being a driver on Uber. Uber pays according to a fixed formula, but many of the parameters of the formula (wait time, accepts-to-pickup distance, trip distance, speed, surge multiplier and incentive payouts) are within the driver’s control. For example, drivers can indirectly affect the surge multiplier and wait times by choosing where and when to work and directly affect their driving speed by simply driving faster. As drivers work more, they can begin to learn optimal driving behaviors to maximize earnings. (Cook et al. 2018, 20)

But the findings in the paper actually show that Uber drivers have a limited ability to raise their earnings, i.e., that there is a pretty low and fixed ceiling on Uber driver earnings.

Cook et al. (2018, 22) note that “[a]ny experience premium results from learning and increased driver productivity,” meaning we can assess the ability of drivers to expand their business by examining how hourly earnings rise with experience.

The paper estimates the growth of earnings as drivers move from having provided less than 100 trips to providing 100–500, 500–1,000, 1,000–2,500, and over 2,500 trips, based on an analysis of the earnings growth of 120,000 Chicago drivers over time. For full-time drivers, the analysis is equivalent to comparing the earnings of those who have been driving for three months, eight months, 18 months, and more than 25 months relative to those who have driven for a month or less.14

The basic finding is that a driver who has completed more than 2,500 trips (i.e., a driver who has at least 25 months of experience) earns 9.0% more (in non-inflation-adjusted dollars) than a driver who has completed 100–500 trips (an average of three months experience).15

A 9.0% non-inflation-adjusted earnings increase over about a two-year interval is not suggestive of entrepreneurial activity. It is even less impressive when one considers that workers, on average, improve their nominal (non-inflation-adjusted) earnings almost as much over a two-year period. The Federal Reserve Bank of Atlanta tracks how much the hourly earnings of different types of workers improve over a year’s time (reflecting the overall growth in earnings and one more year of experience), providing a benchmark for comparison (Federal Reserve Bank of Atlanta 2019a).16 Over the same period as the Uber driver study, 2016–2017, earnings over a year grew 4.0% for job switchers, 3.9% for low-wage workers (those in the bottom 25% of the wage distribution), and 3.7% for prime-age workers. This earnings growth translates into earnings growth over two years of 8.2% for job switchers, 8.0% for low-wage workers, and 7.5% for prime-age workers. Thus, over their first two years of driving, Uber drivers improved their earnings (up 9.0%) just a percentage point or so more than what they otherwise would have gained in other employment (ranging from 7.5% to 8.2%). That is not much entrepreneurial opportunity.

And that limited earnings growth stalls after the first two years. A figure in the Cook et al. 2018 paper shows the growth of driver earnings as experience rises to 5,000 trips, indicating that earnings grow a minimal 1–2% as drivers double their experience from 2,500 trips to 5,000 trips, a further 25 months of experience for a full-time driver. This means that in their third and fourth years of driving, Uber drivers would see their earnings rise substantially more slowly than the earnings of other workers (who, as the Atlanta Federal Reserve Bank data show, see much more earnings growth each year).

Other estimates in Cook et al. 2018 allow us to assess the extent to which particular revenue-generating strategies boost earnings growth. For instance, when the driver’s choice of work hours (when to drive) and where to drive are taken into account, the earnings after completing 2,500 trips compared with earnings after completing just 100–500 trips is reduced to 5.7%: This means that the choice of when and where to drive (including seeking “surges”) provides a very limited return to experienced drivers of just 3.3% (the difference between 9.0% and 5.7%).17 The choice of “when” to drive provides a very limited 0.35% (i.e., less than 1%) boost to earnings after 2,500 trips (relative, again, to 100–500 trips).18

Cook et al. (2018, 24) also note that, “In addition to deciding where and when to drive, drivers can affect their earnings through strategic actions. We consider two such strategic actions: rejecting dispatches and canceling trips.” It turns out, however, that “[c]ontrolling for time and geography, there is a negative impact on earnings of rejecting a dispatch or canceling a trip” (25). So driver strategies in selecting riders, to the extent that they are available, do not even raise earnings. This research makes clear that driving faster or choosing routes that allow faster driving do not raise earnings either.

Learning how to do the job better—exercising the choices available to drivers—has only a limited impact on earnings for Uber drivers: A driver with roughly two years of experience earns about 9% more than a novice driver with about three months of experience, and driving for an additional two years provides only a 1–2% further earnings bump. Basically, each driver gets to choose how many hours to drive but otherwise drivers have no options for greatly expanding their business revenues once they learn the basics. Uber drivers do not have much opportunity to grow their hourly earnings.

Uber drivers have little control over driving

As the previous section showed, the realities of Uber driving are in no way aligned with any standard conception of what it means to be an entrepreneur. But the GC memo’s misrepresentation of these realities extends to Uber drivers’ autonomy. While the GC memo states that drivers have near complete control over their cars and “near-absolute autonomy in performing their daily work,” the facts tell a different story: Drivers are denied the right to decide who to sell to, have limited ability to determine their own routes, and are essentially controlled by the company’s automated tracking and rating systems.

It is difficult to argue that drivers actually control their cars when the company requires they accept a ride without informing them of the destination. In fact, “drivers are not shown the passenger’s destination or how much they could earn on the fare” before being asked to accept a ride assignment: This is accomplished as Uber “enforces blind acceptance of passengers” through penalties or deactivation if drivers fail to have high acceptance rates (the share of rides offered that are accepted) and low cancellation rates (Rosenblat 2016). The GC memo acknowledges this, noting that drivers cannot “routinely reject trips based on expected profitability.” Uber drivers could be locked out of the app temporarily for excessively rejecting trips. Until around 2016, an acceptance rate lower than 80–90% could be a basis for terminating a driver’s relationship with Uber (NLRB GC 2019).

Drivers cannot readily turn down short rides, rides that are not close by (note that the time between dispatch and rider pickup costs time and vehicle expense and provides no revenue to the driver), or rides that go to destinations that will make it difficult to obtain a follow-up ride. When the app alerts a driver to a potential ride, the driver has a very short time—about 15 seconds—to respond as to whether they will accept the rider (Rosenblat and Stark 2016; Rosenblat 2018a). There is no other way to provide a ride through Uber (such as picking up someone hailing you on a street corner) other than this process that is governed by the algorithms and the app (Rosenblat and Stark 2016).

Nor do drivers have much flexibility in how they deliver services—not even in the route they choose to take: Uber reserves the right to retroactively adjust a fare if it determines that an inefficient route was taken. As the GC memo notes, “If a rider complained to Uber about the route taken and Uber determined, based on GPS data, that the route was inefficient, Uber would adjust the fare downward” (NLRB GC 2019). Uber has even added a new “Quiet Mode” option in Uber Black and Uber Black SUV that allows riders to request that the Uber driver not talk to them (Ghajar 2019; Cradeur 2019), an option drivers are required to offer and fulfill, underscoring how little autonomy drivers really have.

Uber drivers are supervised

The GC memo’s misrepresentation of Uber drivers’ realities also extends to their level of supervision. The GC memo falsely claims that Uber does not closely supervise drivers: “With regard to the ‘supervision’ factor, drivers operated without supervision by Uber. They did not report to supervisors and generally interacted with Uber agents only when a problem arose” (NLRB GC 2019).

Just because drivers do not talk to human supervisors does not mean they are unsupervised, as Rosenblat (2018a) amply documents. In reality, Uber “supervises” drivers through management functions accomplished through semi-automated and algorithmic systems. When drivers are on the app, the company heavily monitors their actions. Specifically, Uber tracks driver acceptance rates, time on trips, speed, driving acceleration, routes, and cancellation rates. Uber’s customer rating system further allows the company to automate the evaluation of drivers. These factors may lead to a driver being “deactivated” or kicked off the app. In reality, the algorithm evaluates drivers and serves as Uber’s management. Drivers, in fact, complain that they cannot get in contact with humans to answer questions and resolve issues.19

Uber drivers are ‘entrepreneuring’ at minimum wages—or less

The earnings of Uber drivers are perhaps the most instructive consideration in the analysis of Uber drivers’ entrepreneurial opportunity, though the NLRB’s GC essentially ignores earnings in its evaluation. Hourly earnings of Uber drivers are very low (Mishel 2018). Once you deduct Uber’s commissions (about a third or more of passenger fares), vehicle expenses, self-employment Social Security/Medicare taxes (15.3%, double that of a W-2 worker) and an allowance for a modest benefits package, an Uber driver earns a “wage” of $9–10 an hour. This is lower than the earnings of 90% of all wage earners and below the minimum wage in many major cities. Uber confirmed the expectation of very low earnings in its own Initial Public Offering (IPO) when it compared drivers’ earnings to the earnings of those in “retail, wholesale, or restaurant services or other similar work” (J. Wong 2019). Retail and restaurants are some of the lowest-wage sectors in the U.S., with hourly earnings for nonmanagerial retail and restaurant workers, respectively, 30% and 47% below the private-sector average.20

Most drivers work very limited hours. Uber, in fact, advertises for new drivers by offering a “side hustle” (Uber 2017) and reports that “nearly 60 percent of U.S. drivers use Uber less than 10 hours a week” (Kansal 2018). Most drivers do so primarily to supplement other earnings (Plouffe 2015). There is very high turnover, with 68% of drivers quitting within six months (Cook et al. 2018, Table 1). The average tenure of a driver is three months. Uber driving does not appear to be a business opportunity in which people are growing remunerative small businesses. Not all Uber drivers have limited, part-time hours, however. There are many full-time Uber drivers (defined as driving more than 30 hours each week), and they provide nearly half of the overall rides provided by Uber (Chen et al. 2019).21

Uber’s method of payment exerts control over drivers, further supporting a finding of employee status

As the previous sections have shown, contrary to what the GC memo claims, Uber drivers are not entrepreneurs who control their work driving and act without supervision. Even when the GC memo’s depiction of the realities of Uber driving comes closer to reality, the evidence is then discounted as unimportant.

A prime example involves Uber’s method of payment and its impact on driver classification. Traditionally, a flat fee method of payment supports a finding of independent contractor status, whereas a commission-based method of payment supports a finding of employee status.22 The rationale is that in a flat fee arrangement, it can be inferred that the putative employer company has little motivation to exert control over drivers—as the cost to the company remains the same regardless of how drivers operate. Conversely, a commission-based arrangement creates an inference that the company has reason to exert control over drivers—company profits depend on it.

The GC describes Uber’s payment arrangement as the company retaining “a percentage of fares paid by riders rather than charging drivers a flat fee for the opportunity to use the App.” However, while the GC memo essentially describes Uber’s payment arrangement as commission-based, it simply disregards these considerations and concludes that “the inferences behind the method-of-payment analysis may be overcome by the facts of a particular case. This is such a particular case” (NLRB GC 2019).

Contrary to the GC’s description, Uber’s current method of payment is neither a flat fee nor a pure commission-based model. Uber uses a system of “up-front pricing” in which the company charges riders a predetermined rate based on what the company estimates the trip should cost. But drivers are paid based on the actual miles and minutes of the trip. So Uber may charge a rider $75 for a trip for which a driver earns $30. The transactions are not transparent, at least not to the drivers who rely on information provided though the app and who are not informed in advance of the trip what passengers pay. So, drivers have no way of knowing in advance of committing to the ride the sometimes wide discrepancies between their earnings and the fares the company charges for a particular trip.23

The GC memo does not get the facts of this payment system right in the memo, though it is correct in likening Uber’s method of payment more to a commission-based system than a flat-fee arrangement. While Uber uses neither an exclusively flat fee nor a commission-based payment system, the company clearly retains a financial incentive to control drivers. Uber’s upfront pricing creates a system in which the company has a financial stake in each trip. The company wants to be sure that the fare charged is more than the driver is paid—and it is free to push the limits. In reality, the formula Uber uses to calculate a driver’s earnings is a means of control over drivers. Consider that if a driver takes longer than Uber calculates she should on a given trip, the formula used to calculate her earnings from the trip reduces her pay.

And Uber’s bonuses and promotions give the company additional ways of controlling the way drivers drive and their work schedules. The company provides wage “promotions” for some drivers. Dependent on the driver’s city, the promotions include a cash award for driving a set number of trips or “boost” fares—higher fares in certain locations during busy times of the day (Uber 2019b). And the company has announced that it would offer “bonuses” ranging from $100 to $10,000 to long-serving drivers (Streitfeld 2019).

All of these considerations would be at the heart of a serious evaluation of Uber’s method of payment and how it affects drivers’ classification. Uber’s method of payment points to drivers being covered employees.

Freedom to earn does not overcome control issues

The GC memo minimizes the importance of the realities of Uber driving—such as the methods of payment, the lack of driver control, and failure to have true entrepreneurial opportunity—by invoking driver freedom to choose how to earn:

Drivers’ unlimited freedom to look elsewhere for better earnings also minimized the impact that certain other features of the Uber system would otherwise have on their entrepreneurial opportunity.

The GC memo expands on this and says:

On any given day, and, indeed, at any free moment, drivers could decide how best to serve their own economic objectives: by accepting ride requests through the App, working for a competing ride-share service, or pursuing a different venture altogether. (NLRB GC 2019)

So according to the GC, even though, in reality, Uber exerts key control over earnings obtained “by accepting ride requests through the App,” the fact that drivers can work “for a competing ride-share service” or pursue “a different venture altogether” somehow makes Uber’s control irrelevant. This reasoning stands in stark contrast to the Restatement of the Law, Employment Law, which posits, “The key question is whether a service provider functions as an independent business while performing services on the principal’s behalf” (Estreicher et al. 2015). This suggests that the key question with regard to Uber driver classification is how a driver relates to Uber and not whether the driver can turn on the Lyft app or start a new venture.

One can only respond to the GC’s logic with a combination of bemusement and bewilderment. The matter under discussion is whether Uber drivers should be considered employees of Uber with NLRA rights or be considered contractors obtaining business through the Uber app. How exactly does the opportunity to drive for Lyft or to pursue some self-employment activity affect this determination? Setting aside actual limits on driving for competitors,24 common sense dictates that the freedom to pursue other earnings opportunities has no bearing on the determination of Uber driver work status except in one sense: If Uber could block a driver’s ability to moonlight or seek her own business, then this would weigh in favor of that driver being considered an employee rather than an independent contractor. However, simply having the freedom to moonlight (i.e., drive for Lyft) or start a business does not in itself make someone an independent contractor, or even put them in a gray area. After all, every part-time worker in a job covered by the NLRA that does not have a noncompete agreement has a similar freedom to moonlight or to purse “a different venture” beyond his or her regular job. If one accepts the GC’s logic, then nearly all part-time wage earners covered by the NLRA would be deemed to have “significant control over their earnings” and not be subject to their employer’s control and, therefore, likely to be considered an independent contractor.

The GC’s argument about Uber contracts and lack of access to benefits also does not hold up under scrutiny

The GC argues that “both parties understood their relationship to be one of independent contractors” because the “drivers’ contracts explicitly characterized the relationship this way.” In reality, the Board and the courts have routinely held that these kinds of provisions are not legally dispositive and carry little weight in the employee–independent contractor analysis. Further, independent contractor misclassification is a persistent problem. A 2000 study commissioned by the U.S. Department of Labor found that between 10% and 30% of audited employers misclassified workers and that up to 95% of workers who claimed they were misclassified as independent contractors were reclassified as employees following review (Planmatics 2000). To argue that agreeing to be misclassified would support a determination that the worker is an employee under the NLRA demonstrates a fundamental lack of understanding of workplace realities. In reality, agreements like Uber’s are “contracts of adhesion,” which afford workers no opportunity or leverage to negotiate over provisions or to object to contractual language, even if that language conflicts with controlling law. Consider that in California, nearly 20% of all workers are subject to noncompete agreements, despite the fact that, with limited exceptions, noncompetes are not enforced in that state (U.S. Treasury 2016).

Additionally, the GC memo states that the fact that Uber provided drivers “no benefits, paid leave, or holiday pay” supports independent-contractor status. This assertion reveals a similar lack of awareness of the modern workplace. Consider that over a quarter of all private-industry workers lack access to employer-sponsored health care benefits (BLS-EBS 2018a). Further, nearly a third of all private-industry workers do not have access to a single day of paid sick leave (BLS-EBS 2018b). To imply that a corporation’s unwillingness to provide its workers with access to benefits means that those workers can be classified as independent contractors ignores the reality of the modern workplace. Again and again, the GC gives undue weight to considerations unsupported by the facts to reject an accurate analysis of Uber drivers’ realities and how they reflect employee status.

Conclusion

The GC memo gets the realities of driving for Uber wrong. In reality, Uber drivers do not experience entrepreneurial opportunity. Drivers have no control over the fares Uber charges passengers. Drivers have no ability to grow their business through marketing strategies or subcontracting. In fact, academic research co-authored by Uber’s chief economist demonstrates that drivers have very limited ability to improve their own earnings. Uber monitors drivers while they are on the app and penalizes them for being selective in accepting rides or for taking a route different from Uber’s determined route. Uber supervises drivers through algorithmic controls, including an automated passenger rating system that has meaningful consequences for drivers—including potential deactivation from the app.

The GC memo has little to do with reality—or with meaningful legal analysis. Instead, the memo is the NLRB chief prosecutor’s adoption of a dangerous political argument that companies like Uber have long been making. The argument is that Uber and other digital companies should not have to operate in the world of workplace rules and regulations by virtue of the digital landscape they occupy. These companies want policymakers to focus on the technology they commercialize as opposed to the services they provide. These services—whether rides, errands, or household help—are provided by working men and women whose experience of that work is no different from the experience of workers in the traditional economy; the only difference is that they are assigned the work via an app as opposed to a dispatcher. Like so many workers in the modern economy, the system in which these drivers work is rigged against them and in favor of a business model that relies on paying low wages and denying all responsibility for the workforce. The GC’s advice memo is another example of this. Instead of conducting a fair, fact-based inquiry into the realities of Uber driving to determine if drivers are employees under the NLRA, the GC gets the facts wrong in favor of further rigging the system against workers.

About the authors

Lawrence Mishel is a distinguished fellow and former president of the Economic Policy Institute. He is the co-author of all 12 editions of The State of Working America. His articles have appeared in a variety of academic and nonacademic journals. His areas of research include labor economics, wage and income distribution, industrial relations, productivity growth, and the economics of education. He has a Ph.D. in economics from the University of Wisconsin at Madison.

Celine McNicholas is EPI’s director of government affairs and labor counsel. An attorney, her current areas of work include a wide range of workers’ rights issues, including labor and employment law, collective bargaining, and union organizing. Before joining EPI in 2017, McNicholas served as director of congressional and public affairs and as special counsel for the National Labor Relations Board (NLRB). She has a J.D. from Villanova University School of Law.

Acknowledgments

The authors thank Jennifer Abruzzo, Wilma Liebman, Lynn Rhinehart, Rebecca Smith, and Alex Rosenblat for comments on this report.

Endnotes

1. 29 U.S.C. § 157.

2. 29 U.S.C. § 152(3).

3. In addition to collective bargaining rights under the NLRA, workers who are classified as employees enjoy protections under many other worker protection measures including minimum wage and overtime protections, unemployment insurance, and workers’ compensation. And employers are responsible for payroll tax contributions that fund Social Security and Medicare. Independent contractors are not eligible for those benefits and protections and must pay all required taxes themselves.

4. NLRB v. United Insurance Company of America, 390 U.S. 254 at 256 (1968); NLRB v. United Insurance Company of America, 390 U.S. 254 at 256 (1968); NLRB v. United Insurance Company of America, 390 U.S. 254 at 256 (1968); NLRB v. United Insurance Company of America, 390 U.S. 254 at 256 (1968); NLRB v. United Insurance Company of America, 390 U.S. 254 at 256 (1968).

5. NLRB v. United Insurance Company of America, 390 U.S. 254 at 258 (1968).

6. NLRB v. United Insurance Company of America, 390 U.S. 254 at 258 (1968).

7. The common-law factors include, inter alia, “the extent of control which, by the agreement, the master may exercise over the details of the work”; “the kind of occupation”; whether the worker “supplies the instrumentalities, tools, and the place of work”; “the method of payment, whether by the time or by the job”; whether “the work is a part of the regular business of the employer”; and the intent of the parties. The multifactor test is, in lay terms, the employer’s right to direct how, where, and when a worker does her job.

8. FedEx Home Delivery v. NLRB, 563 F.3d 492, 497 (D.C. Cir. 2009), quoting Corporate Express Delivery Systems v. NLRB, 292 F.3d 777, 780 (D.C. Cir. 2002).

9. FedEx Home Delivery v. NLRB, 563 F.3d 492, 497 (D.C. Cir. 2009), quoting Corporate Express Delivery Systems v. NLRB, 292 F.3d 777, 780 (D.C. Cir. 2002). The dissent pointed out flaws in this approach and noted that the majority was wrong in its review of the NLRB’s case law on this issue. Specifically, the NLRB had not found “control” to be an “animating principle” in the evaluation of worker classification. Further, the dissent emphasized the factual finding that there was “little room for the contractors to influence their income through their own efforts or ingenuity” and that any abstract “rights” to make entrepreneurial decisions, like the use of trucks for shippers other than FedEx or the sale of a route for profit, were in practice prevented.

10. FedEx Home Delivery, 361 NLRB 610 (2014).

11. FedEx Home Delivery v. NLRB, 849 F.3d 1123 (D.C. Cir. 2017).

12. In a legal challenge to drivers’ classification in the United Kingdom for unemployment insurance purposes, an employment tribunal (Aslam v. Uber BV) concluded, “The notion that Uber in London is a mosaic of 30,000 small businesses linked by a common ‘platform’ is to our minds faintly ridiculous. In each case, the ‘business’ consists of a man with a car seeking to make a living by driving it. [Uber representative] Ms. Bertram spoke of Uber assisting the drivers to ‘grow’ their businesses, but no driver is in a position to do anything of the kind, unless growing his business simply means spending more hours at the wheel.” The tribunal’s analysis documents the realities of Uber driving and, among other analyses, asks whether drivers have the ability to make the decisions that small business owners regularly make (U.K. Judiciary 2016, 90).

13. The Uber Rasier Technology Services Agreement (last updated December 11, 2015), Section 2.2, “Provision of Transportation Services,” states: “You shall not contact any Users or use any User’s personal data for any reason other than for the purposes of fulfilling Transportation Services.”

14. Cook et al. (2018) provide data in their Table 1 showing average trips per week (29.83) and average hours per week (17.06): This implies drivers provide 1.75 trips per hour. Assuming a full-time driver drives 40 hours each week, we can then convert the number of trips (the midpoint of the varying ranges) into hours driven, weeks driven, and months of driving experience.

15. Cook et al. (2018) estimate the effects of experience (number of trips completed) on Uber driver earnings. They look at drivers at five different experience levels: less than 100 trips, 100–500, 500–1,000, 1,000–2,500, and more than 2,500. Being in the 100–500 trips range yields a 5.4% earnings bump over having completed less than 100 trips. We choose to assess the impact of experience by comparing those with more than 2,500 trips to those with 100–500 trips, discounting the initial bump as the driver learning the basics. This is measured as the difference between the coefficient (see Cook et al. 2018, Table 6, column (1)) on 2,500 trips (0.1391) and the coefficient on 100–500 trips (0.0530)—a difference of 0.0861. Because the dependent variable is log hourly earnings, we exponentiate 0.0861 and subtract 1.0 to obtain 9.0%. The more experienced Uber drivers (those with 2,500+ trips) are earning roughly 9.0% more than the less experienced drivers (those with 100–500 trips).

16. As the bank notes on its Wage Growth Tracker page, “The Atlanta Fed’s Wage Growth Tracker is a measure of the nominal wage growth of individuals. It is constructed using microdata from the Current Population Survey (CPS), and is the median percent change in the hourly wage of individuals observed 12 months apart. Our measure is based on methodology developed by colleagues at the San Francisco Fed….The methodology is broadly similar to that used by Daly, Hobjin, and Wiles (2012). The earnings data are for wage and salary earners, and refer to an individual’s main job (earnings data are not collected for self-employed people). Earnings are pretax and before other deductions. The Census Bureau reports earnings on either a per-hour or a per-week basis. We convert weekly earnings to hourly by dividing usual weekly earnings by usual weekly hours or actual hours if usual hours is missing” (Federal Reserve Bank of Atlanta 2019b).

17. This can be seen by comparing the results in Cook et al. 2018, Table 6, column (4)—in which location and hours of work are controls—with the results in column (1).

18. The impact of “where” to drive can be seen in the difference in the earnings of those with more than 2,500 trips versus those with 100–500 trips, comparing the results in column (1) versus column (2) and column (3) versus column (4) in Cook et al. 2018, Table 6.

19. A quick perusal of the blog www.rideshareguy.com reveals these and other common complaints by Uber drivers.

20. Calculations are for average hourly earnings of production/nonsupervisory workers from Bureau of Labor Statistics establishment payroll data for February 2019 (BLS-CES 2019).

21. Table 1 in Chen et al. 2019 provides data on the share of driver-weeks by weekly hours driven (0, 1–4, 5–12, 13–20, 21–30, 31–40, 41+) by “active drivers,” those who provided rides for at least one hour in at least 16 of the 36 weeks in which data were collected, during the period from September 2015 to April 2016. These data can be used to derive the total number of hours driven by each hours group (using the midpoint of each range, and assigning the zero hours group at zero hours) and, accordingly, the share of total hours driven by each hours group. Doing so indicates that those driving at least 31 hours each week provided 45.6% of all hours.

22. See Restatement (Second) of Agency §220(2)(g).

23. Uber does show drivers what passengers paid in the drivers’ earnings stubs.

24. Lyft has apparently decided to directly interfere with its drivers’ “freedom to earn” by blocking the use of the third party switching app, Mystro, which lets “drivers drive for both Uber and Lyft automatically” and safely by setting preferences and auto accepting rides, closing out the other app so drivers can focus on their current fare. See Gibbins 2019, which describes how switching apps work: “Simply put, switching apps like Mystro let drivers drive for both Uber and Lyft automatically and they let drivers set terms for preference as far as which app they’d prefer to accept rides from, as well as auto accepting rides and closing out the other app so drivers can focus on what they need to be doing—driving safely. The driver can even tell the app to reject any passenger request under a certain rating. But one thing we’ve been surprised by is how many drivers like using Mystro purely for the safety aspect, since without it, you have to manually accept/reject trips that could come in while driving 45 mph down the street.”

References

Bureau of Labor Statistics, Current Employment Statistics (BLS-CES). 2019. “Table B-3a. Establishment Data: Average Hourly and Weekly Earnings of All Employees on Private Nonfarm Payrolls by Industry Sector, Seasonally Adjusted.” Last modified May 3, 2019.

Bureau of Labor Statistics, Employee Benefits Survey (BLS-EBS). 2018a. “Table 9. Healthcare Benefits: Access, Participation, and Take-Up Rates, Private Industry Workers, March 2018.”

Bureau of Labor Statistics, Employee Benefits Survey (BLS-EBS). 2018b. “Table 32. Leave Benefits: Access, Private Industry Workers, March 2018.”

Campbell, Alexia Fernández. 2019. “Google Employees Fought for Their Right to Sue the Company—and Won.” Vox, February 22, 2019.

Chen, M. Keith, Judith A. Chevalier, Peter E. Rossi, and Emily Oehlsen. 2019. “The Value of Flexible Work: Evidence from Uber Drivers.” NBER Working Paper no. 23296, revised June 2019.

Cook, Cody, Rebecca Diamond, Jonathan Hall, John A. List, and Paul Oyer. 2018. The Gender Earnings Gap in the Gig Economy: Evidence from over a Million Rideshare Drivers. NBER Working Paper no. 24732, June 2018.

Cradeur, Jay. 2019. “Uber Tells Drivers to Shut Up with Quiet Mode” (blog post). The Rideshare Guy, May 20, 2019.

Daly, Mary C., Bart Hobijn, and Theodore S. Wiles. 2012. “Dissecting Aggregate Real Wage Fluctuations: Individual Wage Growth and the Composition Effect.” Federal Reserve Bank of San Francisco Working Paper no. 2011-23, May 2012.

Estreicher, Samuel, Matthew T. Bodie, Michael C. Harper, and Stewart J. Schwab. 2015. Restatement of the Law, Employment Law. Philadelphia: The American Law Institute, 2015.

Federal Reserve Bank of Atlanta. 2019a. “Wage Growth Tracker” (data set). Accessed September 2019.

Federal Reserve Bank of Atlanta. 2019b. “Wage Growth Tracker Methodology” (web page). Accessed September 2019.

Ghajar, Aydin. 2019. “Introducing an Enhanced Uber Black Experience.” Uber Newsroom, May 15, 2019.

Gibbins, Paula. 2019. “Why Is Lyft Blocking Third Party Apps Like Mystro?” (blog post). The Rideshare Guy, August 7, 2019.

Kansal, Sachin. 2018. “Another Step to Prevent Drowsy Driving.” Uber Newsroom, February 12, 2018.

Mishel, Lawrence. 2018. Uber and the Labor Market: Uber Drivers’ Compensation, Wages, and the Scale of Uber and the Gig Economy. Economic Policy Institute, May 2018.

National Employment Law Project (NELP). 2018. Impact of the Fight for $15: $68 Billion in Raises, 22 Million Workers (data brief). November 2018.

National Labor Relations Board, Office of the General Counsel (NLRB GC). 2019. Advice Memorandum, Subject: Uber Technologies Cases 13-CA-163062 et al. April 16, 2019.

Planmatics, Inc. 2000. Independent Contractors: Prevalence and Implications for Unemployment Insurance Programs. Report prepared for the U.S. Department of Labor Employment and Training Administration, February 2000.

Plouffe, David. 2015. “Uber and the American Worker.” Medium, November 5, 2015.

Rhinehart, Lynn. 2019. “Let’s Not Forget Unions and Collective Action When Discussing Victories on Workers’ Rights.” Working Economics Blog (Economic Policy Institute), February 11, 2019.

Rosenblat, Alex. 2016. “The Truth About How Uber’s App Manages Drivers.” Harvard Business Review, April 6, 2016.

Rosenblat, Alex. 2018a. Uberland: How Algorithms Are Rewriting the Rules of Work. Oakland, Calif.: Univ. of California Press.

Rosenblat, Alex. 2018b. “When Your Boss Is an Algorithm.” New York Times, October 12, 2018.

Rosenblat, Alex, and Luke Stark. 2016. “Algorithmic Labor and Information Asymmetries: A Case Study of Uber’s Drivers.” International Journal of Communication 10 (2016): 3758–3784.

Streitfeld, David. 2019. “He Has Driven for Uber Since 2012. He Makes About $40,000 a Year.” New York Times, April 12, 2019.

Uber. 2017. “Side Hustle: Earning” (TV commercial). Last aired September 24, 2017, viewable on iSpot.tv.

Uber. 2019a. “Become and Uber Driver—3 Things to Know About Driving” (web page). Accessed August 6, 2019.

Uber. 2019b. “How Weekly Promotions Work for Drivers” (web page). Accessed August 2019.

U.K. Courts and Tribunals Judiciary (U.K. Judiciary). 2016. Employment Tribunals Between Aslam, Farrar et al. and Uber B.V. et al. (Case No. 2202550/2015 et al.): Reasons for the Reserved Judgment on Preliminary Hearing Sent to the Parties on 28 October 2016.

U.S. Department of the Treasury (U.S. Treasury). 2016. Non-Compete Contracts: Economic Effects and Policy Implications. March 2016.

Wong, Julia Carrie. 2019. “Disgruntled Drivers and ‘Cultural Challenges’: Uber Admits to Its Biggest Risk Factors.” Guardian, April 12, 2019.

Wong, Kristin. 2018. “Should You Tip Your Uber Driver? If So, How Much?” New York Times, October 2, 2018.