By Paul Feinberg
Professor Keith Chen is a behavioral economist whose research blurs the traditional disciplinary boundaries in both subject and methodology, bringing unorthodox tools to bear on problems at the intersection of economics, psychology and biology. He recently returned to UCLA Anderson after a stint as head of economic research at Uber, the near-ubiquitous ride sharing platform.
Chen spoke last month at Anderson’s second annual Big Data Conference. His talk focused on Uber’s current challenge, what the professor called “running a ‘gig’ platform.” Much of Chen’s talk looked at Uber’s experimentation with the disruptive impact of driverless cars. He said that Uber was looking to “create order out of chaos.”
Regarding the value of self-driving cars, Chen noted that “everyone is richer when you don’t need to own a car,” inferring that Americans would have much more disposable income as driverless vehicles become more popular as, naturally, no one would need to actually own a car that drove itself, one would simply request a ride in one. He noted that 3.4 million Americans drive for a living, making up 3 percent of the work force. Of course, there is a concern that self-driving cars render human drivers obsolete. But Uber has done extensive experimentation with such vehicles in Pittsburgh, Pennsylvania, and the data is telling them that self-driving cars reduce the price of rides and increase demand for vehicles on the road. Until there’s a driverless car available for every rider, the demand for human drivers actually increases.
Data scientists at Uber, Chen said, focus on market and mechanism design. On the higher level, they are looking at pricing and dispatch matching to improve the user experience for riders. On a lower level, they study behavior of both drivers and riders. He noted that they spend a lot of time looking at data related to surge pricing and how riders react to different price levels. As an example, Chen said that the data indicated that more people were willing to hail rides at surge prices 2.1 times the normal rate than they were at 2.0 times the rate — a phenomenon that at first seems counterintuitive. Uber’s data scientists also began dividing city maps for drivers into hexagons instead of squares, which provided drivers with more and better information as to which neighborhoods offered better opportunities to pick up more and higher paying riders.
Chen also discussed how his team worked with drivers. Among their advancement, he said, was an instant pay platform that allows drivers to cash out their earnings on demand, instead of waiting for a week. Drivers who have a GoBank checking account are able to withdraw earned funds immediately, with Uber paying the fees. Drivers who use their personal debit accounts may also immediately withdraw funds, though they may incur banking fees. An analysis of the data Uber collected showed them that drivers worked 10 percent more hours after the institution of instant pay.