In September, California legislators passed a bill that requires Lyft, Uber, and other similar service platforms to treat the workers that provide services through them as employees rather than independent contractors. Other states are likely to follow suit.
However, ruling on whether gig workers are employees or contractors just highlights the inadequacy of the traditional binary classification of workers in today’s economy. The truth is that they may be neither.
It is important to recognize that marketplaces for products or services choose a position on a continuum defined by how much control they exert over the interactions or transactions they enable. At one end of the continuum are pure marketplaces, which exert little or no control over the terms of the transactions between independent suppliers or professionals and customers (e.g., Airbnb, Craigslist, eBay, Poshmark, TaskRabbit, Thumbtack, Turo, Upwork).
At the other end of the spectrum are resellers that buy products from suppliers and resell or rent them on terms they completely control to customers (e.g., iTunes, Netflix, Wayfair, Zappos, Zipcar) and employers that hire professionals providing relevant services and almost entirely control how those services are delivered to customers (e.g., Hello Alfred, Infosys, McKinsey). Many firms have chosen to occupy intermediate positions along this continuum (e.g., Apple’s App Store, Gerson Lehrman Group, Handy, Lyft, Postmates, Uber, Wag).
The notion of control over supplier-customer interactions has many dimensions: price, equipment, how the relevant product or service is presented or advertised, how the product or service is delivered or performed, work schedule, and so on. What’s more, the stringency of the rules governing each of these dimensions can vary all over the map. As a result, there is a fine-grained spectrum of intermediate business models between pure marketplace and pure retailer or employer. For example, Postmates has full control over the delivery price charged to customers, but its couriers can choose to use any vehicle they wish for performing their deliveries (including bicycles, cars, and trucks) as well as their work schedules.
Marketplaces are driven to adjust the control dials for the various elements of their business model by a number of different factors: the importance of consistency of the buyer experience (across different sellers), the extent to which the platform vs. the suppliers have better information relevant for different choices (e.g., demand information for setting prices or marketing activities), whether the platform has some cost advantage of centralizing certain activities, and so on. In particular, there may be very good efficiency reasons for choosing different levels of control across various elements: for details, see our previous work.
In this context, trying to force-fit all marketplace companies into one of only two categories will lead to over-inclusive employee classifications in some contexts and under-inclusive employee classifications in others. As firms inevitably adjust in order to avoid being misclassified, this will lead to several inefficiencies, as we discuss in depth in a recent research paper on this issue.
First, firms will most likely “run for the corners” by choosing either a pure marketplace model with as little control as possible over how workers conduct transactions with customers or a pure employment model with full control over workers. This will likely eliminate many intermediate business models (with intermediate degrees of control), which could be more efficient.
The California bill, for instance, heavily favors the employment model—i.e., even companies that are quite close to the pure marketplace model may have their workers classified as employees. For certain services, it may indeed make sense to give workers more benefits and training, and employ them, so they can be dedicated and committed to the service they provide on behalf of the company.
This is the approach chosen by Hello Alfred, whose home managers provide a wide range of in-home services. However, forcing companies to do that for all services defeats many of the advantages and efficiencies of the sharing economy model, which in cases like ride sharing and delivery, provides a way for many individuals to pick up extra work, if and when it fits their schedules.
Second, when businesses that would have preferred to choose an intermediate model are categorized as employers (which comes with 20 percent to 30 percent higher costs), they are likely to take more control over how workers interact with customers (following the California bill, Lyft has already informed its drivers that it may soon restrict them to specific time shifts and geographic areas).
This results in less flexibility for workers and will almost certainly penalize workers that wanted the flexibility of running their own schedule and moving freely between different jobs. This explains why some Uber and Lyft drivers (nearly half of whom drive less than 10 hours a week) are unhappy with the new California law.
Third, when businesses that would have preferred to choose an intermediate model are driven to the pure marketplace model by the threat of higher costs associated with the employment model, they might cease investing in activities that would have made workers or customers (or both) better off. Two examples are worker training (e.g., Postmates offers its drivers free access to online college course and professional certifications) and group health care plans (such as the one offered by Uber): Because providing training and health care coverage are among the criteria that make an intermediary more likely to be classified an employer, companies that wish to remain marketplaces will become more reluctant to provide such benefits, to the detriment of workers.
Fourth, specifically in the case of transportation platforms like Lyft and Uber, it is important to recognize that this is a very competitive industry, mainly due to the fact that drivers have the ability to “multi-home,” i.e. drive for two or more platforms, as many do. However, if Lyft and Uber drivers are classified as employees, then each platform will be able to legally restrict its drivers from driving for a competing platform. Perversely, this would limit drivers’ choices (again, following the California bill, Lyft has informed its drivers that it may soon require them to drive for Lyft only). And ultimately, as the platform that manages to lock up more drivers becomes more attractive for riders through lower wait times, and more riders attract more exclusive drivers, it could lead to tipping whereby one platform dominates (in any given city).
In an ideal world, firms would be able to choose among all possible intermediate steps between pure marketplace and pure employer—subject to the constraint that their costs (including the benefits paid to workers) will increase relative to some aggregate measure of the control exerted.
Needless to say, this is not practically feasible: There are just too many possible intermediate configurations, and it would be prohibitively complex to assign a different legal status to each of them. But it is not too much to ask for the introduction of at least one intermediate step such as “dependent contractors,” for whom firms would cover some costs but not others (such a third category already exists in Canada, Germany, Italy, South Korea, and the U.K.)
Critically, any worker expenses or benefits that the new category requires firms to cover should be proportional to the work that these workers actually do with each of the firms (e.g., the number of hours worked or the value of the jobs carried out). Any expenses or benefits that are fixed irrespective of the actual work carried out would make engaging workers for short shifts unattractive and incentivize the firms to engage their workers exclusively to prevent rival firms from free-riding on their investments.
A case in point is the requirement that gig workers be paid a minimum wage per hour they are available on the marketplace rather than based on the actual work carried out. A recent case in Australia illustrates the point, in which a Uber Eats driver claimed she was not receiving the minimum wage because she worked as long as 96 hours in some weeks—most of it spent waiting for orders to be placed via the Uber Eats app—but earned as little as $300 for those periods. However, the Fair Work Commission in Australia rejected her claim, pointing out that she had rejected more than 550 food delivery requests and cancelled a further 240 after having accepted them. Requiring a minimum wage that is based on the hours a worker is available on the platform would almost certainly push each firm to require that workers meet a minimum number of hours on its platform and that they do not work for rivals during these “shifts”.
Yes, there would still be the problem of drawing the boundaries between the intermediate status and the other two. Nevertheless, this would be a big step forward in terms of freeing firms to explore a variety of intermediate business models and arrangements with their workers.
Authors Andrei Hagiu is an associate professor of information systems at Boston University’s Questrom School of Business. Julian Wright is a professor of economics at National University of Singapore.
