For the most part the assertions of the need for a new category have had little substance beyond a claim that technology has changed the nature of the work that gig workers (or “app workers”) perform, making current legal categories antiquated. But a recent proposal by Seth Harris and Alan Krueger for a third status called “independent worker” (Harris and Krueger 2015a) has elevated this policy discussion in several ways. First, they provide specific empirical claims as to why they believe the “[n]ew and emerging work relationships arising in the ‘online gig economy’ do not fit the existing legal definitions of ‘employee’ or ‘independent contractor’ status.” Harris and Krueger offer Lyft and Uber drivers as exemplars of these new relationships. Second, their proposal delineates the legal protections that so-called independent workers should have and those that do not fit gig work. Harris and Krueger argue that clarifying gig workers’ legal status would facilitate innovation and would provide greater protections for workers than the courts might if forced to choose between two ill-fitting categories.
In this paper we examine Harris and Krueger’s empirical claim that the “immeasurability of work hours” for gig workers places them in a gray area between employee and independent contractor and negates the possibility of applying the Fair Labor Standards Act (FLSA) to work done through some digital apps. Denial of these rights is central to their policy proposal and of major consequence, but the labor protections that so-called independent workers would not enjoy compared with bona fide employees include, in addition to coverage by the FLSA, rights under the National Labor Relations Act, state workers’ compensation laws, and unemployment insurance.
This paper is limited to the issues of measuring and controlling drivers’ hours and the implications for establishing the need for a third status of independent worker. We also challenge Harris and Krueger’s proposal to deny certain protections to independent workers. This paper does not examine all of the issues required to assess whether Uber drivers are employees rather than independent contractors, as do Harris and Krueger (2015a), Sachs (2015), and Rosenblat (2016). We do believe for a host of reasons that Uber drivers are employees—for example, they don’t set their own fares or freely choose their own customers, their performance is measured and controlled by Uber, their driving is essential to Uber’s business, and the economic reality is that they are not independent businesses but small cogs in Uber’s powerful multinational business. But we leave this analysis for further work and note that early court and state agency rulings are coming down on both sides of the issue.
We find:
- The claim that the immeasurability of work hours for some digital app relationships requires a third employment status other than employee or independent contractor is empirically flawed. Uber is the leading case, and Uber can and does measure the time drivers have their apps on, to the minute. Uber has a guaranteed wage program that demonstrates that hours tracking and minimum-wage obligations can be administered effectively. Uber takes disciplinary action for having a low acceptance rate, so drivers must respond quickly to ride requests or face serious consequences. Therefore, drivers are free to choose when they turn on the app but are not free to ignore the app and do personal chores when it is turned on.
- A driver can be an employee regardless of whether waiting time is compensable or charged to any particular employer; just because that problem seems complicated doesn’t mean we need a new category of worker. Moreover, there is an easy solution to the problem of allocating waiting time among one or more employers: the employer whose driver accepts the ride pays for the waiting time leading up to the ride.
- We disagree with the proposal to deny minimum-wage and overtime protections to Uber drivers and see no need for a third employment status. Our conclusion that Uber exerts substantial controls over a driver’s time while the driver is on the app also has implications for whether Uber drivers should be considered employees. Rather than pursue a legislative fix along the lines offered by Harris and Krueger, a better approach is simply to establish that Uber and Lyft drivers and similar workers are employees with all attendant rights.
Are work hours immeasurable for app workers?
One of Harris and Krueger’s principal claims for why some work relationships governed by digital are qualitatively “new” and therefore require a third status of independent worker is that, because “[a] worker in the online gig economy could be primarily engaged in personal tasks while one or more intermediaries’ apps are turned on,” any determination of work hours for a particular employer is impossible (Harris and Krueger 2015a, 13). They ask us to imagine a driver who has apps open for both Lyft and Uber while also conducting personal tasks, and point to the difficulty of determining which employer, if any, is associated with the period of waiting time before a ride is requested. (See the appendix for excerpts from the Harris and Krueger paper and the follow-up event discussing the findings.)
The work-hour measurability issue presented by Harris and Krueger is both an empirical and a legal matter. We discuss the legal issues in the next section. On the empirical front, we examine the situation of Uber drivers—the prototypical gig workers—and find that the Harris and Krueger description of the work situation avoids three important details:
- Both Lyft and Uber track their drivers’ work hours and readily identify drivers’ hours in many public statements.
- Uber drivers cannot keep their app on and monitoring potential riders and refuse to accept rides without incurring serious consequences, including being deactivated (i.e., fired) for having too low an “acceptance rate.”
- It is clearly possible for Uber work hours to be measurable for the establishment of a minimum wage. In fact, since mid-2014, Uber has had a “guaranteed pay” system that functions like a minimum wage and that demonstrates the company’s ability to apply the FLSA in its work relationship with drivers.
We discuss these three points in turn.
Uber and Lyft already measure drivers’ work hours and describe them in public statements
Uber and Lyft have no problem measuring driver work hours when they present information to the public. For instance, David Plouffe, Uber chief advisor and board member, in a speech in November 2015 noted:
On average, half of all drivers in the U.S. drive fewer than 10 hours a week. More than 40 percent drive fewer than 8 hours per week. And the average number of hours driven continues to fall; in fact, it’s down more than 10 percent since the beginning of the year. This is crucial: for most people, driving on Uber is not even a part-time job…it’s just driving an hour or two a day, here or there, to help pay the bills. (Plouffe 2015)
In early 2015, Alan Krueger published a paper with Uber Research Director Jonathan Hall that was based on “comprehensive data on driver-partners’ trips, fares, and time using the Uber app” (Hall and Krueger 2015). The measurement of work hours was based on the time the Uber app was turned on. There is no discussion in the paper indicating that this hours metric was uncertain or biased in any way.
In Table 2 of their paper, Hall and Krueger present tabulations of the “Distribution and Median Hourly Earnings of UberX Driver-Partners by Hours Worked,” breaking the sample into the following ranges of work hours: 1–15; 16–34; 35–49; and 50 and over. They also analyze changes in work hours and find:
Figure 8 [“Distribution of Changes in Work Hours from Week to Week”] shows that driver-partners vary the number of hours in which they use the Uber platform by a considerable amount from week to week. In any given week, well more than half (65 percent) of driver-partners drive more than 25 percent, or less than 25 percent, than the amount they drove in the previous week. Only 17 percent of driver-partners tend to drive within 10 percent of the amount of time that they drove in the previous week. (Hall and Krueger 2015)
Last, the authors use the hours data to compute hourly earnings of Uber drivers for comparison with the hourly wages of taxi drivers (presented in their Table 6).
Harris and Krueger’s analysis of work time suggests that an hour with an app turned on oversates the work done for Uber because it includes time waiting to be engaged (for a ride request). If so, Krueger and Hall’s analysis overstates Uber work hours and understates Uber’s hourly gross pay.2
As for Lyft, in its press release regarding a court settlement in January 2016, the company noted: “Roughly 80 percent of drivers who use the Lyft platform drive 15 hours per week or less to supplement their incomes.”
The ‘acceptance rate’ threat means that when a driver’s app is on, the driver is ‘engaged to wait’
In the actual work lives of Uber drivers they cannot turn on their Uber app and ignore ride requests while they do personal chores without potentially serious consequences. This is because Uber monitors each driver’s acceptance rate, reports it to drivers weekly, and dismisses drivers who do not maintain a sufficient rate. Moreover, being online with the app is not a passive experience. Uber will signal a driver if there is a rider it wants the driver to pick up by pinging the driver’s phone with beeps and flashing the screen. If the ride request is not accepted within 15 seconds, Uber treats it as refused and lowers the driver’s acceptance rate. Uber is clear that it expects drivers to engage most ride requests (see its training video) and disciplines drivers who do not do so (Thawfeek 2015). A high acceptance rate is obviously important for Uber’s interest in providing quick service.
On its website Uber provides information on the acceptance rate:
Acceptance rates are calculated as a percentage of the total number of requests you accept out of those sent to you while online. Maintaining a high acceptance rate keeps the Uber system reliable for riders and drivers. You should accept at least 80% of trip requests to retain your account status. (Uber Technologies)
Uber gives drivers warnings when their acceptance rates fall below 80 percent, and if rates don’t improve quickly the driver is deactivated (Mishel and Eisenbrey correspondence with Harry Campbell, “The Rideshare Guy”).
According to Rosenblat and Stark (2015), drivers are given their acceptance rate weekly, along with a report of their hours online, passengers’ ratings, number of trips, and other information. The authors describe the effect of these control measures:
Uber’s system enforces blind acceptance of passengers by drivers, who are not shown the passenger’s destination or how much they could earn on the fare. Drivers risk “deactivation” (being suspended or removed permanently from the system) for cancelling unprofitable fares. The Uber system requires drivers to maintain a low cancellation rate, such as 5% in San Francisco (as of July 2015), and a high acceptance rate, such as 80% or 90%. Drivers absorb the risk of unknown fares, even though Uber promotes the idea that they are entrepreneurs who are knowingly investing in such risk. (Rosenblat and Stark 2015, 9)
In July 2015 the “Ride Sharing Driver” blog, an information resource for drivers, posted the “top reasons why Uber drivers get deactivated” (DougH 2015). The third reason (of seven) was:
You don’t accept enough rides. You’ll be deactivated if you let too many ride requests expire. Uber typically warns drivers about low acceptance rates before deactivation happens, so keep an eye out for emails and texts from Uber. (DougH 2015, emphasis in original)
Uber’s Australian newsroom notes that:
Part of providing a 5 star service is providing a reliable ride. Even if your rating is very good, if you’re not accepting every trip request that comes through to you you’re making the Uber system unreliable. You should be aiming for an acceptance rate of at least 90%. If your acceptance rate is below average you may lose access to the Uber Partner app. (Denman 2014)
In O’Connor v. Uber Techs. (N.D. Cal. 2015), a federal district court examined the evidence on the record concerning Uber drivers’ ability to reject a ride request:
For instance, while Uber argues that drivers never actually have to accept ride requests when logged in to the Uber application, Plaintiffs provided an Uber Driver Handbook that expressly states: “We expect on-duty drivers to accept all [ride] requests.” Handbook at 1. The Handbook goes on to state that “[w]e consider a dispatch that is not accepted to be a rejection,” and we “will follow-up with all drivers that are rejecting trips.” Id. The Handbook further notes that Uber considers “[r]ejecting too many trips” to be a performance issue that could lead to possible termination from the Uber platform. Id. at 8; see also Docket No. 223-57 (email from Uber to driver stating that the driver’s “dispatch acceptance rate [of 60%] is too low.…Please work towards a dispatch acceptance rate of 80%. If you are unable to significantly improve your dispatch acceptance rate, Uber may suspend your account”).
The time a driver has to accept a ride is typically very short; dispatch requests expire within 15 seconds.
One suspects, then, that few Uber drivers keep their Uber app on while vacuuming or doing other household chores.
The pressure to maintain a high acceptance rate is so great that it also leads drivers to undertake business that they would not pursue if they truly were independent contractors and demonstrates Uber’s control over a driver’s time. Consider this example provided by the Rideshare Guy blogger Harry Campbell: 3
As an Uber driver in LA, a minimum fare ride (any ride around 1.75 miles or less) will cost the passenger $4.65. After Uber takes its $1.65 booking fee and then 20% commission on the remaining $3, a driver is only left with $2.40. Drivers refer to these rides as income-killers since you could potentially wait 5-10 mins for a ride, drive 5-10 mins to get the passenger, wait 0-5 minutes for the passenger to come outside and then drive 5-10 minutes with the passenger (only part that is paid), all for just $2.40. As a true independent contractor, I would never accept this type of ride but drivers have to since they don’t know where the passenger is headed until they arrive and start the trip. But a driver has to accept such trips in order to maintain a sufficient acceptance rate.
Uber’s guaranteed pay program already establishes a minimum wage
Uber has provided its drivers with a minimum-guaranteed-wage program since at least mid-2014 (Huet 2014). In January 2015 Uber’s newsroom noted:
In the past, Uber has implemented price cuts without a guarantee for drivers. However, this time around, we’re trying a new approach. We’re so confident in the earnings gains drivers will see that we’re making earnings guarantees in every city where we’re cutting prices. We feel that it is important for drivers to have this kind of certainty and comfort going into a price cut. (Zara 2015, emphasis in original)
In January 2016 a similar announcement accompanied a new round of price cuts (Rachel 2016).
The guaranteed payments program has been covered by the press (Huet 2014) and discussed by leading Uber-driver blogsites (Campbell 2015; Perea 2015; Van Maldegiam 2015b).
Uber, in effect, has a minimum-wage program that differs by city and by the particular hour driven. Specifically, Uber guarantees a certain minimum hourly rate4 with requirements that typically include, besides the customer rating requirement: a certain acceptance rate (90 percent); an average of at least one trip/hour; and online presence for 50 minutes of every hour worked (Ferenstein 2015). Clearly, Uber knows (to the minute) when a driver is working for Uber and can guarantee a minimum wage.
A further indication of Uber’s ability to monitor drivers’ activities to the minute is its cancellation policy, which charges a fare to riders who cancel a ride more than five minutes after a driver has accepted it.
Krueger acknowledges that every minute the driver is engaged with Uber might be trackable by Uber, but he believes that enforcing the obligation to pay would be too difficult. “You know, take the minimum wage. If it could be applied, I think it would be just fine. But if somebody is working for two employers at once, it’s not obvious to me how you’re going to administer the minimum wage. If you could do it, I think it would be great” (Harris and Krueger 2015c). Yet experience demonstrates that in fact Uber can and does administer a minimum-wage program.
The legal issues of apportioning waiting time
Once a driver accepts a ride, she is engaged by Uber, working for Uber, and entitled to be paid by Uber. Uber contests whether this creates an employer-employee relationship, but the Uber contract with its drivers leaves no doubt that the driver’s pay for this time will come from Uber, which sets the fare, receives payment from the rider, takes its commission and fees, and passes the remainder to the driver (Technology Services Agreement 2015).
But what about the on-call waiting time—the period when the app is turned on but the driver has yet to receive a ride request? And what about drivers who have two apps turned on simultaneously? Are they working for both companies, only one or the other, or neither?
The compensability of waiting time has a rich history under the FLSA, going back to cases in the 1940s, such as Armour & Co. v. Wantock, et al., 323 U.S. 126 (1944). In Armour, the Supreme Court wrote, “Of course an employer, if he chooses, may hire a man to do nothing, or to do nothing but wait for something to happen. Refraining from other activity often is a factor of instant readiness to serve, and idleness plays a part in all employments in a stand-by capacity. Readiness to serve may be hired, quite as much as service itself.” The Court held that “time spent in playing cards and other amusements, or in idleness” was nevertheless working time and compensable under the FLSA.
The courts consider each fact situation on a case-by-case basis, always trying to discern the extent to which employees have given up their freedom and transferred control of their time to the employer. Key factors in making this determination with respect to on-call employees like Uber drivers include the amount of time an employee has to respond to a call; the frequency with which an employee is called to return to work; and the extent to which an employee may decline a particular request to return to work without facing adverse employment consequences. In the words of a Department of Labor opinion letter, “If the calls are so frequent or the on-call time conditions so restrictive that the employee cannot effectively use the on-call time for his or her own purposes, the on-call waiting time would constitute hours worked” (Wage and Hour Division 2008).
Here are two examples of federal appeals court cases where employee waiting time was found compensable (Kearns n.d.):
- City firefighters required to be able to report to the station house within 20 minutes of being paged in appropriate physical condition to work, who were called back to work an average of three to five times per 24 hour on-call period, could trade on-call shifts only with great difficulty, and were effectively precluded by their schedules from obtaining secondary employment (Renfro v. City of Emporia)
- Forestry service employees required to remain within 50 miles of the work site, who could not participate in social or other activities that would have prevented them from monitoring radio transmissions, had to respond to an emergency call within 30 minutes, and could not obtain relief from the on-call status because they were subject to call 24 hours per day (Cross v. Arkansas Forestry Comm’n)
And here are three cases to the contrary, holding that waiting time was not compensable for the following employees:
- City water and sewer department employees who could wear pagers, could not consume alcoholic beverages, and were called back to duty an average of less than once per day (Gilligan v. City of Emporia)
- City police detectives called less than twice per week, who could be reached by pager, and who had to remain sober and report to duty within 20 minutes of responding to a pager (Armitage v. City of Emporia)
- Pulp mill mechanics who only had to respond to one-third of the calls they received, who accepted an average of six calls per year, were issued pagers, and had to call their employer within 10 minutes of receiving a call (Owens v. Local No. 169)5
The U.S. Department of Labor and the courts will eventually sort out how much, if any, of Uber drivers’ waiting time is compensable, based on well-established legal principles. The facts demonstrate that drivers must respond to ride requests within 15 seconds in order to avoid the kind of acceptance rate that could cause them to be fired; they might get several ride requests every hour their app is turned on; and they might simultaneously have another company’s app turned on.
Krueger and Harris think that the time an Uber driver waits for a ride request with his or her apps turned on should not be compensable, that the driver is “waiting to be engaged”:
She can turn off the apps at will and go to her traditional job, undertake another moneymaking activity, drive her children to school, or park by the side of the road and take a nap. Even if she does not turn off either app, she is not obligated to pick up any particular customer. She can wait for a customer of her choosing, or until after she has completed her personal activities, whatever they might be. (Harris and Krueger 2015a)
But we think their description of the driver’s situation is misinformed and their argument flawed, and the driver is actually “engaged to wait,” and her time if compensable, for two reasons.
First, if she turns off her app, no one would say that her personal time is compensable. At that point, Uber gets no benefit from her time; it is all hers. But if she does not turn off the app, she cannot “go to her traditional job, undertake another moneymaking activity, drive her children to school, or park by the side of the road and take a nap.” Given that she has to monitor the app and respond to pings within 15 seconds, none of those activities is a realistic possibility.
Second, while she can refuse customers or wait until after she has completed her personal activities to accept a ride request, she cannot do so without adverse consequences, including possible termination. She has to respond to requests within 15 seconds or lower her acceptance rate, and a lower acceptance rate can foreclose Uber’s guaranteed pay rate and lead to deactivation.
Harris and Krueger raise the issue of which employer is responsible for paying for the time a driver is engaged to wait if the driver is monitoring two apps simultaneously. They suggest that the question is unanswerable:
We should not pretend that existing FLSA doctrine answers this question since there is no analogy in the employee–employer relationship to a driver with two simultaneously open apps for different services. This situation is not joint employment. If anything, Uber and Lyft are competitors for the driver’s services, not co-employers. The best legal answer seems to be that there is not a good answer.
We, however, believe there is a good answer. Both employers want the driver to be waiting for their ping, and the one that should pay is the one that will benefit—the company whose app provides the first accepted rider (Van Maldegiam 2015a).
Conclusion
Harris and Krueger have laid out a case for what makes gig work different enough to warrant different treatment under law, and their proposal for an independent worker category between employee and independent contractor has elevated the policy debate about the appropriate labor protections for gig workers. They forthrightly articulate the policy context and the issues of concern in this debate:
Technology is creating exciting new opportunities to link workers who provide services directly to customers, with potentially large gains in the quality, speed, and efficiency of service. From an economic and societal perspective, however, it is important that, if these new intermediaries are to succeed and expand, it is a result of their superior technology, efficiency, or service, not because their technology or business model enables regulatory arbitrage. For example, if an intermediary succeeds by displacing traditional employers who offer the same service because the intermediary gains a cost advantage by avoiding provision of certain legally mandated benefits and protections, then welfare is reduced by the innovation. (Harris and Krueger 2015a)
Nonetheless, our analysis of their assertion that the work hours of gig workers are immeasurable suggests that their case for a third work category is flawed.
Harris and Krueger’s primary example rests on a mistaken assessment of an Uber driver’s situation: Uber drivers cannot realistically keep their apps on without quickly responding to pings to pick up riders lest they earn a low acceptance rate and get deactivated. Thus, as a practical matter, these particular app workers do not enjoy a special, unprecedented multitasking capability that, in itself, meaningfully distinguishes them from other contingent or casual workers.
In addition, Uber drivers can clearly be subject to a minimum-wage requirement because Uber itself maintains one, offered to provide some income security to drivers as Uber keeps reducing fares.
We have also argued that the seemingly unique challenge of drivers availing themselves of two apps simultaneously (Van Maldegiam 2015a), or otherwise occupying themselves while waiting for work, can be addressed within the current legal scheme, and it is not central to determining whether there is an employment relationship.
Finally, we wish to note two troubling aspects of the developing discussion about labor policies toward digital platforms, and Uber in particular. First, this discussion does not now include the voices of current or former Uber drivers. Drivers have a wide variety of views, as anyone reading online comments can attest. Their experiences with the actual work relationships with their digital platforms are important to hear. Events like the forum on the Harris-Krueger paper had a large audience, including many people who had been Uber riders, but no one who had been a driver. Had there been drivers in attendance we suspect that we would have heard about acceptance rates and minimum-pay guarantees.
Second, it is apparent that Uber fails to provide a clear delineation of its policies to its drivers in a handbook (the Uber Movement site shows what is available). This lack of documentation may be driven by fear of litigation and liability. Regardless, Uber presents itself as providing a spot market for riders and drivers, yet a well-functioning market requires ample information. That does not seem to be the case right now.
—The authors appreciate the comments provided by Jules Bernstein, Harry Campbell, Kathy Krieger, Shannon Liss-Riordan, Alex Rosenblat, Kelly Ross, Catherine Ruckelshaus, Ben Sachs, Julia Simon-Mishel, Rebecca Smith, and the assistance of Jessica Scheider and Patrick Watson in the preparation of this paper.
Appendix
Harris and Krueger on ‘independent workers’ and work hours for app workers
Excerpts from their 2015 paper and a presentation at the Hamilton Project’s National Press Club event.
“Independent workers can work for multiple intermediaries simultaneously, or engage in personal tasks while working (“waiting to engage” vs. “engaged to wait”); their hours are easy to measure with technology but often conceptually impossible to assign to an employer.”(Harris and Krueger 2015b)
Krueger: “The independent worker can work for simultaneous intermediaries at the same time and can also conduct personal tasks while waiting to do work for an intermediary. In this sense, it’s conceptually very difficult and often impossible to assign the work hours of the independent worker to a particular employer. Technologically you can measure aspects of the work hours, but conceptually, the worker has the freedom to turn off the app and do whatever he or she wants, or to use two apps at the same time for different intermediaries, there is a fundamental difficulty when it comes to determining whether this individual is working for a particular employer, which makes it, we think, conceptually impossible to apply parts of the labor law to this type of the relationship if it were treated as an employment relationship.” (Harris and Krueger 2015c)
Krueger: “You know, take the minimum wage. If it could be applied, I think it would be just fine. But if somebody is working for two employers at once, it’s not obvious to me how you’re going to administer the minimum wage. If you could do it, I think it would be great. If you’re going to say, ‘Well we’re going to do it for the six minutes that someone has a passenger in the car,’ then it’s hollow, because it’s never going to be binding. It’s never actually going to help anyone.” (Harris and Krueger 2015c)
Krueger: “[I]t’s impossible to attribute their hours to a particular employer, if they can work for simultaneous employers at once or on personal tasks.” (Harris and Krueger 2015c)
Harris: “As I just described, one of the distinguishing characteristics—and Alan also talked about this—one of the distinguishing characteristics of the gig economy is that it is very difficult to distinguish work from nonwork and to assign responsibility for a worker’s time to a particular intermediary. Sometimes that can be impossible. Sometimes it’s just very conceptually difficult. As a result, we would not have the law impose a minimum wage or overtime requirement or require participation in the UI system.” (Harris and Krueger 2015c)
Harris: “And the reason why a third category is going to result in less misclassification, is because it is a more accurate description of the work circumstances.” (Harris and Krueger 2015c)
Harris: “Does the central problem exist? And there is a focus, and perhaps this is—I’ll take responsibility for this part of the paper that we used. We focused people on the countability of the time. Can you count it? And I thought we said in the paper, and I think we did, that actually, with technology, it’s easier to count time, much easier than even using a watch or a clock, to precision.” (Harris and Krueger 2015c)
That’s not the problem. The problem is the point that I emphasized in my presentation, and I think was emphasized in the paper. And that is, you cannot accurately say that the intermediaries control the time of the worker. So Craig, you talked about the app being the foreman. You can’t turn off a foreman, at least not without going to jail for assault. You can turn off the app. You can also have multiple apps, and you choose them. You choose when they’re on, when they’re not on. You can within some broad boundaries decide, in the case of Uber and Lyft, which ride you’re going to take, which you’re not… . And it’s not countability, it’s whether or not it’s fair to assign responsibility, it is able, it’s possible to assign responsibility for a time in the way that you clearly can in work relationships, which are never simultaneous.” (Harris and Krueger 2015c)
Harris and Krueger: “If a worker works for two intermediaries at the same time, as illustrated by the example of a driver who uses apps for Lyft and Uber simultaneously, it is unclear how the law would or should apportion total work hours between the two companies. Moreover, a worker could spend time at home with her app turned on, waiting for a possible work opportunity, while primarily performing work for another intermediary or engaging in nonwork activities.” (Harris and Krueger 2015c)
Harris and Krueger: “If work hours cannot be apportioned and measured for the purposes of assigning benefits or assessing hourly earnings, we think it makes little sense to require intermediaries to provide hours-based benefits, such as overtime and the minimum wage.” (Harris and Krueger 2015c)
