Lower Fares Don't Move More Riders. A Pricier Shared Seat Does.

Based on the research ofMing Hu, Jianfu Wang, Hengda Wen, and Zhoupeng (Jack) Zhang, "Shared or Solo? Platform Pricing and Rider Choices in Ride-Hailing," Management Science, 2026

A platform trying to move the most riders should, in one specific place, charge more for the ride rather than less: pricing the shared seat higher is what turns solo riders into sharers, and sharing is what manufactures the driver capacity needed to move everyone else.

Ming Hu, Jianfu Wang, Hengda Wen, and Zhoupeng (Jack) Zhang model a ride-hailing platform as a queueing system offering two products, solo and shared, under three possible objectives: maximize revenue, maximize ride volume, or maximize social welfare. Riders are self-interested, deciding whether to request a ride at all and, if so, whether to share it. Standard pricing intuition says a platform chasing more volume should cut prices to pull in more demand. The paper's central result inverts that: a platform maximizing volume or welfare can rationally set higher prices on both the solo and the shared product than a revenue-maximizing platform would, because a shared ride is not just a discounted trip, it is two riders served on the driver-hours of one. Push more riders toward sharing and the platform is not simply reallocating existing demand across a fixed number of cars, it is manufacturing supply. The authors calibrate the model against Chicago ride-hailing trip data, Chicago is one of the few cities whose public data portal has published trip-level records since November 2018 that flag whether a ride was authorized to be shared, alongside pickup, drop-off, fare, and timing detail.

The Two-For-One Trick Hiding Inside a Shared Fare

The mechanism only works because a shared trip is a genuinely different unit of supply than a solo one. Uber's own pricing for UberX Share (the current name for the product that started as UberPool) makes the structure visible: riders get an upfront discount for choosing to share even before a match is found, then an additional discount, up to 20 percent off the total fare, once they are actually paired with a co-rider along the way, according to Uber. That two-part price is not a marketing flourish. It is the platform separately pricing the choice to share and the capacity created by an actual match, which is precisely the lever the paper's model turns. Before the pandemic, shared rides were not a niche product: Uber reported that more than 20 percent of its rides worldwide were shared trips, a scale at which the capacity effect stops being a rounding error and starts materially changing how many total rides the platform can clear through its existing driver pool. Grab's equivalent product, GrabShare, prices up to 30 percent below GrabCar Economy fares and pairs at most two bookings onto one trip with no more than two stops, operating under regional names such as JustSave in Malaysia and GrabCar Bareng in Indonesia. Different markets, same instrument: a visible price gap engineered to pull riders out of the solo queue and onto shared driver-trips.

What a Revenue-Maximizer Leaves on the Table

A pure revenue-maximizer prices each trip to capture what that rider will pay, without crediting itself for the extra capacity a sharing decision frees up elsewhere in the system, the rides that capacity enables somewhere else in the network don't register as this trip's revenue, so the incentive to push the sharing price up is weaker. The flip side of that coupling between price and capacity shows up starkly whenever capacity actually runs short. In 2021, as Uber and Lyft's driver pools failed to recover with ridership, fares did what queueing theory predicts happens when capacity shrinks and pricing exists only to ration it: they rose. CNBC, citing Rakuten Intelligence data, reported that the cost of a typical Uber or Lyft ride rose 92 percent between January 2018 and July 2021, and that by early July 2021 Uber and Lyft driver supply was running about 40 percent below the level needed to meet demand. That is the same queueing system the paper models, but observed from the demand-rationing side rather than the capacity-creating side: price rose because supply had shrunk, and volume did not rise alongside it, it fell. Raising the shared-ride price does something categorically different from that kind of scarcity pricing, it raises price and expands effective capacity in the same motion, which is exactly why a platform optimizing for volume, not just revenue, would want to do more of it, not less.

Lyft's Accidental Experiment in Removing the Lever

Lyft supplied an unplanned test of what happens when a platform removes the sharing lever altogether. In May 2023, under new CEO David Risher, Lyft discontinued its shared-ride product entirely, with Risher telling reporters that "the problem with shared trips is that they take people out of their way" and that the company needed to "pay attention to what your customers want," as Bloomberg reported. Cutting the shared option didn't just remove a cheap fare tier, it removed the only channel through which pricing could pull riders into carrying two fares on one driver-trip. Two years later, in May 2025, Lyft reversed course, reintroducing pooled rides at airports in five cities, Austin, Chicago, Denver, Las Vegas, and New Orleans, explicitly as a cheaper option amid rising fares, according to Fortune, following a survey earlier that year finding a majority of riders said they would cut back further if prices climbed beyond the 7.2 percent increase Lyft and Uber fares had already seen in 2024. Read against the paper's logic, Lyft's round trip looks less like a change of taste and more like rediscovering that the shared-ride price gap is a capacity instrument, not just a rider-facing discount, and that giving it up constrains how much total volume a platform can serve without leaning entirely on higher solo prices.

The Lever Depends on Trust, Not Just Price

The sharing lever is also fragile in a way pure pricing models don't capture: it depends on riders trusting the product enough to use it, and that trust can vanish overnight. DiDi suspended its Hitch carpooling service in China in August 2018, after a driver was accused of raping and killing a passenger, the second such killing of a female Hitch passenger in a matter of months; regulators said DiDi had "lost control" of its drivers and vehicles. A service that had logged more than a billion rides went dark and stayed dark until a trial relaunch in November 2019, in seven cities including Beijing and Harbin, with new real-time trip monitoring, an in-app Safety Assistant, and hours for women riders restricted to a cutoff of 8 p.m. rather than the 11 p.m. cutoff for men, as TechCrunch and CNN reported. No price adjustment could have substituted for the lost trust; the capacity-expanding channel simply stopped existing for over a year, regardless of what price DiDi might otherwise have set on it. Any platform treating the shared-ride price purely as a demand-management dial is missing that the dial only turns anything if riders still believe the product is safe to use.

A shared ride matched onto a driver-trip that would otherwise carry one rider is capacity created out of nothing, and creating it is worth paying for, on either side of the fare.

The Diagnostic Hiding in a Platform's Price Sheet

The direction a platform moves its shared-ride price relative to its solo price is a more honest tell of what it is actually optimizing than any mission statement. A platform quietly narrowing that gap to compete on headline affordability is optimizing something closer to revenue or short-run bookings. A platform willing to widen it, accepting a higher average fare in exchange for pulling more riders into shared trips, is using price to build capacity it does not have to hire, which is the behavior of an operator maximizing total rides served or, in the welfare-maximizing case modeled by Hu, Wang, Wen, and Zhang, total surplus across riders, drivers, and the platform together. Chicago's public, trip-level, share-flagged data is what makes that diagnostic checkable from outside the company: because the city records whether each ride was authorized to be shared, a regulator, a journalist, or a competing platform can, in principle, verify which logic a given operator's pricing actually follows, rather than relying on what the platform says about it.