Robotaxis Aren’t Sure Route to Ride-Hailing Profits

A model weighs vehicle cost and ownership to assess benefits to riders, drivers and companies

After years of testing, delays and setbacks, self-driving vehicles are finally on their way to the big ride-hailing services. Lyft, in a partnership with Ford Motor Co. and an artificial intelligence startup, plans to start a robotaxi service in Miami by the end of 2021. Uber, in a separate joint venture, expects to have autonomous cabs on the road by 2024.

The companies have invested heavily in autonomous-vehicle technology for good reason. They believe widespread use of AVs would boost profits by dramatically reducing their biggest cost center: wages paid to human drivers.

That might not be the case. In a study forthcoming in the journal Manufacturing and Service Operations Management, UCLA Anderson’s Auyon Siddiq and UC Berkeley’s Terry A.  Taylor develop a game-theory model to discover how access to AVs will affect the ride-hailing companies, their drivers and their riders. Their findings suggest that AVs aren’t a sure road to profitability.

“Stakeholders in the ride-hailing industry should be mindful of the possibility that AVs can actually hurt profitability in some markets, as a result of intensified price competition,” Siddiq says in an email exchange.

In their model, the authors consider three sets of alternatives:

• Do the companies own or lease the vehicles, or do they deploy AVs owned by private individuals as needed? Owning or leasing AVs commits the companies to an upfront investment, while relying on individually owned vehicles allows them to bring cars into service only as market conditions require.
• Is it more or less expensive to access AVs? If the cost of bringing AVs into a company’s fleet is relatively low, it can put more on the streets. But so can its competitor.
• Is competition more intense for riders or for drivers? Or looked at another way, how much do companies have to cut fares to attract riders or raise wages to attract drivers?

Who benefits from the adoption of self-driving taxis, the authors found, depends on different combinations of these variables.

Are Autonomous Vehicles Profitable?

Ride-hailing companies can profit from owning AVs, the findings suggest, but only if the vehicles are relatively cheap and if the cost of paying drivers is relatively high. A fleet of AVs would be unprofitable if the vehicles are expensive and if competition for riders is intense, driving down fares.

The reason, the model suggests, is that owning the vehicles immediately requires a company in a competitive market to reduce fares in order to attract more riders, which in turn forces its rival to respond with price cuts. When competition for riders is more intense, the price war could cause fare revenues to fall faster than the savings from lower driver pay. If the cars are expensive, profits are squeezed.

Using individually owned vehicles, surprisingly, is profitable only if the AVs are expensive. That’s because, if AVs can be cheaply called into service as needed, a rival ride-hailing company can easily enter into a costly price war, hurting profits for both companies.

“Depending on the cost of AVs and market conditions, these lower prices can end up reducing profit of both platforms, because the intensified price competition outweighs the cost benefits of AVs,” Siddiq says in an email. “If AVs are cheap, then it’s possible both platforms come out ahead.” 

Riders generally benefit from having more AVs in service since rides become cheaper and more readily available. For drivers, the harm is obvious — after all, replacing humans is the whole point, and even a limited supply of robotaxis can drive down wages. Another question the authors consider is: Do the benefits to riders outweigh the losses suffered by drivers?

They do — but again, only if AVs are relatively cheap. Cheaper vehicles mean more get put into service, lowering fares. Since there are more riders than drivers, the combined welfare of the two groups comes out to a net positive. When AVs are expensive, driver wages still suffer, but less so, but not enough riders benefit from low fares to offset those losses.

Do Autonomous Vehicles  Increase Overall Social Welfare? 

In the model, ride-hailing companies sometimes profit and sometimes don’t, depending on how the vehicles are owned.  “Agent” welfare — the net effect of AVs on riders and drivers — always is positive when AVs are cheap, no matter who owns them. Is there any situation where the combination of all three — companies, riders and drivers — sees a net benefit from the adoption of AVs?

Here, access to individually owned AVs is the clear winner. When they’re cheap, the benefit to riders is greater than the loss in company profits, with a net gain. When they’re expensive, profits are high enough to compensate for reduced benefits to riders.

This is a situation envisioned by Tesla CEO Elon Musk, who has proposed a system in which Tesla owners make their idle vehicles available to a ride-sharing service in exchange for a portion of the fares. It, however, isn’t the approach being taken by companies, at least not initially.

Instead, Uber and Lyft will introduce their first robotaxis in partnerships with automakers and AV startups. While the ride-hailing companies won’t own the vehicles outright, Siddiq says that this approach is closer to its company-owned model, since the joint venture makes a decision about how many vehicles to deploy ahead of any demand signals.

If policymakers want to encourage adoption of AVs, the findings suggest, their best bet is to do whatever possible to reduce their costs. If that fails, they can try to make it easier for companies to use vehicles owned by individuals. And since drivers invariably face the loss of jobs and wages, their advocates can either try to obstruct adoption of AVs or push for laws designed to protect drivers’ interests.