Cars take up way too much space in cities. New technology could change that.

By: Brad Plumer

Javier Zarracina/Vox

When we talk about the problems associated with cars and transportation, we often focus on fatal accidents, or air pollution, or traffic jams.

We less frequently consider how much sheer space cars take up in America’s cities. But let’s pause to give this some thought.

There’s the space the cars themselves occupy. The average car, two hulking tons of steel, is 80 percent empty when it’s being driven by a single person. And most of the day, cars are totally empty, sitting unused. That, of course, requires space for parking: There are a billion parking spots across the United States, four for every car in existence. Plus, there are all the paved roads crisscrossing our cities. Add it up, and many downtowns devote 50 to 60 percent of their scarce real estate to vehicles:

Satellite view of downtown Buffalo, New York.

It all seems rather inefficient and wasteful. If cities could reclaim even a fraction of this land from vehicles, they could build more housing, or stores, or parks, or plazas. For cities struggling with housing shortages and soaring rents, such as San Francisco and New York City, the gains would be staggering.

Some cities are already tinkering around the margins here — looking, for instance, to cut down on excessive parking requirements or boost mass transit and free up land for development. But new technology could push this even further. Ride-sharing services like Uber and Lyft already hint at a world in which cars are utilized more efficiently and take up less room in aggregate. And if self-driving vehicles become widespread, cities could in theory shrink their transportation footprint even more dramatically.

This world no longer seems so far away. A recent report from the Rocky Mountain Institute argued that the era of private car ownership may peak within a decade, as new networks of shared, electric, possibly autonomous vehicles become cheaper. Instead of buying a car, you can simply buy a ride whenever you need one. That shift has the potential, at least, to revolutionize our streets.

The trick is figuring out how to redesign cities accordingly. Recently, San Francisco sketched out a forward-looking plan to take advantage of these new transportation options and shrink the amount of space devoted to cars. With smaller streets and fewer parking spots, the city would have more land to work with — to build more affordable housing, say. If it works, it could be the start of an important new trend.

San Francisco has an audacious plan to reclaim land from cars

An illustration from San Francisco’s Smart City proposal. On the left are streets currently occupied by cars and street parking. By shifting to shared, autonomous vehicles, much more area could be reclaimed for parks and housing. (SFMTA)

Earlier this year, the Department of Transportation held a contest asking dozens of local governments to submit visions for a "city of the future" that incorporate things like self-driving cars to tackle problems like congestion and climate change.

Columbus, Ohio, ended up winning the contest with a detailed plan to improve mobility in low-income areas. But let’s take a closer look at San Francisco’s submission, because it’s a great exploration of how cities might use new tech and business models to take back scarce land from cars.

The proposal starts by observing that San Francisco currently has 440,000 on-street parking spaces — the same amount of land as the Golden Gate Park and 120 Transamerica buildings. And much of that land sits empty much of the time. "Our plan," the proposal notes, "would phase in innovative technologies that allow us to repurpose public space currently under-utilized as parking into affordable housing, small parks and pedestrian amenities."

The first step would be to make ride-sharing services (including, but not limited to, Uber and Lyft) more convenient and accessible to residents. After all, if people can use these cars for their transportation needs, or combine them with mass transit options, they wouldn’t need nearly as much street parking. And because each of these cars serves multiple passengers, they take up less space on the road.

In phase one, San Francisco planned to shift 10 percent of single-occupancy vehicle trips to transit and ride hailing. To do so, the city proposed partnering with the University of California Berkeley and various tech companies to work out ways to:

1) Provide incentives to shift people from their own cars into car sharing: That might mean designating certain road lanes as only available for ride sharing, making them the faster option. It might also entail seamlessly integrating car sharing, bike sharing, and public transit by creating a single simple mobile app that combines routing, scheduling, and payment for all of those services.

2) Make these services more affordable: That might involve providing low-income residents with access to smartphones and banking services, as well as providing free public wifi so that all could use these services. It would also mean finding ways to lower the price of car sharing — say, by deploying larger six-person passenger vans to cut costs below what an Uber or Lyft ride currently costs.

3) Eventually move to automated electric vehicles: If self-driving cars and buses eventually become a reality, they too could be connected into a centralized network, making sharing even easier. In theory, these vehicles could also reduce fatal collisions (assuming that self-driving technology proves safer) and would also eliminate air pollution (assuming that the cars were all electrified rather than running on gasoline).

The proposal offered an illustration of how these different phases would unfold. Note that over time, the amount of land required for parking shrinks as people move from a vehicle ownership model to a transportation service model that encompassed everything from cars to Muni buses to delivery vans:

Timothy Papandreou, the former head of the San Francisco Municipal Transportation Agency’s Office of Innovation, described the goal to the Washington Post in June: "We can move the same amount of people with a tenth of the vehicles. ... It’s really going to open up our minds. We’re not going to need to have all that excess road space."

Now, San Francisco didn’t win the federal contest or the $50 million grant that came with it. But SFMTA spokesperson Paul Rose told me they’re still looking to move forward with many of the plan’s components. The city is applying for another federal grant to launch pilot programs around "connected carpool lanes, smart traffic signals, autonomous shuttles, dynamic carpool pick-up curbs, connected Vision Zero corridor, and Congestion Toll System," Rose said. "We expect a decision this month on this application."

We shouldn’t underrate the challenges here. Reducing the number of vehicles on the road will require persuading people to give up their privately owned cars and shift to a pure sharing model. That’s not easy. It involves changing some deep-seated behaviors, and policymakers and companies will have to get the incentives just right. If ride sharing remains unaffordable, or if people simply don’t want to give up their cars, the plan could easily stumble.

Still, San Francisco has all the reason in the world to try. The city has sharp geographic constraints, and skyrocketing housing prices are making the area unaffordable for many. Reformers often focus on changing the Bay Area’s zoning laws to build more housing on existing land, and that’s no doubt part of a solution. But reclaiming vehicle space for housing could prove an equally appealing concept.

Self-driving cars could free up an enormous amount of room

San Francisco’s transit officials aren’t the only ones thinking about how new tech might decrease the amount of space that cars take up. A fascinating recent study by two British engineering firms, Farrells and WSP|Parsons Brinckerhoff, looked at how London’s streets might be entirely redesigned if self-driving cars ever became a reality.

That study imagined a world in which the autonomous vehicles (AVs) of the future are shared rather than owned — you call for an AV, and it zips right to your door. The AVs themselves are either always on the road, picking up and dropping off passengers, or charging/refueling/parking in a few centralized locations. As such, there’s simply less need for street parking.

What’s more, if all the cars on the road were autonomous, they could take up far less space on the road. Vehicles navigated by robots could nestle closer together without fear of rear-ending each other. If collisions became more rare, the cars themselves could be smaller and thinner, taking up less space. City planners could reduce the width of streets or even cut back on the number of lanes without greatly affecting travel times.

If you pushed this far enough, the study notes, a city like London could gain another 15 to 20 percent of developable area. "This is primarily due to the removal of almost all parking spaces, but also because of roadspace simplification that will save space."

The authors of the paper sketch out a few visions of what this might look like in London. Like so:

Or like so:

"Of the estimated 8,000 hectares of central London land occupied by parked cars today, it is reasonable to assume that 50-70% — potentially more than 5,000 hectares — could be released once AVs are commonly in use," the study notes. Replacing those spots with housing or other structures would be worth tens of billions of dollars.

Of course, this is just a vision of what could be — someday. There are tons of hurdles in getting there. For starters, despite all the hype, there are no autonomous vehicles yet available that can handle the range of surprises that might pop up in an urban environment on a daily basis. As I’ve written before, the toughest thing for an autonomous vehicle to handle is other people — particularly reading and reacting to pedestrians, cyclists, and other human drivers. So we’re a ways from true self-driving cars that require no human intervention. It might be years; it might be decades.

What’s more, the transition is likely to be messy. Self-driving cars are most valuable when all the cars on the road are self-driving (that’s when you can get these cars to platoon closely together, for instance, or move more quickly through intersections). As long as there are still some human drivers on the road, though, it’s much harder to get the full benefit from autonomous vehicles. Perhaps the shift will happen naturally, as insurance rates for "manually driven" vehicles go up. Or perhaps cities will have to force the transition through policy.

So for now, think of this report as more of a utopian daydream — the culmination of a slow change in vehicle technology that will eventually let cities devote less space to vehicles and more space to, well, everybody else.

Cities can reclaim vehicle space in low-tech ways, too

Up until now, we’ve mostly been looking at newfangled technology: connected ride-sharing systems and autonomous vehicles. But it’s worth adding that cities don’t have to wait for Silicon Valley to come along before they can reclaim space from cars. There are plenty of low-tech solutions, too, from boosting mass transit to promoting walking and cycling to simple changes in parking policy.

Just as one example, Donald Shoup, an economist at UCLA, has long argued that cities have overbuilt and over-mandated parking. They do this partly by providing free street parking for all. But perhaps more importantly, many cities require all new developments to include specified large numbers of added parking spaces.

This is essentially a mandate for more parking — even if the demand isn’t there. In Washington, DC, the underground spots many developers build to comply with these minimum requirements cost between $30,000 and $50,000 each. It ends up driving up housing prices. And, of course, it means less space for other purposes.

Shoup has argued for a whole spate of changes in parking policy. But one of his simplest recommendations is to simply do away with minimum requirements for off-street parking for new buildings. "I'm pro-choice," he told Vox. "Let the developers build however many parking spots they want." (Developers would no doubt still build parking spots to accommodate demand — they just wouldn’t be required to build more than the market could bear.)

It’s not as sexy as a city full of shared, autonomous, connected, electric vehicles. But it’s the same principle. Cars take up a lot of space. One way to make cities better and more prosperous would be to find ways to reduce that space.

Further reading:

This story is part of The new new economy, a series on what the 21st century holds for how we live, travel, and work.