Sharing is Not Caring: Reassessing Shared Mobility

Over the past several years, city planners, public officials, and car opponents have bought uncritically into a vision of shared mobility peddled by venture-backed Silicon Valley firms and free-market advocates. With transit agencies struggling to modernize, environmental stakes rising, and techno-optimism drowning out sober consideration of impacts from mobility services, we need to reassess our approach to new market entrants. Mounting evidence demonstrates that shared mobility is at best a mixed bag, and unconstrained may represent a net negative for many of America’s largest metros.

The False Promise of TNCs

There are some unambiguous benefits to the growth of TNCs (Uber/Lyft), starting with the much improved experience of ridehailing services over that of taxicabs. Racial discrimination by taxi drivers, a longstanding problem for Blacks and Latinos throughout the United States, appears to have decreased with ridehailing alternatives. A team of researchers at UCLA found wait times and driver-initiated cancellations for Uber and Lyft rides were significantly lower than those for taxis. And during the debate over New York’s moratorium on new TNC vehicle licenses, proponents and opponents alike praised TNCs for boosting mobility in underserved neighborhoods. Drunk driving has decreased significantly since Uber and Lyft emerged on the national scene, though the extent of credit we owe them is not entirely clear. Road fatalities have increased in markets where UberX was introduced, but this correlates with increased car sales and personal ownership.

The labor picture is also unclear. While TNCs have created hundreds of thousands of new jobs nationally, historically low unemployment and significant labor shortages across other industries should raise questions about the value of these jobs. Many Uber and Lyft drivers are unaware of full ownership and operating costs associated with driving, and net income is significantly lower than TNCs would have their drivers believe. Driver net income has halved since its 2014 peak, as large signup bonuses and frequently high surge pricing have all but disappeared. Presented with a more realistic assessment of earnings and costs versus other employment alternatives, it’s unclear how many full-time TNC drivers would continue. Despite perpetual promises and ambitious deployment timelines, it may be many years before we see labor cease to be a factor in urban mobility.

However, on the other side of the ledger, TNC growth is generating a significant increase in vehicle miles travelled (VMT) in dense urban areas. In San Francisco, a recent study attributed half of the observed congestion growth from 2010-2016 to Uber and Lyft. The shorter-term reasons are clear: lower-cost and more readily on-demand transportation has increased demand for direct end-to-end travel. Uber and Lyft claim they provide a significant share of rides that originate or terminate at transit stations, though since neither provides complete O-D (origin-destination) data, these claims are impossible to independently verify. Similarly, studies of transit impacts have been mixed and inconclusive due to poor data availability. Logic suggests TNCs provide little last-mile connectivity in the areas already best served by transit, and instead undercut it by offering increased convenience for choice riders.

Will Scooters Save Us?

It’s true that dockless mobility devices (e-scooters and e-bikes) are likely siphoning ridership from ridehailing services and some personal vehicle trips. However, even accepting these devices are pulling mode share from cars, the speed, range, and trip length data available suggest they are cannibalizing transit, pedal bicycles, and walking at much higher rates. They also appear to be generating many new recreational trips altogether. Surveys to date measuring travel behavior of e-scooter users and the devices’ impact on other modes have been dubious. Those run by Lime, The Portland Bureau of Transportation, and Populus featured poor sampling methods, but have shown a large majority of trips are either entirely new or coming at the expense of active modes.

The e-scooter firms are trying to increase longevity of their devices, from a reported average of 23 days in at least one market, to those that can withstand regular use for months. That said, nobody has performed a lifecycle emissions analysis on fleet devices in their current or ideal states. Calling these devices “zero emissions” is deceptive marketing that masks what are almost certainly awful GHG impacts from manufacturing lithium-ion batteries, transporting them, and regularly (re)distributing them to maximize utilization.

In the short term, there may be a real benefit to shared fleets as a gateway drug of sorts into micromobility. Rather than invest several hundred dollars purchasing a personal device, at $2-$3 a ride, access is cheap, convenient, and commitment-free. In the long term however, these services are not only unsustainable, but highly wasteful. The shared model incentivizes carelessness in both parking and riding (who cares if you damage a device you’re not liable for?), exposure to elements that reduce unit life, and general bad user behavior. Personal or pooled ownership may be a more socially desirable model for our cities.

The Case for Ownership

If certain modes offer real utility, we should want consistent travel patterns for strategic planning purposes. As much as “sharing” increases the attractiveness of a mode, the temporal and spatial challenges of ensuring a device or vehicle that can only be in a single place at a given time requires oversupply to ensure reliable availability on demand. Unconstrained by regulations, dockless e-scooter firms would flood sidewalks with devices, as real-time rebalancing wouldn’t be cost effective for such low-revenue trips. Most of the VMT generated by Uber and Lyft vehicles are “deadheading” miles, when drivers are carrying no passengers in between revenue trips. Anyone who says “Big Data” can solve these problems either doesn’t understand urban geometry or is selling product.

Instead, let’s focus on the real enemy, wasteful consumption-driven services. TNCs offload operations expenses on their drivers, and therefore neither pay for VMT nor parking. Uber only pays its drivers when vehicles are generating revenue, and therefore cares about operations expenses only to the extent it needs to retain drivers. If Uber’s “partners” spend too much on gas and vehicle maintenance, they may recognize the arrangement is unfair to them, and quit.

As such, TNCs neither pay for the emissions and pavement wear-and-tear they generate through dead-heading miles, nor the emissions cost of manufacturing more vehicles demanded by depreciation. These are effectively socialized losses, meaning the general public subsidizes TNC operations. Moreover, personal vehicles typically pay for at least destination parking, and in certain metros, origin (home) parking as well. TNCs pay for neither, which is why ride-based fees should take precedence over congestion charges (Jim McPherson lays out a thorough case here). That’s not to say increased parking fees and congestion charges aren’t also desirable, but costs with lower visibility to voters will always carry greater political viability.

We also need to recognize relatively few Americans live in the types of places amenable to the urbanist utopian vision. At least half the social cost of car parking is baked in for all but a handful of US metros; lot sizes and parking availability in the suburbs of Sun Belt and Midwest cities won’t shift with increased shared mobility. The cost of supplying reliable on-demand transportation is unreasonable in low-density areas during non-peak travel hours, and makes vehicle ownership necessary for a semblance of “anytime” convenience. As Blair Schlecter notes, too many trip types still require personal vehicles, and fleet economics won’t support a number of use cases for a very long time.

A more reasonable vision would focus on 1) steadily increasing residential and commercial density, 2) reallocating rights of way to high-capacity modes, and 3) improving transit operations to reduce single-occupancy trips during peak commuting hours. And of course investing in both pedestrian and bicycle infrastructures that incentivize active transportation. If the conditions warrant it (distance, temperature), electric micromobility would benefit as well.

Unlike cars, which require significant parking capacity, scooters (and bikes) are easy to store in an apartment or an office. Or in designated bike parking areas, the likes of which are finally emerging as common amenities in newer multifamily developments. Furniture zones and parking racks can accommodate incidental use, but micromobility commuters should be able to store their devices at home or in a sheltered area provided by their place of work. A static surplus of devices that last years, and not weeks or months, is far preferable to competing fleets, where operators consistently replace outdated-but-still-highly-functional devices with shinier and more attractive toys.

On a final note, we need to consider the impacts on transit, which require significant taxpayer support and long-term public investment in capital assets. Mandates from both the feds and their own charters require transit agencies to provide consistent service in all weather conditions and accommodate not just choice riders, but the poor, elderly, and disabled. Forcing the same coverage and design mandates on private mobility services would ruin their already questionable economics. Allowing private services to undercut transit systems today may permanently degrade them, and should private services falter financially, they will prioritize the interests of remote shareholders over the welfare of urban denizens.

Ultimately, our cities are not governed like a game of SimCity – legacy infrastructure and land use matter, as do our web of decentralized jurisdictions and their diverse demographics and political orientations. Shifts in land use patterns will take decades, and far-off visions of a shared mobility ecosystem for the masses are extremely far off. Eventually we can repurpose much of our urban and high-density suburban parking, but flooding our rights-of-way with wasteful ridehailing vehicles and dockless mobility devices will cause more harm than benefit. We should prioritize personal ownership, with its sunk infrastructure costs, while progressively densifying our communities, boosting high-capacity transit, and reallocating rights-of-way to transit and active modes.

When Silicon Valley firms market the “sharing economy,” they seek to evoke a communitarian ethos that breeds efficiency and sustainability. We need to start recognizing this false promise for what it is.