Urban transportation is a question of geometry. We simply can’t fit as many cars onto a city street as there are people.
You never see a driver in a car commercial stuck in gridlock. There’s no footage of frantically circling a city block to hunt for a parking spot, or a driver with road rage lobbing slurs at a cyclist. In fact, few ads even take place in the city: exquisite aerial shots show vehicles hurtling through a secluded forest or navigating hairpin turns down a mountain valley (usually with a “closed course” or “professional driver” warning tucked at the bottom).1 Universally hated day-to-day problems with automobility only make it to the screen as a means of demonstrating how this vehicle is the means to avoid such situations; the underlying message is freedom, tranquility, efficiency.
Of course, these are fanciful depictions of what we hope driving will be. Commuting by car tends to be alarmingly frustrating. Congestion leads to calamitous effects: mental duress, missed appointments, higher fuel consumption, road rage, and impeded emergency vehicles.2 In 2018, traffic congestion cost the U.S. an estimated $87 billion in lost productivity.3 The chief policy response has been the constant expansion of roadways, which itself only incites more demand for driving and sprawl.4 Urban road mileage grew by 77 percent between 1980 and 2014, while the U.S. population increased by only 41 percent; roads and highways receive far greater funding, with considerably less bureaucratic delay, than public transit.5
The average automobile is used for only an hour a day; it’s parked for a stunning 96 percent of the time.6 As each vehicle takes up about 100 cubic feet, that requires an enormous quantity of space.7 In 2010, for example, some 14 percent of Los Angeles County’s incorporated land was committed to parking.8 This institutionalization of the car is deeply embedded in the U.S. federal tax system. The country spends about $7.3 billion per year to encourage people to drive to work through the federal income tax exclusion for commuter parking, which principally benefits higher-income earners.9 Most cities also mandate that property developers install a certain number of off-street parking spots. About this policy of “mandatory parking minimums,” TransitCenter wrote: “Few planning practices do so much harm to cities: Parking mandates induce more driving, increase housing costs, and put climate goals out of reach.” Cities including San Diego, Minneapolis, and Sacramento are working to abolish parking mandates near transit stops.10
Driving isn’t only frustrating and spatially inefficient—it’s exceedingly dangerous. About 40,000 people in the U.S. died in motor vehicle collisions in 2018, while over 4.5 million people were injured badly enough to require medical attention.11 Globally, an estimated 1.35 million people die each year in crashes.12 Car deaths are worst in states with high speed limits (the state with the highest relative road fatality rate, Montana, has a rural interstate speed limit of 80 miles per hour), as well as in poor states where drivers are more likely to use older vehicles without crash protection.13 Other key factors include alcohol, seat belt laws, and distracted driving. A full 10 percent of fatal crashes in the U.S. and 15 percent of crashes causing injuries resulted from distracted driving including cell phone use, eating, talking to other passengers, and adjusting the radio.14 As National Safety Council president Deborah A.P. Hersman noted in 2018: “The price we are paying for mobility is 40,000 lives each year.”15
It’s not just drivers who are affected: 16 percent of Americans killed in motor vehicle crashes in 2016 were pedestrians.16 While motorist deaths have declined slightly in percentage, the number of pedestrians killed by vehicles has increased by 35 percent over the last decade.17 Pedestrians using wheelchairs are 36 percent more likely to die in a crash than pedestrians who are not in wheelchairs.18
These deaths are also highly racialized. Between 2005 and 2014, Native American pedestrians were almost five times more likely to be killed than white ones, while Black pedestrians were almost twice as likely to be killed.19 Similarly, cyclists in communities of colour are more likely to be hit by a car due to lack of safe biking infrastructure.20 That’s despite polling that suggests low-income communities of colour rely the most on biking for transportation and recreation, and have a greater desire to bike in protected lanes than white people do.21
Much of the danger comes down to speed and size of vehicles. Gregory Shill of the University of Iowa has observed that speeding kills about as many people as drunk driving, yet there are no national campaigns to reduce speeding.22 There’s a direct correlation between the speed of a vehicle and the chance of a pedestrian surviving a collision; in Philadelphia, there are ten times more deaths on roads with average speeds of 45 miles per hour than on roads with speed limits of 25 miles per hour.23
The increased sale of larger vehicles further worsens the chance of a pedestrian or cyclist surviving a crash, as the person can be pushed under the higher vehicle body. Pedestrians are two to three times more likely to die when hit by an SUV or truck than by a passenger car.24 In response to the unveiling of Tesla’s controversial Cybertruck in November 2019, the head of Australia’s crash test authority said its shape and materials would likely pose increased risk to pedestrians and cyclists.25 A cycling website dubbed it “a rolling tank against cyclists & pedestrians.”26
Many advocates of next-generation automobility are genuinely concerned about the realities of congestion, sprawl, and fatalities. The Shared Mobility Principles for Livable Cities, a manifesto of sorts created by a group of transportation NGOs and signed onto by Uber and Lyft, emphasizes the need to design compact cities, limit single-passenger and oversized vehicles, and promote mobility for pedestrians and cyclists.27 No one transportation mode would dominate: users will plan out their trip on smartphones in a single app, switching between everything from scooters to ride-hailing services to public transit to get to their destination.
Most new vehicles, including electric vehicles, feature significant upgrades in safety software due to small advances in autonomous technologies. The proliferation of partial automation crash-avoidance features—including blind spot monitoring, lane departure warning, and forward collision warning—is expected to help prevent 133,000 injury-causing crashes and 10,100 fatal crashes in the U.S. every year.28 The issue of electric vehicles being too quiet, endangering pedestrians and especially those with limited visibility, can be fixed by requiring a sound emitter. With that said, new features like a media player, climate control, and cell phone—available by using the seventeen-inch touchscreen display in a Tesla Model S, with expandable menus that require attention be taken off the road to use them—may compromise safety improvements with novel distractions.29 A study by the American Automobile Association found that programming navigation can distract a driver for an average of forty seconds; a driver only needs to take their eyes off the road for two seconds to double the risk of a crash.30
The sharing component particularly excites next- generation automobility enthusiasts on this front. Rather than having to own a personal electric vehicle, users are soon projected to be able to summon a car with their phone that will be shared with other riders and, at least in the short term, driven by the contractor who owns the vehicle. This scenario is expected to result in many current car owners selling off their vehicles, in turn freeing up space that would otherwise be consumed by congested traffic and parking. Ride-hailing executives have embraced this rhetoric as a marketing opportunity. Uber CEO Dara Khosrowshahi said the company’s entire plan is “aimed at eventually replacing car ownership itself. Cars are unused 95% of the time and take up enormous amounts of space, in parking etc—we want to give that space back to the city.”31 In 2016, Zipcar announced that it had surpassed the one million membership mark and had assisted in pulling over 400,000 vehicles off the road.32
Lyft has followed a similar trajectory in its marketing approach, emphasizing the dent its services have made in personal car ownership. In one promotion, it offered credits to users who gave up their vehicles for a month.33 Being able to count on reliable transportation modes that don’t require a single-occupancy vehicle adding to congestion, parking, and potential collisions seems an incontestable net positive.
The hype for fully autonomous vehicles goes a step further. In the future, according to this vision, all cars will interact with each other to ensure perfectly planned movements, eliminating most safety issues and congestion as we know it as each vehicle determines the most efficient means of getting to a destination. Downtown parking spots will be rendered irrelevant, as the autonomous vehicle can move riders at pretty much all times. In turn, this radical shift will allow for the densification of urban areas. Shared ride-hailing will no longer depend on the availability of drivers, so riders can be picked up within minutes at any time of the day. They can then spend their stress-free rides reading, working, or catching up on a Netflix show.
Because such a high percentage of crashes are caused at least in part by human error, the industry claims that fatalities and injuries will drop by at least 90 percent with the proliferation of their technology.34 The National Highway Traffic Safety Administration has calculated that the “critical reason” for 94 percent of crashes can be attributed to human error. That doesn’t mean it’s the only cause; vehicle-related causes, including tire and brake failures, and slick roads also play significant roles. Rather, human error is the “last failure in the causal chain of events leading up to the crash.” In General Motors’ 2018 report on self-driving vehicles, it told readers to “imagine a world with no car crashes.”35 Almost every negative thing we associate with automobility, we’re led to believe, will be erased by a highly intelligent robot.
Yet fundamentally, the crisis of urban transportation boils down to a spatial argument. It’s about the efficient use of a limited amount of square footage. Regardless of what’s powering the individual automobile, the vehicle itself takes up an unavoidably large amount of space. Space in a city is a zero-sum game; every square foot that’s prioritized to one form undermines the possibility of another. While demand- responsive services like Uber or Lyft don’t rely on street parking in the same way that a personal vehicle does, they still require cars to drive in a meandering path instead of the straight line that a fixed-route bus drives—making them a highly inefficient mode of transportation.
Private motor vehicles can only transport between 600 to 1,600 people per hour.36 On-street bus lanes can move between 4,000 and 8,000 people. A dedicated light rail or bus rapid transit lane can serve between 10,000 and 25,000 people per hour—or a full fifteen times what motor vehicles can in the best-case scenario. Peter Norton told me: “When we’re talking about cars in cities, it’s very often like trying to drive a nail with a wrench. It can be done. It just makes no sense to be done that way.”
This spatial sensibility is already playing out in real time.37 An estimated 59 percent of trips taken with a ride-hailing service add additional cars to Boston’s roads (the remaining percentage would have either taken a taxi or a private vehicle).38 Traffic congestion in San Francisco increased by 62 percent between 2010 and 2016, with Uber and Lyft bearing responsibility for over half that increase; travel time would have increased by only 22 percent without ride-hailing.39
Driving from one part of town to another for a pickup doesn’t remove the presence of a vehicle: it only substitutes it. Ride-hailing vehicles are also notorious for blocking transit lanes during pickup and drop-off; ride-hailing vehicles made up two-thirds of the violations recorded in San Francisco in three months of 2017 for driving in transit lanes.40 Over three-quarters of violations in that same period for obstructing a bike lane or traffic lane were against ride-hailing vehicles. Not only are companies like Uber and Lyft worsening gridlock by adding more cars to the street, they’re slowing down and endangering more environmentally and socially beneficial modes of transportation in the process of loading or dropping off riders.41
Partial autonomous technologies in newer cars do represent advances in vehicle safety—at least in theory. But recent testing suggests that test cars equipped with automatic emergency braking and pedestrian detection alerts strike dummy pedestrians 60 percent of the time during daylight, 89 percent of the time when it’s a child-sized version.42
And such technologies don’t yet apply to ride-hailing, as many drivers are using older vehicles. Traffic fatalities in the largest 100 cities in the U.S. have not declined since the arrival of Uber and Lyft; researchers in 2016 concluded this statistic “should provoke skepticism of broad claims regarding the citywide effects of rideshare services in reducing traffic fatalities.”43 More recent studies indicate the launching of Uber or Lyft in a city actually increased traffic deaths by between 2 and 3 percent since 2011.44
While ride-hailing vehicles aren’t inherently more dangerous, they do add more vehicles to the road and make it easier for people to be in a car, resulting in a range of negative effects, including fatalities.45 One in every six vehicles being used by Uber and Lyft drivers in New York City and Seattle have outstanding recalls, including for issues like faulty airbags and potential engine failure; a safety advocate recommended that before using a ride-hailing vehicle, riders use the myCARFAX phone app to look up its licence plate for open safety recalls on an online database.46
There are a wide array of safety concerns with autonomous vehicles, including failures of hardware and software, increased risk-taking by passengers such as not wearing seatbelts, higher risks resulting from platooning (when autonomous vehicles travel closely behind each other to minimize drag), and the overall increase in travel by vehicles—which itself increases the chance of a crash.47 Todd Litman of the Victoria Transport Policy Institute wrote that successfully operating an autonomous vehicle on public roads will require “orders of magnitude more complex software than aircraft” with extremely high stakes in case of failure.48
That’s further complicated by weather conditions such as heavy rain or snow that may get in the way of essential radar and lidar sensors. “Even a half-second delay in processing could result in death,” Litman told me. “There’s good reason to be a little bit skeptical.” Think of how many times your computer or smartphone may have glitched or crashed; now imagine that technology is controlling the vehicle that you’re travelling in at high speed behind a semitrailer. There’s also concern among autonomous vehicle watchers that cars might be hacked and hijacked.49
A massive challenge facing the industry is the mixing of human drivers with computer control in the mid levels of automation. Humans have proven to be very inadequate when they are required to monitor the driving but not control it at all times. Elaine Herzberg, a forty-nine-year-old unhoused woman, was killed by a “self-driving” Uber in Arizona when its backup safety driver was distracted, watching a TV show on her phone.50 Several drivers of Tesla cars have died in crashes while relying on the vehicle’s Autopilot system.51 Tesla CEO Elon Musk has blamed several incidents on driver complacency, rather than the technology.52 Yet Musk’s claims of the unprecedented safety benefits of Autopilot are unverified outside of Tesla itself; government regulators lack sufficient data to confirm them.53
Up to 11 billion miles of autonomous vehicle testing will be required before completely reliable statistics are produced about average fatality rates compared to human-driven vehicles, a process that would take some 500 years of testing—and require 100 autonomous vehicles driving 24/7 at 25 miles per hour.54 To help put that figure in perspective, Waymo, by far the most advanced company on this front, reached only 10 million miles on public roads by October 2018 with its 600 autonomous vehicles.55 As a result, alternative methods of testing have been recommended such as accelerated testing, virtual testing, and simulations and mathematical modelling. Even then, Rand Corporation experts have concluded: “It may not be possible to establish the safety of autonomous vehicles prior to making them available for public use.”56
A poll of 1,000 U.S. residents in May 2018 concluded that 73 percent of drivers wouldn’t feel comfortable riding in an autonomous vehicle. That was up from 63 percent in late 2017; the spike was likely caused in part by a series of high-profile fatal accidents involving Uber and Tesla autonomous vehicles.57 Fifty-four percent of the American public prefer that autonomous vehicle testing not be conducted in their municipality.58 Autonomous vehicle testing by Waymo has also been interfered with by local residents in Arizona. The vehicles have been the recipients of dozens of attacks, including slashed tires, rock throwing, and other vehicles trying to run them off the road.59 In late 2019, the Washington Post profiled Karen Brenchley, a veteran computer scientist living in Silicon Valley who is extremely concerned about autonomous vehicle testing: “The problem isn’t that she doesn’t understand the technology. It’s that she does, and she knows how flawed nascent technology can be.”60
Autonomous vehicle proponents have already indicated that successful deployment of the technology will require persuading pedestrians not to jaywalk and to “behave less erratically.”61 In fact, it will likely necessitate a complete overhaul of cities and towns, making them even more hostile to pedestrian and cyclist traffic.62 Gill Pratt, head of the Toyota Research Institute, has said “the trouble with self-driving cars is people”—a suggestion consistent with the century-long history of automobility in North America.63
Andrew Ng, an autonomous vehicle investor, also received significant criticism for his response to the “pogo stick problem,” in which self-driving cars are expected to be able to sense and stop for a pedestrian on a pogo stick in the middle of a highway.64 “Rather than building AI to solve the pogo stick problem, we should partner with the government to ask people to be lawful and considerate,” Ng said in 2018. “Safety isn’t just about the quality of the AI technology.” Renowned roboticist Rodney Brooks wrote a scathing blog post in response to Ng’s comments, suggesting they “completely upended the whole rationale for self-driving cars” and the idea of government forcing change would turn pedestrians into “the potential literal roadkill in the self-satisfaction your actual customers will experience knowing that they have gotten just the latest gee whiz technology all for themselves.”65 In 2016, Mercedes-Benz’s manager of driverless car safety told Car and Driver that in the situation of a crash between an autonomous vehicle and pedestrian, the car would be programmed to save the driver over the pedestrian every time.66
There are growing ethical concerns about the programming of autonomous vehicles when it comes to decision-making about who lives or dies in a potential crash. Such dilemmas have included extreme hypotheticals such as “whether to run over a group of schoolchildren or plunge off a cliff, killing [the car’s] own occupants.”67 Others have suggested that racial in equity is hardwired into autonomous vehicles, as they are less likely to detect dark-skinned pedestrians.68
Preliminary reports from the National Transportation Safety Board indicate that the self-driving Uber vehicle that killed Herzberg detected her a full six seconds and called for an emergency brake 1.3 seconds before contact. Uber had programmed the vehicle not to stop in that situation because frequent braking could make for an “uncomfortable and jerky” ride for passengers, which may undermine ridership and profits.69 Arizona prosecutors announced in early 2019 that they wouldn’t pursue criminal charges against Uber for the killing of Herzberg, which Angie Schmitt of Streetsblog argued “signals that tech companies won’t be punished for taking egregious risks with their untested technology even when the worst happens.”70 The safety board’s report released in late 2019 found the vehicle failed to classify Herzberg as a human because “the system design did not include consideration for jaywalking pedestrians.”71
Ashley Nunes of Massachusetts Institute of Technology has suggested that these are signs of things to come, that autonomous vehicle promises are “on life support,” and that
even if robotaxis had an unrealistically high utilization rate and even if their investors lowered their profit expectations, the cost of providing safety oversight would need to be substantially reduced (to below existing minimum-wage levels) in order for robotaxi fares to be cost competitive with owning an older vehicle.72
“It’s the worst-case scenario,” Greg Lindsay, senior fellow of NewCities, told me. “Autonomous vehicles are going to literally run roughshod over people and there’s not going to be any real oversight over it.”
To recoup the technology’s significant costs, autonomous vehicles will likely have to be on the road almost 24/7.73 Another fear for industry watchers is “zombie vehicles,” or personal autonomous cars that continue driving around instead of finding a parking spot in order to avoid parking costs. While such an approach would feasibly reduce the amount of geographic space required for parking, it could massively increase emissions and gridlock. Without a punitive tax like congestion pricing to deter the practice, autonomous vehicles are very likely to cruise at low speeds rather than park. In a best-case modelling of San Francisco, only 2,000 autonomous vehicles would slow traffic to less than two miles per hour.74
Daniel Aldana Cohen warns that autonomous vehicles could help draw even more residents into ever-sprawling suburbs, rather than finding ways to live in more dense environments and making those areas equitable. He says that we should be “very, very worried” about cities repeating their histories of suburbanization and segregation in new iterations of “eco-apartheid.” This kind of future could be highly lucrative for autonomous vehicle companies using data collection—every minute spent in a car is sellable information—as well as to developers seeking the next round of areas to gentrify.
Other questions remain about who would pay for road maintenance and construction in the future, particularly if sprawl continues unchecked. Autonomous vehicles will depend on high-tech roads that integrate sensors and 5G broadband—but advocates rarely address how such upgrades would be paid for.75 If realized, autonomous electric vehicles would rely on this existing infrastructure while eliminating the gas and diesel tax, a main source of roads funding.76
“It’s scandalous that drivers are getting away with socializing the cost of their pollution. But it’s even more scandalous that autonomous vehicle companies are basically skipping that entire line in their assumptions. These are smart people. I think they’re perfectly aware that cars don’t pay the cost of their infrastructure.”
—Matthew Lewis, climate and energy policy consultant, Berkeley
Funding for roads is already in a difficult spot, as the U.S. gas tax has stagnated since 1993, decreasing its purchasing power by 40 percent.77 Since 2008, $140 billion has been transferred to the Highway Trust Fund from general revenues to make up for the shortfall from the gas tax, with $70 billion transferred in 2016 alone.78 An average household in the U.S. spends $597 per year in general tax revenue to fund road construction and repair, as well as more in subsidies.79 Almost 90 percent of local roads—meaning non-highway infrastructure—were paid for by the public, and only 11 percent specifically by motorists.80 “In major cities, capacity remains the major issue,” Yonah Freemark of the Transport Politic told me. “We’d have to build new highways if we wanted to rely on autonomous vehicles for all our transportation needs.”
Urban sprawl and road dominance also make climate- change-exacerbated extreme weather events more difficult to cope with. Rising temperatures will put stress on infrastructure like roads—many of which are already in poor shape. Since concrete is a highly impermeable surface, heavy rainfall isn’t absorbed by soil and instead overloads stormwater drainage systems and basements. In recent decades, wetlands and marshes, which offer natural water storage and filtration services, have been filled in for agriculture and paved over to build roads and suburbs. This worsens the effects of heavy precipitation.81
Next-generation automobiles fail to mitigate this growing crisis of public subsidies and resulting ecological destruction. The promise of mostly shared vehicles that prioritize integration with publicly owned modes of transportation ignores decades of history demonstrating the singular focus by automotive interests on securing market share and profits. The risks that cities and towns are being forced to absorb as result of these experiments could lead to greater congestion and traffic fatalities, and encourage sprawl, exacerbating flooding.82 Given a long history of racist discrimination and segregation, it’s very likely that these impacts will disproportionately harm low income communities of colour.
There’s a popular meme that compares transportation modes by the amount of space that they require to move sixty people. The cars stretch into the distance, taking up all lanes of traffic. The bikes line up only slightly beyond the group of people huddled together for the photo. And, quite sensibly, the same number of riders fit into less space than the bus takes up. Variations of the meme will add extra frames of electric vehicles, or Uber rides, or autonomous cars. Of course, they take up the same amount of space as regular cars.
As Jarrett Walker often points out, urban transportation is a question of geometry. We simply can’t fit as many cars onto a city street as there are people. Building new roads to ease congestion only leads to induced demand—and more congestion. The solution is to use space more efficiently and in a way that prioritizes people over vehicles. For while congestion is frustrating to everyone, it’s often people who can’t afford to live in or near the downtown that are punished the most by poor planning that encourages dangerous and inefficient modes of transportation. Luckily, these problems can be very simple and affordable to fix.
Toronto’s year-long King Street streetcar pilot project—which ran from November 2017 to December 2018—is a glowing example of what is possible. At the outset, drivers represented only 16 percent of users on the downtown corridor but were allocated 64 percent of the space. Meanwhile, pedestrians (half of the users) got only 25 percent of the space, while transit was about evenly matched at one-third of users and space.83 The presence of private cars has historically impeded speed and reliability of the busy streetcar route; as a result, the pilot project prohibited private vehicles from using intersections that crossed the streetcar route in the downtown core, giving full priority to streetcars. On-street parking was also banned in the pilot area.
As you might expect, the results were extremely positive. Streetcar travel time improved by five minutes in each direction. The slowest travel time during the pilot was in the same range as the average travel time before it.84 Average ridership went up by 17 percent. Driving time on adjacent streets was impacted by a mere minute. Newly opened up public space along the curb lanes hosted dozens of amenities including cafes, public art, and seating areas.
In April 2019, Toronto city council voted to make the pilot project permanent.85 Costing only $1.5 million to implement and another $1.5 million to make permanent, this project is a textbook example of how small changes that give transit modes priority can massively improve reliability and ridership.
Dedicated bus lanes succeed based on the same merits, removing the less spatially efficient private vehicles from their path and ensuring dependable transit. New York City’s 14th Street Transit & Truck Priority Pilot Project, introduced in October 2019, reserved a lengthy stretch of the major Manhattan street for buses, trucks, and emergency vehicles between the hours of 6 a.m. and 10 p.m. A report written by renowned transportation engineer Sam Schwartz found that the dedicated lane resulted in a 36 percent improvement in average weekday travel times for buses and a 24 percent increase in bus ridership.86 Automobile travel times on most adjacent streets increased by a single minute.
This isn’t to suggest that pilot projects are embraced by all constituents: many are vehemently opposed, especially by business interests who fear loss of affluent customer access by car. Several businesses on Toronto’s King Street erected ice sculptures of a middle finger facing the streetcar line, with one owner alleging that his delivery business had declined by 75 percent since the pilot project began.87
Promotion of transit and active transportation—rather than doubling down on automobility with a new generation of technologies—guarantees safer streets as well. In 2018, 251 people were killed by transit crashes in the U.S.88 That represents a mere 0.7 percent of people killed in transportation- related incidents.89 A 2017 study of a decade’s worth of police reports along major traffic corridors in Montreal found that private automobiles were responsible for 95 percent of pedestrian and cyclist injuries, and that a 50 percent shift from people driving to using the bus would reduce transportation-related injury rates by 35 percent.90
There are many reasons for the inherent safety of public transportation, ranging from the expertise of transit drivers, to frequently maintained vehicles, to physically separated tracks or lanes. Reliable service can also reduce the chances of drunk driving by providing options other than a personal vehicle; as a 2014 CityLab article asked: “What If the Best Way to End Drunk Driving Is to End Driving?”91 The more infrastructure built for transit—whether it be dedicated bus and streetcar lanes, or light rail lines—the safer it becomes, as such modes aren’t required to interact with automobiles. The same goes for wider sidewalks and protected cycle tracks, which allow people to get places without fear of potentially fatal interactions with cars.
Not that transit is invincible or flawless. Several Amtrak trains have derailed in recent years, resulting in several deaths and hundreds injured.92 A brutal crash in early 2019 by an Ottawa transit double-decker bus into a station overhang resulted in three deaths and twenty-three injuries.93 The American Public Transit Association has identified a wide range of safety improvements that can further protect riders and other street users, including more driver training, new transit vehicles, grade separation, and new station design.94 Increased labour power and a requirement for transit agencies and city councils to take transit operators’ demands seriously would also improve safety for everyone.
Transit crashes do happen, but at a far lower rate than for automobiles. Litman wrote in his 2018 report about safety that “disproportionate media coverage can also stimulate transit fear. Because transit accidents and assaults are infrequent, they tend to receive significant media coverage.”95 A 2018 investigation found that 1,235 people have died on subway and rail tracks in Canada since 2007, a “vast majority” of them ruled as suicides.96 But there’s a simple fix to help reduce such fatalities: platform edge barriers. These sliding doors that open only once a train has stopped have reduced deaths on transit by more than half in places where they have been deployed, like Hong Kong. Toronto’s transit agency has estimated that it would cost between $1 billion to $1.5 billion to install platform barriers on all seventy-five subway stations, a seemingly small price to pay for reducing deaths.97
Public transportation is subject to extreme weather impacts. Hurricane Sandy in 2012 flooded the New York subway system, causing billions of dollars in damage that is taking years to repair. Other cities can follow the example provided by New York City, which has spent the last several years attempting to floodproof the thousands of street-level openings to the subway with flexible stairwell covers, portable vent covers, and waterproof gates.98 While continually sprawling cities undermine attempts at such planning, more centralized communities and transportation systems allow for more coordinated infrastructure—flood walls, elevated tracks or roadways, dikes—to ensure mobility during and after extreme weather events.
Billy Fields, a political science professor at Texas State University and an expert in adaptation urbanism, told me that massive flooding experienced in Copenhagen in 2011 and 2014 encouraged the Danish capital to rethink its design of streets and open spaces. In one area, the city tore up asphalt for parking and redesigned it to better regulate the flow of stormwater into drainage systems. Such “green-blue” infrastructure includes flowerbeds and play areas with graduated levels that store water in a flooding event.99
But when large cities like New York City do think about building transit, it’s often in illogical and climate-change denying ways. In 2016, Mayor Bill de Blasio proposed building a $2.5 billion streetcar route from Queens to Brooklyn that would pass through a major floodplain. An investigation by the Village Voice argued that the plan “isn’t just a clichéd amenity for developers and a waste of the administration’s political capital; it also threatens to undermine the city’s efforts to plan for the devastating and imminent effects of climate change.”100
Transit and active transportation aren’t one-stop shops for communities. It’s going to require a much broader set of changes to accomplish improved safety and climate mitigation in a way that is genuinely just and protects low-income communities from gentrification and policing. Retrofitting suburbs and commuter towns for a less car-dominated future is an underappreciated step in this process. Currently, many transit agencies don’t even try to provide adequate service to the suburbs; the logic is that the automobile reigns supreme today, so it always will. But these qualities are not inherent to the geographies of suburbia or rural communities; such places can boast high-quality and dependable transit.
The decision to render low-density communities as unserviceable is chiefly a political one, and it ignores the reality that many residents require or prefer non-automotive transportation. In other countries around the world, an explicit decision was made in the 1960s to expand transit systems to new suburbs.101 In a similar vein, many newer communities, built around the private car, are defined by cul-de-sacs and curving roads that, compared to the grid-like planning in older cities, make it difficult to access transit as a pedestrian.
It will cost a lot to play catch-up and foster the conditions for a cultural shift toward transit, but it can be done. Paul Mees—an Australian transit expert—argued that the appeal by urbanists to density as a prerequisite for quality transit was a fundamentally anti-transit argument. As he put it in his 2010 book Transport for Suburbia: “The notion that urban form, rather than transport policy, determines transport outcomes is convenient for [highway planning agencies]. It can also suit those responsible for providing public transport, because it pins the blame for poor services on suburban residents rather than public transport providers.”102
Imagining that a minimum density threshold is required to support sufficient transit usage effectively writes off a vast portion of residents in communities across the continent. It doesn’t mean density and urban form don’t matter—only that they’re not the end of the story. Just over half of U.S. residents, 55 percent, live in a suburban county, compared to only 31 percent in an urban county and 14 percent in a rural.103 In Canada, an estimated 67 percent of the country’s metropolitan population live in “automobile suburbs,” while only 14 percent are in “active cores.”104 Demographics in the suburbs have been shifting dramatically in recent years, with steady increases in rates of poverty and new immigrants.105 Commuter towns, or “exurbs,” are rapidly growing in population.106
Plenty of reasonable cultural, environmental, and financial critiques can be made of suburbia. This pattern of urban development is unsustainable on many fronts, and should be seriously curtailed.107 But given the persistence of urban sprawl and the fact that poverty is now suburbanizing, it appears critical for governments to make a serious effort to provide public transit for people living there.
Mees was by no means an advocate for suburban sprawl. But for him, the fundamental challenge wasn’t urban form—which he said was used as an excuse for not addressing service quality—but urban structure, requiring transit service that used buses to shuttle people to more efficient rail service.108 He called this “the network effect,” when “public transport imitates the flexibility of the car by knitting different routes and modes into a single, multi-modal network.”109 While transit agencies obsessed with ridership numbers might worry about the fiscal feasibility of such a coverage-oriented approach, Mees and others have argued that it’s the key to providing enough consistent ridership for high-efficiency rail lines to also succeed.
An example of this is happening in the suburban city of Brampton, Ontario, which has made serious progress in rehabilitating the car-heavy landscape for transit options by establishing a grid-like system and two-hour transfers.110 In 2010, the city introduced an express bus service called Züm. It includes large heated bus shelters and traffic signal priority that allows buses to jump the queue at intersections. Ridership has tripled since 2006.111 Residents who previously drove are retiring their cars for commutes and relying instead on frequent and dependable public transit.
Meanwhile, Calgary—often thought of as hopelessly sprawled—now has the second-highest light rail ridership on the continent, spanning to many parts of the farthest communities (although recent budget cuts to bus service now compromise that system).112 Of course, it’s going to take a lot more work to turn the suburbs into sustainable human-scale communities, particularly through government commitments to fund significant public housing.113 These are objectives to be sustained over the next several decades.
But there is plenty of potential for interim measures—and mandated automobility, which disproportionately discriminates against low-income communities, isn’t inevitable. To again quote Mees: “Transport policy can be changed more quickly and cheaply, and with less disruption, than city density, so it might even be possible to make the necessary changes in time to save the planet.”114
It’s something that we should take seriously. The ultimate objective is to densify existing residential areas with public housing and nearby healthcare, education, and social services to reduce the distance people have to travel to get what they need. That will rely on the power of activists fighting for public housing, universal rent control, anti-eviction measures, and community land trusts. But politicians searching for more excuses to cut public funding to transit shouldn’t be allowed to weaponize an existing lack of density as a reason not to build transit now.
Next-generation automobiles—personal electrics, ride- hailing services, autonomous technologies—will almost certainly reproduce the spatial inefficiencies and wastefulness that enabled sprawl and tens of thousands of traffic fatalities per year in the first place. The profit margins of the venture-capital-backed companies depend on such wastefulness. Resisting the widespread proliferation of such modes and fighting for better access to transit will put us in a far better position to build denser communities rooted in public housing, climate mitigation, and community services.