New Year’s Eve saw the now-annual ritual of complaining about Uber’s surge pricing—its practice of raising prices when demand for rides outstrips supply. One passenger, for example, paid $125.90 for a 14-mile ride rather than the normal fare of $21.34. Uber warns customers in advance when it implements surge pricing, and it makes sure they tap their consent to the price multiplier on the app when they call for a car. Nonetheless, everyone complains about being Ubercharged.
Sophisticates in the blogosphere say that the customers are whiners who don’t understand economics. Surge pricing is just a reflection of the laws of supply and demand, which we accept in most areas of economic life. At 1 a.m. on New Year’s Day, most Uber drivers would rather be in bed. Others may have part-time jobs like bartending that pay them more than the normal Uber fare does, particularly on a holiday. Increased demand for rides is fueled by drunken celebrants staggering out of bars. Taxis are in short supply for just this reason. If you want a ride but you’re not willing to compensate an Uber driver for the trouble, don’t complain.
Meanwhile, Uber has convinced many commentators that the taxi systems in cities are just cartels. They overcharge you for lousy service while not offering enough cars when demand peaks. Uber is winning over customers because it offers better service at a lower price—or, on New Year’s Eve, because it offers service, period. Its fans say Uber works so well because it operates outside of the outdated and inefficient taxi regulation system.
While Uber has done much good, its defenders are wrong to argue that Uber should not be regulated. So far, critics have focused on safety and data privacy, which are important but unrelated to the surge pricing controversy. There is an even more basic problem with Uber that so far its critics have Uberlooked.
To understand this problem, think about the market for car passenger service. This market is what economists call a “two-sided platform.” Uber’s technology plays a role just like that of other newfangled technologies, like video games, computer operating systems, and smartphones. On one side of the platform are the customers. Thousands or millions of people want to use an identical product—video games, a ride from here to there. On the other side of the platform are service providers—developers in the case of video games, drivers in the case of car service. The platform technology connects these groups of people.
What is special about markets with platforms is that it is relatively hard for people to switch from one platform to another. In the 1990s, the U.S. government sued Microsoft because consumers using Windows could not easily switch to the Apple operating system, allowing Microsoft to compel the fauna of the Windows ecosystem to use other Microsoft products linked to the operating system. While Apple served a niche market, Microsoft served the masses, and it was able to overcharge people once they were locked into its system, which most people chose because it allowed them to work with other Windows users.
Uber provides a similar sort of platform. While it is not impossible for consumers to switch among Uber, Lyft, and local taxis, nor for drivers to switch between the systems, it is a lot more convenient for everyone if everyone uses the same system. And the more people who use Uber alone, the further that Uber can spread its fixed costs—the cost of developing and maintaining software—among a large population, allowing it to charge customers less than its competitors can. Because Uber has a huge head start on Lyft and offers a service vastly superior to taxis, it has Uberwhelmed the competition, and made enormous progress in dominating markets. (For a contrary view, see this article by Aswath Damodaran, but his skepticism about investors’ bets on Uber may reflect his failure to take account of Uber’s potential to displace all competitors in a platform-dominated market, which is probably why investors see it as a vast cash cow.)
The underlying reason for the perhaps surprisingly platformlike nature of car services is that it is very difficult for passengers to negotiate with drivers. When you need a ride, you probably have no idea what a reasonable price for it is. This is in part because the cost for the driver is not always the same; it’s much lower, for instance, if he anticipates a return fare from someone located at your destination. Meanwhile, drivers who might pick you up don’t know how much you are willing to pay. Shopping around and haggling—the solution to this problem in other market settings—doesn’t work when people need a car here and now, where thousands of people are frantically trying to hail thousands of drivers. (That’s also why no one has proposed an eBay for taxi services in which drivers bid for your patronage.) The historical solution of this problem, as economists Edward Gallick and David Sisk explain, was price regulation by the government. Under this system, drivers charge the average cost of a trip (on a per-mile basis, once meters were invented). The price is always predictable (which pleases the customer) while the driver’s gain on low-cost trips offsets his loss on high-cost trips.
Nothing is new under the sun, and that includes the Uber controversy. Price regulation of taxis was thrown into disarray in the 1920s when mass production drove down the cost of cars. When that happened, anyone, including part-timers, could pick up a passenger for a high fare during times of peak demand, depriving the taxis of fares from low-cost trips that they needed to survive and reintroducing price confusion. Governments invented the medallion system to block the part-timers from unraveling the market by siphoning off the lowest-cost fares.
Uber is replaying this story. Its technology allows part-timers to re-enter the market and grab the low-cost passengers away from taxis. Eventually, Uber may displace taxis altogether. But Uber itself uses a version of average-cost pricing; all Uber drivers charge you the same amount, based on congestion as well as distance. So unless Uber can block competition the way taxis have, its own system will unravel, and we’ll be back where we started.
As Uber disrupts the market, it reintroduces the price confusion that led to taxi regulation in the first place, just as in the 1920s. The Uber passengers who complain about surge pricing are right to be annoyed. They have no idea whether they are paying a reasonable price or are being ripped off. Because taxi drivers cannot raise prices on New Year’s, most of them stay home. And Lyft is not much of a presence. That means that Uber is the only game in town, and as a monopolist, it can charge a price well above what is necessary to persuade drivers to climb out of bed and into the driver’s seat. The recurrent rejoinder—you don’t have to use Uber if you don’t want to pay the price—is a response that any monopolist could make, and has never been justification for charging people above cost. Of course, we don’t know that Uber overcharges us. But since Uber has made no secret that it is a profit-maximizing entity—indeed, it demonstrates a kind of infantile pride about its ruthlessness—we should, taking it at its word, assume that it charges the highest price it can get away with.
But whether or not Uber does overcharge people now, sooner or later—once it displaces taxis and dominates markets—it will. At the same time, new entrants will also try to unravel Uber just as Uber is unraveling the taxi system. Unfortunately, Uber can’t counter this problem just by charging the market price. At that point, price regulation and licensing will be necessary.
Uber and its defenders have made much of the virtues of the free market. And the cronyish relationship between taxi medallion owners and local governments, which has resulted in high prices and bad service, makes for an easy target. Indeed, as economists E. Glen Weyl and Alexander White point out, one can think of local governments themselves as platform-providers. They set the rules that the taxi companies must follow, just as Uber itself regulates the prices charged by its drivers and the terms on which they offer us service. It’s these two “platforms” that are in competition. Local governments want to kill off Uber because Uber threatens the political benefits of the taxi cartel.
Weyl and White also point out that one of the problems with platform-dominated markets is that new companies may be overly deterred from entering them, leading to excessive market fragmentation. The taxi market is plausibly a natural monopoly—a market in which a single firm (or an entity like a local government that coordinates drivers) can charge lower prices than a group of competitors while still charging people well above the cost of providing services. This means that it will be expensive for entrants that, like Uber, develop a new technology to dislodge incumbents, who enjoy familiarity among customers, and hence a cost advantage. A fragmented market, in which Uber competes with taxis and Lyft rather than rules the market alone, may be worse than a monopolized market where costs are much lower (and price confusion is minimized). For this reason, Weyl and White argue that, if anything, governments should subsidize companies that challenge natural monopolists.
But that is only when they enter the market. Once they become established, they resemble the rebel who overthrows the tyrant only to become a tyrant himself, a stock character in all those Greek and Elizabethan dramas that Uber seems intent on replaying. Weyl and White would probably tolerate Uber’s surge pricing today—even if Uber charges well above cost—because these profits help offset the cost of Uberthrowing the incumbent taxi services. But in the long run—and we may already be there—Uber will need to be regulated just like taxis.
There is a last twist. If local governments are Uber’s competitors, it might not be wise for the local governments to do the regulating. Weyl and White argue that the federal government will need to step in and set the rules for competition over the right to offer car-service platforms in different municipalities because, in their view, the city-taxi cartels effectively offer platforms in competition with Uber.
I’m not sure whether they are right. The federal government is not always good at regulating local activities; and a national organization like Uber can cozy up to the federal government, just as local taxi companies have cozied up with local governments. The key, it seems to me, is that when monopolies exist, the people who set the prices should not profit from them—that’s why we turn to government regulation despite its many imperfections. But whatever level of government ultimately acts, it will need to set licensing requirements and, yes, price caps. It will also need to lend a helping hand to the next round of entrants, the disruptors who want to disrupt Uber once it becomes a tired old monopoly that Ubercharges us like Microsoft.