With a presence in 65 countries, ride-sharing company Uber has conducted about 10 billion trips in its lifetime — about 15 million per day. Paul Solman looks at how economists are using this treasure trove of data.
Read the Full Transcript
Amna Nawaz: Uber’s impending initial public offering is expected to be one of the largest ever. Our economics correspondent, Paul Solman, recently visited their headquarters to better understand what happens when a San Francisco company puts economists in the driver’s seat. It’s the latest installment of Making Sense, our weekly economics series. Cory Kendrick: We’re in 65 countries, and we see about 15 million trips a day. Paul Solman: Uber, these days literally all over the map. Cory Kendrick: So, you will see the biggest circles are where we see the most trips happening on our platform. Paul Solman: Uber data scientist Cory Kendrick can track every ride worldwide, and there have been tons of them. Cory Kendrick: About 10 billion trips since we started. Paul Solman: Here at Uber’s San Francisco headquarters, the resulting data is analyzed by a team of in-house economists. Jonathan Hall: My guess is, overall, between 20 and 30 Ph.Ds. Paul Solman: Jonathan Hall is Uber’s chief economist. His mission? Jonathan Hall: Writing papers, publishing them, and trying to establish ground truth on tricky issues like the value of flexibility to the work force. Paul Solman: Of course, Uber’s economists need to protect its bottom line, help it stay ahead of the competition. But Uber presents an unusual opportunity, the realization of many an economist’s dream, efficient markets. Case in point, surge pricing. Jonathan Hall: Economists really don’t like the idea of the price not being able to change when supply or demand changes. We have this algorithm called the surge algorithm whose purpose is to identify imbalances between supply and demand and solve them in a way that maximizes the efficiency of the market. Paul Solman: So when more passengers demand rides, prices go up to entice more drivers to supply them. Or take Uber’s controversial, oft-criticized labor model. Jonathan Hall: The drivers, who are independent contractors, can choose to work whenever they want or wherever they want, with essentially no restrictions. And that’s very exciting to an economist who believes in open markets and competition. Paul Solman: So you are the embodiment of what economists would like an economy to look like? (LAUGHTER) Jonathan Hall: I think we strive for that. Paul Solman: So Hall’s economics squad is mining the vast data collected by the Uber app to understand driver and passenger behavior, and to tweak it. Jonathan Hall: Women, on average, tend to do fewer trips per unit of time than men. Paul Solman: One study analyzed data from over a million drivers to find a 7 percent gender earnings gap. Paul Oyer: We know it’s not discrimination. There’s a formula that determines how much they make. Paul Solman: Stanford economist Paul Oyer, who did the research with Uber, says women make less for a less obvious reason. Paul Oyer: They drive slower. So, the faster you drive, the more revenue you’re generating, because you’re getting more rides in per hour. Paul Solman: And since men give more rides, they get more experience. Paul Oyer: So if you go out and get in an Uber right now, and there’s a man driving, and then you go out and get in an Uber with a woman driving, on average, that man will have had more experience driving for Uber, and therefore he will be better at it and he will earn more money. Paul Solman: While we were there, the Uber team explored one possible tweak. Alison Stein: Having this per-mile component essentially mechanically rewards for speed. So, what this intervention considers is weighting less heavily on this per-mile and more heavily on this per-minute. Paul Solman: Data is the driver, about tipping, for example, introduced in 2017. Jonathan Hall: We consider the rollout of tipping to be an important experiment and opportunity that — to get it right. And so you will see changes happening over time. Paul Solman: Because you’re constantly testing, does this make people more generous towards drivers or less inclined to use Uber, or both? Jonathan Hall: Both. Paul Solman: But using the key metric of economics, the cost of losing customers, was outweighed by the benefit of more motivated drivers. Stanford economist Susan Athey many tech companies now employ economists to run the numbers. Susan Athey: You have a bit of this image of Silicon Valley that somebody is sitting there, you know, thinking of a brilliant idea, you know, alone in a room, and then putting it into practice. But, actually, the real innovation that happens, especially for the larger tech firms, is just lots and lots and lots of incremental innovation, randomized controlled tests, thousands of them a year, to try to improve algorithms, to try to improve systems to make them better. Paul Solman: Better, which ultimately means more profitable, whether consumers are on board or not. Jonathan Hall and I took an Uber with Hayley McGonigle, who says the algorithms keep her busy, happily busy. Hayley McGonigle: I think a lot of the improvements lately have been to keep drivers constantly busy. Any time you don’t have a rider, that’s, of course, lost wages right there. Paul Solman: McGonigle was provided by Uber. She works morning rush hours, says she makes $35 an hour. But for some San Francisco drivers, the economist’s dream can be the worker’s nightmare. Mostafa Maklad: They have been decreasing how much money they pay to drivers year after year. Paul Solman: Mostafa Maklad has been driving for almost four years. Mostafa Maklad: Instead of working like 40 hours a week, I’m forced to work at least 60, 70, 80 hours a week in order to pay my bills and pay all my expenses. Paul Solman: Bay Area drivers for Uber and rival Lyft have been protesting low pay with no benefits. Meanwhile, the IPOs of both companies will make many of their actual employees rich. Michael Martinez: There’s a lot of frustration out there. Paul Solman: Michael Martinez drives for Uber about 20 to 25 hours a week. Michael Martinez: If you include all of the costs, and you also include the constant rate cuts that have been going on over the last four years I have been doing it, I’m probably getting pretty close to just what San Francisco’s minimum wage would be, to be honest. Paul Solman: Uber says there have been rate changes in both directions. But that doesn’t mitigate the impotence drivers feel in the hands of a faceless, forever-tweaked algorithm. Mostafa Maklad: Whenever they do any changes, we are forced to accept and agree to their terms in order to go online and start driving. Michael Martinez: We have absolutely no power whatsoever. Uber does as they will, and we have no choice but to accept it. Paul Solman: But, despite their gripes, both Maklad and Martinez continue to drive. Michael Martinez: The main draw is the flexibility. Mostafa Maklad: To be honest, it’s the only job that I can do while going to school. Paul Solman: Unsurprisingly, Uber economists like Libby Mishkin studied the value of flexibility. Elizabeth Mishkin: Drivers worked more as a result of this flexibility, because they don’t have to commit to doing something that they might not want to do later. They can just choose to work whenever they feel like it. Paul Solman: The economists also use data to answer critics on a myriad of issues, like the charge that uber increases congestion in cities like New York, Uber rides lighting up the city on this map. Cory Kendrick: There’s really mixed evidence. And so one thing we’re trying to do is make sure there’s more rigorous research out there, trying to get at this question. It’s really hard to study, for a number of reasons. One is that there has been a lot of economic growth in the last 10 years, and there’s more traffic when that happens. Paul Solman: Suppose you did a study and found that Uber was the main contributor to congestion problems in New York, San Francisco, and elsewhere. Wouldn’t you be constrained from telling everybody that that’s the case? Cory Kendrick: Some of these questions may be tough questions for us, but we would rather know the answer than not know. Congestion is bad for Uber too. Our cars will be going slower. We also have bikes and scooters, which are new modes on the platform that are trying to get people out of cars and into active modes. Paul Solman: In short, more tweaking, with the data from mega-millions of customers leading the way. And that’s the new world of economics, says Susan Athey. Susan Athey: Right now, it’s the Googles and the Amazons and the Ubers, but it’s going to be the banks and it’s going to be other manufacturing firms. So, really, all parts of the economy are going to digitize and start optimizing in a more scientific way. Eventually, it will all get routine. Paul Solman: Because of the information you and I provide. From San Francisco, this is “PBS NewsHour” business and economics correspondent Paul Solman.
