Quantifying the Value of Flexibility for Uber’s Driver-Partners
By Judy Chevalier (William S. Beinecke Professor of Finance and Economics at Yale University) and Emily Oehlsen (Uber)
We’re excited to share findings from the release of our paper jointly authored with UCLA professors Keith Chen and Peter Rossi, The Value of Flexible Work: Evidence from Uber Drivers. In this paper, we aim to understand the extent to which drivers benefit from Uber’s flexibility−the ability to work if, when, where, and for however long they choose.
We’ve heard from drivers that this flexibility matters to them and can be difficult to find in other forms of work, and we wanted to understand the benefits of Uber’s flexibility in quantitative terms. We consider two dimensions of flexibility: the ability to set a personalized schedule (or no specific schedule at all), and the ability to adjust that schedule in real time. In other words, flexibility is not only the opportunity to work the hours you see fit — perhaps around another job, college courses, or responsibilities at home — but also the ability to change your schedule at a moment’s notice with no penalty — perhaps to study for an upcoming test or take care of a sick child.
It has been widely reported that the typical Uber driver-partner does not drive conventional 9 to 5 hours, five days a week. Consider the picture below, which shows every hour of the week. In red, you can see the fraction of employed males working any particular hour according to the American Time Use Survey. In grey, you can see the driving patterns of Uber driver-partners who are active in a particular week. The fraction of Uber driver-partners driving peaks on Saturdays and evenings — the exact times when people are less likely to be working at conventional jobs.
Our paper uses the economic concept of a “reservation wage,” defined as the earnings level below which a person will not work. If a driver has a good idea of how lucrative different hours are going to be in her city — since earnings may vary by time of day — we can get a good measure of her reservation wage by looking at the pattern of hours she chooses to work. For example, if a driver tends to work in hours when she can make $20 an hour but never works in hours when she can make $15 an hour, the driver has a reservation wage between $15 and $20.
We can also look at whether drivers have different reservations wages at different times. Consider a driver who often drives during the day, but has children to care for from 3 to 5 pm, and thus, almost never drives in those hours. Even if those hours offer relatively high earnings, we estimate that the driver has an even higher reservation wage between 3 and 5 pm, but a lower reservation wage at other times during the day.
Uber driver-partners can choose their hours, so they work only when the money they make from driving with Uber is larger than their reservation wages. This is not always possible in a traditional job: you are often committed to a predetermined schedule (e.g., 9 to 5 pm, Monday to Friday), and only in unusual circumstances can you miss an hour, day, or week of work because something else comes up (i.e., your reservation wage is high).
We use an econometric model to estimate reservation wages for U.S. drivers in the top 20 Uber cities between August 2015 and May 2016. We find that many Uber driver-partners do make an effort to drive more in lucrative hours. However, we also find that drivers have a lot of variation in their reservation wages over the course of the typical week. For a typical driver, there are relatively slow hours when she is still willing to work, and relatively lucrative hours when the same driver is not willing to work — indirectly revealing us to that the driver has a high reservation wage for that hour.
We calculate a measure that economists call surplus — that’s the difference between the wages the driver actually earns over a period of time and her reservation wages, the minimum it would have taken to induce the driver to work. After calculating this surplus, we then estimate how much it would change if drivers had to commit to their schedules at the start of the day or week and could not adjust, as is often the case in traditional jobs.
The median driver in our sample drove between 10 and 15 hours and earned about $400 per week. Of that $400, we estimate that 40%, or $160, is surplus — defined as any money earned above the absolute minimum that it would have taken to induce the driver to drive at the times she chose.
Constraints on Flexibility
We also find that this $160 surplus estimate falls dramatically (by two-thirds) if wages stay the same, but the driver is not able to adjust her schedule on an hourly basis. This is because we estimate that the typical driver has reservation wages that move around a lot — that is, there are some hours in which the driver’s outside activity is sufficiently important that she won’t drive, even if the payout is relatively high.
If drivers had to commit to pre-set schedules, we estimate they would end up driving hours they would not want to drive (earning less than the reservation wage they would require if they could choose to drive hour by hour). We also estimate that the typical driver would drive fewer hours than she drives now: she would not want to precommit to drive in an hour when her expected take-home was lower than her reservation wage for that hour.
We also examine alternative scenarios to determine what would happen if Uber were to put restrictions on driver schedules. For example, we find that most Uber driver-partners would simply not drive with Uber if they had to commit to driving full 8-hour shifts.
One of the attractions of Uber is the flexibility afforded to drivers. Our research documents an important source of value in flexible work arrangements — the ability to adapt work schedules to the demands of everyday life. Perhaps not surprisingly, this adaptability has high value to individuals who have chosen to drive and earn with the Uber platform.
The full paper, by Keith Chen (UCLA Anderson School of Management), Judy Chevalier (Yale School of Management), Peter Rossi (UCLA Anderson School of Management), and Emily Oehlsen (Uber Technologies), is available as an NBER Working Paper here
 For instance, in a 2014 survey drivers were asked to select major and minor reasons for partnering with Uber; combining major and minor reasons, 87 percent chose “to be my own boss and set my own schedule”, and 85 percent chose “to have more flexibility in my schedule and balance my work with my life and family” (for more details, see: Taking Another Look at the Labor Market for Uber’s Driver-Partners).