I’m perennially fascinated by the way in which different people read the same news story. After reading Eric Newcomer’s Bloomberg piece on Uber pricing, Marshall Steinbaum wrote a short tweetstorm about what it implies for Uber’s incentives, which I suspect is exactly wrong.

The gist of the article is not particularly new: it’s based on the fact that the amount a rider gets charged, for any particular ride, is increasingly getting divorced from the amount that the driver gets paid. Once upon a time, the driver would simply receive 80% of the fare, which was calculated on a miles-and-minutes basis. And indeed drivers still get paid on that basis. But riders are increasingly charged not according to miles-and-minutes, but rather according to opaque Uber algorithms. If Uber ends up charging them substantially more than it’s paying the driver for that fare, it can start making money rather than losing it.

What does this mean for Uber’s incentives? Steinbaum says that it “probably entails” Uber “withholding service as prices are bid up, especially during surges”. He also says that “empty cars are costless for Uber”.


I think that both of these assertions are wrong. For one thing, empty cars are not costless for Uber: the opportunity cost of keeping a car empty, when it might otherwise be generating revenue, is very real.

More interesting is the idea that the way that Uber can start charging more fo fares is by trying to maximize the number of surges. And I think that what Steinbaum is missing here is that if Uber wants more surges, it doesn’t need to keep cars artificially empty: it just needs to program in more surges. It can raise fares in any neighborhood, at any time, for any reason, or none.

Newcomer’s article, meanwhile, never once mentions surge pricing. Instead, his headline simply says that Uber is beginning to charge “what it thinks you’re willing to pay”. Which is something very different.


As I said back in October, the move to fixed fares was fantastic for Uber, because

it meant that the company could effectively hide surge multipliers, and indeed could start charging a premium not only when there was greater than usual demand, but also according to anything else that might prove profitable. Want to get picked up at a bar after midnight? It’s a pretty good bet that your normal price sensitivity might be impaired. Using a corporate card? Maybe you won’t mind so much if a few extra dollars are tacked on to the fare.

Surge, in other words, is an extremely blunt instrument. Uber now has much more sophisticated tools to maximize the revenue per ride: the example that Newcomer gives is that if you’re riding from an expensive neighborhood to another expensive neighborhood, Uber will charge you more than if you’re riding between poorer neighborhoods. (This is known as “route-based pricing”.)

But the logical end-point here is that Uber will charge not only according to routes, but also according to individual information that it has about each rider. The less price-sensitive you are, the more you’re likely to get charged. That’s a much more effective way for Uber to maximize revenues than any tactic which involves keeping drivers’ cars artificially empty.


Would that be an antitrust violation? I have no idea. I do think that it does bring Uber one step closer to being the drivers’ employer, since the drivers are effectively being paid a flat wage for generating a variable revenue stream. But I don’t think that Uber has any real incentive to keep cars empty. If it can charge all riders something close to the most they’re willing to pay, it will always want to do that in as many cars as possible.