In 1956, US taxpayers gave General Motors, Ford, and the American car industry one of the world’s most expensive gifts: A $26 billion interstate highway system (eventually costing $129 billion) and a transportation policy designed around the car. It’s a gift that has kept on giving, says Peter Norton, an associate professor at the University of Virginia. “The effort served a utopian vision of cities in which anyone could drive anywhere, anytime, and park at the destination,” he writes. “It destroyed much of the urban America it was meant to serve.”
Today, the bulk of America’s transportation expenditures go to fueling, parking, and moving our cars around. That former sense of freedom has proved to be a prison for commuters trying to get to and from their jobs. Cities once had walkable layouts, mixed zoning, and efficient mass transit. Public transit, which now represents just 1% of passenger miles in the US, is starved for funds and inaccessible for many. Every working day, the average American driver spends (pdf) almost an hour going between work and home, up 20% since 1980, while spending $8,800 per vehicle per year, about one-fifth of their household budget.
Autonomous cars are America’s chance for a transportation “do-over.” We have two paths in front of us—utopian or dystopian. Which it will be is a matter of policy as much as technology or economics. We’ve inherited a lot of assumptions from the 20th century. Americans love to drive. Americans hate transit. People want to live in the suburbs. Americans chose the automobile long ago. But for many that’s a myth, argues Norton.
“It’s akin to locking a man in a 7-11 and studying his diet,” he says. Unless they’re offered viable alternatives, it’s not at all clear that Americans would opt for their car culture. Many, in fact, are doing the opposite by flocking to cities and walkable communities. “If we rebuild the landscape for autonomous vehicles we may make it unsuitable for anything else—including walking,” says Norton. “Before we make autonomous cars the solution, we must formulate the problem correctly. Until we do, we risk accelerating a journey to the wrong destination.”
🌇 Utopia around the bend
Autonomous vehicles (AVs) promise a transportation paradise (pdf), a revived Futurama in which we’ll hardly have to look up from our entertainment while we get where we’re going. Algorithms will deliver more people to their destination with fewer vehicles thanks to sharing. Our cars, hired out while not in use, will pay for themselves. Congestion will ease as more cars travel at higher speeds closer together than humans could ever manage. Square miles of parking lots and garages will be converted into public parks and spaces. In cities, private car ownership will fade as a few thousand robo-taxis replace private vehicles fleets. Recent modeling studies (pdf; 68) suggest Austin, Texas; New York, and Lisbon, Portugal could replace 90% of personal cars with modest fleets of robo-taxis. As these are electrified, the greenhouse gases they generate could fall by up to 82% by 2030, estimates Lawrence Berkeley National Laboratory.
All these benefits are theoretically possible. Only the data suggests that’s not where we are headed. In a business-as-usual scenario, automating vehicles are poised to double down on the ills of American mobility, at least in its cities. Here’s how that’s playing out.
Today, Americans are dependent on their cars as a matter of infrastructure and urban design. It’s the result of a century of hard-won campaigns by automakers to put cars at the center of American life. What worked for a period in the new suburbs and wide-open highways of the mid-20th century is turning into a gridlocked disaster in the sprawling cities where more than half the US population now lives. Even today, billionaires who can afford the most expensive vehicles on the market can’t escape horrific traffic jams in America’s cities. “Being stuck in traffic is soul-destroying,” lamented Los Angeles-based Elon Musk, the CEO of Tesla and co-founder of The Boring Company which is building tunneling machines as a way to escape such congestion underground (a dubious approach).
As long as the vision for AVs is to replicate the experience of personal car ownership, the future will only double down on the problems of today. “Self-driving cars will actually make it worse by making vehicle travel more affordable,” said Musk.
What might a soul-crushing AV future look like? To some degree, it’s already visible. Evidence is building that even partial automation from Tesla’s Autopilot is prompting owners to drive more miles, says Daniel Sperling, the founder of the Institute of Transportation Studies at the University of California, Davis. To test the effect of full automation on driving habits, researchers at the University of California, Davis, and Berkeley gave car owners personal chauffeurs to use as they pleased. On average, according to research presented last year at the Transportation Research Board, vehicle usage soared by 83%. Once millions of autonomous cars are redesigned expressly for this application, they expect miles to rise further.
Ride-hailing is unexpectedly worsening congestion as well. Between 2010 and 2016, San Francisco’s traffic delays rose by about 60%, estimates the city’s transportation agency, and half that increase can be laid at the feet of Lyft and Uber. While the ride-hailing firms (designated as Transportation Network Companies or TNCs by local governments) have promoted sharing rides in dense urban cores, the services are still encouraging people to use cars instead of public transit, walking or not going at all. When cars burn empty miles searching for passengers, known as deadheading, it makes it even worse.
To see a city planners’ worst nightmare, imagine a future where thousands of self-driving cars circle endlessly or search for their owners. Out-of-reach towns hundreds of miles from city centers suddenly turn into bedroom suburbs. Highways are full most of the night as commuters take advantage of sleeping cabins to make longer and longer commutes. People who can afford them own luxury self-driving vehicles, and spend more hours and miles in cars. Family hours spent in the car soar as AVs routinely rack up 60,000 miles per year.
If the price of driving falls as expected, the number of miles will keep rising. Just as people spend more time online as connection speeds increase, and more highways mean more traffic, AVs may lengthen commuting distances rather than relieve traffic. Even the environmental benefits could evaporate. Power demands from sensors and constant driving mean most AVs will likely be hybrid or even conventional gasoline vehicles at first (see Waymo’s hybrid Chrysler Pacific minivans). Autonomous cars could triple the 100 billion gallons Americans burn each year, erasing efficiency gains of the last fifty years, and more. The accounting firm KPMG (pdf) estimates AVs will add an additional trillion or more vehicles miles traveled annually by 2050—and four times more than that if autonomous vehicles are not shared and allowed to travel without passengers.
🛑 Emergency brake
How do we escape this trap? It only took 30 years or so for cars to replace horses in major cities during the early 20th century. And we’re still living with transportation decisions made then. Today’s decisions will likewise shape how AVs are deployed for decades to come.
The clock is ticking. By 2040, one of two worlds is likely to have emerged, says Sperling. “UberX or UberPool? That’s really the question,” he states. “To be the good scenario, it has to be UberPool without the driver. If cars personally-owned automated vehicles are superimposed on today’s patterns of usage, that’s the one that will lead to the hell scenario.”
To avoid the worst, here’s what planners say we need to do to rethink our cities’ transportation strategy from curb to highway.
Get ready, cities
“I can’t tell you if cities are prepared because we don’t know if we need to prepare for the utopian or the dystopian future,” says Warren Logan, Oakland’s mobility policy director and a former San Francisco transportation planner. But the uninvited arrival of Lyft and Uber taught cities they can’t wait for a new technology before they respond. Without a strong plan, cities will see their transportation goals go up in smoke. “[Uber and Lyft] had a honeymoon period. It was pretty short.” says Marcel Moran, a PhD candidate in urban planning at the University of California, Berkeley. Personal convenience for some has led to problems for all. Congestion went up. Transit ridership went down. “The context [for autonomous vehicles] is fairly oppositional,” he said. “It’s not a blank slate.”
AV pilots are underway in at least 36 states, and more are announced every month. Moran expected cut-and-paste policies before he started studying municipal plans for AVs. Yet he and his colleagues found a wide variety in cities’ approaches to the technology. “We find there is little consensus in terms of what cities should do regarding AVs, and it has been established that the vast majority of municipalities have not carried out planning for AVs,” he wrote in research to be published later this year. For those that are taking action, however, “cities understand AVs are a possible lever for other long-standing goals,” says Moran. “That’s the key.” A survey by Bloomberg Philanthropies found that cities want AVs to buttress their transit systems above all else, even more than taxis.
Getting companies to play ball may not be easy. Most have resisted divulging much data to cities citing privacy and “proprietary” concerns. While eager to present their bonafides to promote sharing and a (belated) willingness not to spar with cities over-regulation, they don’t want to reveal too much data to governments or competitors lest they erode their market position. If the data show, as New York’s research has, that TNCs are competing with transit more than personal cars, cities may be less welcoming.
For now, TNCs are pressing the charm offensive. “We want to do this with the community, not at the community,” said Dan Ammann, the CEO of Cruise which is launching a (now-delayed) robo-taxi service in San Francisco. “The traditional Silicon Valley approach of move fast and break things is not the right approach for bringing this technology to life…Even in the early stages, we want this to be something you can use as your primary mode of transportation.”
Yet even that message stands in contrast to what Cruise has been telling city planners, that it wants to buttress public transit, not replace it. “They’re trying to straddle this line,” says Moran. “[TNCs] are saying, ‘We’re providing these benefits such as last-mile transportation and electrification.’ Yeah, but you still want people in cars.”
For the last half-century, most roads have been free. Tolls were originally proposed but rejected for the interstate system. Especially in cities, that may be coming to an end. “In the context of AVs, you can’t have unpriced roads,” says Moran. The reason is known generally as the tragedy of the commons. When AVs slash the cost and effort to drive a mile, everyone will travel more. Unless a price is placed on what is effectively a very expensive public good, congestion, maintenance, pollution, and other consequences will increase. In 2016 alone, the US spent more than $530 per citizen maintaining its highways.
To adapt, cities and states are exploring two pricing options: congestion charges and a tax on miles.
Congestion pricing is well established. Singapore, Stockholm, London and other global metropolises all impose hefty fees on cars to access city centers during peak hours. Singapore’s goes back to 1975. But the idea is now spreading to America where New York will be the first American city to impose congestion pricing in 2021. Once AVs hit the road in real numbers, experts suggest most urban centers will have to follow suit to avoid choking traffic.
Taxing vehicles based on distance traveled is harder. Individuals’ mileage on public roads would be taxed. That would replace gasoline taxes expected to fall with the arrival of electric vehicles (a crucial source of infrastructure funding) and account for greater usage by AVs. Politically, and technically, this would be hard to implement. But Uber and Lyft have already signaled they’re open to the idea and legislatures are testing it out. (Washington state has a pilot underway.)
There’s an uphill battle ahead. “Americans are capitalist except when it comes to roads, where they’re staunch communists,” said Jeffrey Tumlin of the transportation planning firm NelsonNygaard Consulting to The Information (paywall). “All other forms of travel we use the market to balance supply and demand.”
“Let’s be clear,” writes Sperling about a future with AVs in his book Three Revolutions. “There is no alternative to pooling.” Cities have limited road capacity. Not everyone can drive personal vehicles at the same time, no matter how close together an algorithm spaces them. This simple fact vexes urban planners who see today’s ride-hailing services exacerbating miles traveled and congestion.
Uber recognized this years ago. When it launched its shared UberPool service in 2014, it called it “a bold social experiment” to reduce cars, congestion, pollution and parking challenges. But it hasn’t quite worked out that way. Most people still opt to travel alone over today’s carpooling options. Pooling rates peaked in the US in the midst of the 1970s energy crisis. It has fallen from 20% in 1970 to just 9% in 2016. Even in San Francisco, where half of Lyft and Uber passengers opt for shared rides, the reduction in miles traveled and congestion hasn’t materialized. The reason is the inefficiency of personal cars, which carry relatively few passengers compared to transit, and the empty miles as cars search for new passengers. Many people who would have taken transit are the ones using these services. Sperling cites a 2016 study in the Bay Area showing 75% of casual carpoolers were once public transit riders, while only 10% had driven alone.
“City planners need to explain how individual actors who feel like they’re making a positive choice [by taking shared rides] are in fact making a negative impact on the community,” says Logan, in comparison to avoiding personal vehicles. “They think, ‘I’m not driving, it must be good.’ That’s false. Someone is driving you.”
With Lyft Line and UberPool not yet easing congestion, players will have to do far more to encourage more efficient pooling by more people and increase transit use to realize benefits from AVs.
Citizens love to vote for public transit. Seventy percent of public transit ballot measures pass, a far higher rate than the average tax referenda. But ridership has not followed suit falling by 7% (pdf) over the past decade outside New York (the region is the lone exception with rising ridership). As AVs promise to cut the cost of riding in personal vehicles down to a few cents per mile, even as the cost of expanding today’s transit systems remains expensive, public transit could empty out.
But AVs could, theoretically, boost ailing transit systems by providing last-mile service, operating shuttles, and partnering with cities to provide cheaper alternatives on the systems’ margins. All of these things are being tested now. Lyft is partnering on last-mile services with transit agencies from North Carolina to California. May Mobility is operating six-passenger shuttles on popular routes across the Midwest. Five Central Florida cities have subsidized Uber rides to and from rail stations, as well as trips within city limits. The future of autonomous ride-hailing in dense cities is probably synonymous with great public transit and last-mile transportation.
🌉 Rethinking our cities
At their best, AVs will make it possible to travel further, faster, and safer at much lower costs. Highways will rumble with cheap freight on autonomous trucks. Distant cities can enjoy robust regional transportation options. Cities public transit systems could be revitalized, not supplemented, by autonomous vehicles. It’s a heady vision. Smart policy can go some way toward realizing it.
But at the end of the day, fixing our transportation system is really about more than cars. It’s about land use and where we connect our homes and workplaces. Silicon Valley is the cautionary tale when decisions made decades ago fail to keep up with a growing population. In the 1940s, most of Silicon Valley was blanketed in farms and fruit orchards. As the economy expanded from defense spending and the nascent tech industry, single-family homes and isolated office parks spread across former fields along the major highway. Cars were prioritized. Highways were built. Streetcar lines were ripped out. A single commuter line (CalTrain) snakes through a sprawling metropolis that’s home to more than 4 million people. In many neighborhoods, neither a sidewalk nor a bike lane can be found. San Francisco’s major highways now beat out Los Angeles for congestion. Commuters spend an average of 79 hours per year in traffic during peak hours, and 12% of that time idling in gridlock.
Fundamental changes are needed. From expanded, dignified transit to interlinked transportation connecting cities and affordable housing, AVs will only work as part of a larger system. We need to rethink the design of our environment, not just technology, to untangle cities’ gridlock. If Uber and Waymo transport more young urban professionals who would otherwise have walked or biked (as appears to be the case), our cities won’t necessarily be better off, and neither will many people using them (see the health effects of lower transit usage).
Perhaps the choice between a utopia and a dystopia will ultimately depend on who can make the decisions. That’s the lesson famed urbanist Jane Jacobs drew in The Death and Life of Great American Cities (1961). She lamented the Garden City movement, a well-intentioned effort by Ebenezer Howard in the 1890s that sought to create vibrant, affordable cities ringed by circles of civic buildings and public gardens. The vision, she argued, failed to account for the needs and desires of people who would actually be living there. “As in all Utopias,” she wrote, “the right to have plans of any significance belonged only to the planner in charge.”