Uber is asking drivers to ante up for their future earnings.
The company is offering some drivers in Houston, Texas, a chance to “Celebrate Halloween early” by buying a week of “accelerated earnings” for an upfront payment of $115. The promotion, spotted by Data & Society researcher Alex Rosenblat, promises drivers 33% more on every trip between Oct. 23 and Oct. 29. “As long as your weekly earnings exceed $349, you’ll come out ahead,” it states. Drivers selected for the promo have until this Saturday, Oct. 21, to opt in.
Uber’s promotion.
Uber told Rosenblat (and later Quartz) that the promo is related to a study it’s conducting with two economists at MIT. A working paper on the study was first published in the National Bureau of Economic Research in September. It described a randomized experiment called the “earnings accelerator.” According to the paper, which was revised this month, the experiment targeted drivers in Boston in July 2016 who had averaged five to 25 hours of driving per week in the previous month. Roughly 45% of Boston Uber drivers were eligible and 1,600 were picked to participate.
Here’s how the experiment worked. First, Uber offered all 1,600 drivers the option to drive “fee free” for a week, meaning Uber waived its normal commission on rides.1 About 65%, or 1,031 drivers accepted the offer (the researchers believe the others simply may have missed it). Those drivers earned on average $100 more than the rest of the group during their “fee free” week and drove for an additional four hours.
Next, some of the 1,031 drivers who opted into the first offer were given a new choice: pay a certain amount upfront in exchange for another week of “accelerated” Uber-ing. The upfront payments ranged from $15 to $165 and were presented three different ways. Some drivers were told again that they could buy a week of fee-free driving. Others were given the chance to buy reduced Uber fees. Still others were offered a week of increased earnings. This time, the drivers who opted into those arrangements worked and earned on average 30% more.
The study is framed as a comparison between Uber and taxis. In a traditional taxi model, drivers keep 100% of their fares but pay a fee to work, typically in the form of a lease on a car or taxi medallion. Uber doesn’t charge most drivers anything to work (though some drivers do rent or lease their cars) but instead takes a cut of each trip. The experiment explores how much drivers need to earn to make a “lease” model worthwhile.
Of course, there’s an important caveat to all of this: Uber is always in control. Drivers pick when they work, but it’s Uber’s app that assigns them trips. They might get a long trip or a short one, a cheap fare or a more lucrative one. They might be booked for a private ride or hailed for a shared UberPool. Uber drivers’ earnings depend on minutes and miles but also on unpredictable financial incentives, like “boost” (guaranteed hourly wages) and “quest” (bonuses paid for completing enough rides in a certain period of time).
“Uber and its ride-hailing competitors routinely offer drivers temporary increases in pay known to drivers as promotions,” the NBER paper states. “Promotions are used to increase trip supply, equilibrating supply and demand without the need for surge pricing.” Driving for Uber can feel a lot like playing the slot machines.
Uber’s real-time data is catnip to economists. The company has 600,000 drivers in the US and about 2 million globally. It does millions of trips per day with prices that vary continuously based on location, traffic, and demand, among other factors. Influential MIT economist Josh Angrist is one of the researchers behind the earnings accelerator. Uber has previously recruited econ superstar and former Obama administration adviser Alan Krueger as a collaborator.
To the people who worked on it, the earnings accelerator is just another part of the ongoing experiment that is Uber, a chance to pull some levers and test an interesting economic hypothesis. To the part-time drivers who were randomly selected to participate, it’s their livelihood. Imagine if someone bought a week of accelerated earnings for $100 and then was in a car accident, or had to care for a sick child. They can’t get that down payment back. Previous research has found that “gig” economy platforms like Uber attract lower-income workers whose earnings are somewhat unstable.
“Under a taxi-style lease contract if you get sick or if something unexpected happens, you may lose money. That’s part of what we’re exploring,” Sydnee Caldwell, the other MIT researcher working on the study, wrote Quartz in an email. “The vast majority of drivers who accepted lease contracts indeed benefited in the sense that they made more money by opting in than they would have if they had not accepted the offer.”
Uber said its study was approved by MIT’s Institutional Review Board and that drivers selected to participate were sent an opt-in form with a disclaimer about academic research by either text, email, or a pop-up in the Uber app. The disclaimer shown to Boston drivers said the study could last through the end of 2016.
“This academic study is part of broader efforts to better understand the extent to which drivers benefit from Uber’s flexible work model in quantitative terms,” an Uber spokesman said in an emailed statement.
It’s hard not to feel like Uber’s definition of “flexible” increasingly means “able to be manipulated by short-term incentives.” That, after all, is what this study is really probing: what it takes to get drivers to rearrange their schedules and even wager present dollars on the possibility of future earnings. “People respond a lot to changes in their hourly earnings,” Caldwell told US News.
And what of the current “accelerated earnings” promotion in Houston? Academic research aside, it’s also a convenient way for Uber to lock in a supply of drivers ahead of the busy Halloween weekend.
Sign up for the Quartz Daily Brief, our free daily newsletter with the world’s most important and interesting news.
More stories from Quartz: