Customers sure love Uber.  If you ask them to describe their experience with the ride-share firm, most Uber passengers will gladly tick off a long list of superlatives: Innovative! Economical! Revolutionary!

But a less-flattering picture of Uber has recently surfaced in courtrooms across the country.  Told by aggrieved drivers, this countervailing narrative depicts Uber as a company that cheats its workers out of wages and denies them basic workplace rights.  In fact, earlier this year, Uber agreed to pay upwards of $100 million to drivers in California and Massachusetts for alleged employment law violations.

So which is it?  Is Uber a boon to passengers or just another abusive employer wrapped in an app?

Although the question appears in a novel setting (the sharing economy) the legal issue is actually a familiar one to employment lawyers: whether the law should treat workers as “employees” or “independent contractors.”  Because only “employees” enjoy most workplace rights such as overtime and antidiscrimination protections, businesses like Uber attempt to avoid these responsibilities by denying their status as employers.

The independent contract categorization not only separates workers from employment protections it also denies them basic workplace benefits as well.  As putative “independent contractors” Uber drivers do not receive unemployment insurance, workers’ compensation, or employer contributions to Social Security.  The ride-share firm gains immediate economic advantages from this strategy, as it can save up to thirty percent in payroll costs simply by classifying its drivers as nonemployees.

Although Uber is the most visible firm engaged in this practice, up to thirty percent of American businesses misclassify their workers as “independent contractors.”  But just because a company calls its workers “independent contractors” does not mean that the law will actually treat them as such.  Although the tests for determining employee status differ somewhat between the states, the concept of control usually sits at the center of most judicial analyses of the topic.  The more control that a firm exercises over working conditions, the more likely that the worker is an employee.

Many companies such as Uber disclaim their status as employers by asserting that they do not exercise daily, direct control over workers.  But such a binary approach to control unnecessarily constrains the meaning of employment.  In fact, employment status has never depended on whether firms control the minutia of workplace details.  Rather, businesses today become employers when they meaningfully influence working conditions, even if contractual layers obscure that power.  Courts can adopt this broader understanding of employment by shifting the focus of any control-based analysis from daily supervision to a firm’s influence over all working conditions.  For example, if a large company such as Uber controls key aspects of its relationship with so-called “independent contractors” (such as the manner in which a service is delivered, the price paid for the service, and the timeline for delivering the service) then this direct control over these labor-based outcomes establishes effective control over the manner and means of work.

In addition to expanding the subjects of control, courts should also evaluate the direction that control travels between firms and workers.  When a business reserves most relevant forms of power over workers (such as how they do their jobs or how much they get paid), this type of one-way control gives rise to employment responsibilities given that workers have no meaningful say over their jobs.  Conversely, evidence of bidirectional control—in which workers and firms codetermine working conditions—suggests that workers are more like entrepreneurs and less like employees.

Reflecting the broader vision of control outlined here, two recent federal court decisions involving ride-share defendants provide a roadmap for evaluating employment relationships in the sharing economy.  Last year, in O’Connor v. Uber Technologies, Inc., a federal judge in California held that employment determinations involving ride-share drivers should focus less on drivers’ control over hours and more on Uber’s control over drivers once they log on the Uber app.

By examining other ways in which Uber exercised control over drivers beyond scheduling, the O’Connor court expanded the relevant subjects of control.  For example, the court noted that Uber set the price that customers paid for rides and prohibited drivers from negotiating their own wages.  In addition, the court explained how Uber maintained a detailed performance protocol that directed drivers to dress professionally, send clients texts, open doors for clients, and turn the radio off or play soft jazz or NPR.

The O’Connor court explained how Uber monitored its “partners’” compliance with these mandates with customer star ratings and made deactivation decisions based on this feedback.  By casting customers as supervisors, Uber retained a great deal of control over drivers, just as traditional employers monitor their employees’ performance.

Like Uber, the ride-hailing company Lyft recently agreed to settle a class action brought by drivers soon after receiving an unfavorable court ruling on the issue of employee misclassification (both the Lyft and Uber settlements are currently pending court approval).  In Cotter v. Lyft, Inc., a federal court in California explained how Lyft gave drivers its “Rules of the Road,” which directed them to keep their cars clean, help passengers with luggage, hold umbrellas for passengers, not accept cash, play the passengers’ preferred type of music, and greet passengers with a fist bump and smile.

Although Lyft exercised a great deal of control over the workers when they were on duty, drivers also enjoyed many freedoms such as the ability to choose their work hours and select the neighborhoods where they would drive.  Based on these factors, the Cotter court held that the control analysis favored the drivers, but not so much as to conclusively categorize them as Lyft’s employees.

As these two leading decisions from the sharing economy made clear, peer-to-peer work can expand the control analysis in several new directions.  For example, on the issue of scheduling, ride-share drivers behave somewhat like independent contractors who choose their own hours.  On the other hand, once drivers report to work, on-demand platforms expect them to accept new assignments and reserve the right to deactivate members who decline such assignments—a far cry from the type of entrepreneurial power retained by independent contractors who can accept and reject jobs as they see fit (not to mention negotiate their own pay rates).

The sharing economy also expands the possible subjects and methods of control.  For example, although no supervisor physically observes drivers’ work, the star-rating system used by ride-hailing platforms enables firms to monitor and enforce their conduct codes.  Finally, the direction of control travels differently with peer-to-peer work.  While sharing platforms hold unidirectional control over some subjects (pay, performance, and supervision), control takes a bidirectional course with regard to other subjects (hours and work location).

The Uber and Lyft litigation represents only the beginning of a much larger wave of cases brought by peer-to-peer workers.  Indeed, soon after announcing settlements in California and Massachusetts earlier this year, Uber was hit with a nationwide lawsuit involving drivers in forty-eight other states.

As both the O’Connor and Cotter decisions make clear, companies that disclaim their status as employers may still effectively control workers’ daily existence.  Whether it is Uber telling its drivers to play soft jazz or Lyft ordering its drivers to greet passengers with a fist bump, businesses that control contractual outcomes frequently control working conditions as well.  Given these realities, courts must fully assess all aspects of workplace control, from its direction to its diverse subjects.  Armed with this refocused vision of control, courts can assign responsibility to firms that retain ultimate authority over the manner and means of modern work.

The preceding post comes to us from Keith Cunningham-Parmeter, Professor of Law at Willamette University.  The post is based on his paper, which is entitled “From Amazon to Uber: Defining Employment in the Modern Economy”, forthcoming in the Boston University Law Review and available here.