Insurgent Transit

In Human Transit, Jarrett Walker describes his consulting work as being a domain expert who asks communities questions about the transit networks that they want to build. The questions present two endpoints to a spectrum. Each end has advantages and disadvantages. The position a transit agency takes depends on what the community that it serves values; there are no wrong answers. Walker then can help design a transit network conforming to these positions. There are four overarching questions.

• Ridership or coverage?
• Connections or direct service?
• Peak-first or base-first?
• Exclusive or compromised right of way?

I propose a fifth question: cooperation or insurgence?

A cooperative network anticipates that fixed-route public transit will operate alongside other motorized modes of transportation, like privately-owned automobiles and for-profit ride hailing services. In a cooperative network, public transit focuses on scenarios where it can most efficiently operate, or fulfill needs unmet by the free market. This typically yields a transit network that is used for congregation—like for sporting events and for commutes to business districts—and as a lifeline for those unable to use other modes of transportation due to poverty or disability. Most American transit agencies operate cooperatively.

An insurgent network is designed to take over nearly all non-commercial uses of motor vehicles within an area. The goal is to create regions where the household rate of car ownership is low, and car use within its boundaries is considered unusual.

The Case for Insurgence

Sharing space with other vehicles makes public transit service more difficult and expensive to operate. Reducing congestion on roads allows transit vehicles that run in mixed traffic to move faster between their stops. Faster speed means a lower cost to serve the same number of locations. This allows investment in more transit service. Since this investment makes transit more attractive to potential riders, it can further reduce congestion. Thus, reducing car use kicks off a self-reinforcing cycle of transit use. Reducing congestion also makes transit agencies look more competent. Buses bunching when roads are congested generally isn’t an agency’s fault, but looks wasteful and inept. Free-flowing transit vehicles, unimpeded by private car traffic, don’t have this problem, and are more likely to be on time. Fewer cars on the road sets transit agencies up for success.

Reducing car ownership, rather than just use, has additional value to agencies. When transit agencies undertake large infrastructure projects like subway systems, they must contend with the impact that construction has on existing roads, and the uproar this causes during planning. They must present a variety of alternatives around levels of disruption, adding cost and delays. Ultimately, they may be forced to locate stations in areas suboptimal for attracting riders or supporting transfers in order to avoid disruptions to drivers during construction. Even in less extensive projects, potential bus lanes may get ceded for parking or general traffic. Producing car-free households builds a reliably pro-transit constituency that can counter the car-centric voices that complicate or derail transit projects.

Outside of the direct benefits to transit agencies, reducing car ownership has a positive impact on social welfare. Owning a car is expensive, with AAA estimating that it costs Americans an average of $12,297 per year. Replacing this cost with a transit pass could be a literal life saver for some; according to the Federal Reserve, only 63% of American adults can cover a $400 emergency expense. Because the cost per trip of using a car is smaller than the cost of acquiring and maintaining one, reducing ownership offers an outsized benefit over just reducing use.

Widespread car ownership is a public health hazard. According to the CDC, car crashes are the leading cause of death in the United States. The EPA reports that transportation contributes 28% of greenhouse gas emissions in the county, with over half of that contribution coming from light-duty vehicles. Driving enables a sedentary lifestyle. Neither automated nor electric vehicles solve these problems. There is an inherent danger in making heavy, fast-moving vehicles, driven by non-professionals, ubiquitous in the same spaces where many people live.

Embracing insurgence would shift the mission of transit agencies, but not in an unprecedented way for a government division. Transit agencies typically have a mindset of accommodation. They run service that allows people to do what they want to do, and are not in the business of opposing behavior and compelling people to change. Yet public health departments categorically denounce smoking and provide resources for quitting, without regard for the profitability of tobacco companies or the feelings of smokers uninterested in quitting. Even in a society that values individual liberty, sometimes the role of government is to change harmful social norms. The harms of car ownership are manifold, and public transit agencies should not shy away from putting themselves in an oppositional position.

Designing for Insurgence

A network designed around insurgence establishes an insurgence area, where car ownership will be severely reduced. This is unlikely to be the transit agency’s full service area, but could be a city, part of one, or several of them. Most importantly, people should think of the insurgence area as a contiguous, coherent thing, and give it a name. It should contain everything that those within it need to live, and nearly everything that they could want. Residents leave it occasionally, but not often. It is useful for the area to have a unified government, such that non-transit policies discouraging car use can also be put in place. Choosing the area is more a matter of gauging perception than knowing demographics. When the people of an area reach the consensus that “people here do not need a car”, insurgence has been achieved. Progress is measured by the rate of household car ownership there.

In my access analyses, I’ve used Seattle as an insurgence area. It is a coherent area that people reference by name. It contains the necessities for living, like jobs for people with varying abilities and expertise, a variety of types of housing, medical facilities, grocery stores, and schools. There are museums, parks of various sizes and types, professional sports, music venues, restaurants, and retail establishments. A person can live a full and varied life there without leaving it regularly.

Some aspects of driving cannot be replicated with fixed-route public transit, even when considering only non-commercial use within an insurgence area. Public transit can never feel as private, nor can it make all trips door-to-door. There are some recreational uses where a car is the only choice; the musician who lugs a drum kit to a few shows per month will likely remain a car owner. Cases where privacy, door-to-door service, or hauling capability are truly necessary make up a tiny share of urban car trips. Most people could make nearly all their trips without these capabilities, with short-term car rentals filling the rare gaps. The aspects of a car that public transit cannot replicate are rarely necessities, and thus not reasons to deem insurgence impractical.

Yet an insurgent public transit network must provide the core freedom that a car bestows on drivers. Even in cities, people are acclimated to the way a car can get them around, and do not want to give up that freedom. A car can get a person from one arbitrary point to another at any time of day, typically in an amount of time that feels reasonable and predictable. Drivers are unconcerned that the destination that they want to reach is not a popular one. As long as there are roads, they can reach exactly what they want. Drivers don’t worry that the roads that they use will close in the middle of the night. Even when the roads are used less, they remain open. Drivers do not worry about a schedule. They can start their trip at any time. Achieving those properties with public transit requires a specific network design.

The ability of public transit to get a person from wherever they are, to wherever they want to be, whatever the time of day, in a reasonable amount of time is measured with unweighted access. A totally insurgence-oriented network maximizes this measurement within the insurgence area. The region is divided into a grid of uniformly sized sectors. The measurement sums, for each sector and each minute of the day, how many other sectors can be reached within a time budget while walking and using transit. This is similar to the population-weighted access to destinations measurements associated with “ridership networks”—those which favor ridership on Jarrett Walker’s ridership-coverage spectrum. The unweighted measurement differs by considering all parts of an area as equally valuable. The advantage of not weighting is counterintuitive; obviously more people originate in densely populated areas, and some destinations are more popular than others. A measurement without weighting better reflects how well transit allows an arbitrary individual to replicate the freedom that they would have if they were driving. As a consequence of designing around unweighted access, the transit network within the insurgence area tends to have routes that are spread evenly, but not too closely, and run as frequently as possible, all day and night.

A fully insurgence-oriented network would concentrate service entirely in insurgence areas, to the detriment of other place that the transit agency serves. Like Jarrett Walker’s questions, though, cooperation or insurgence is not a binary choice. A transit agency starting to shift its resources toward insurgence would design a network that serves insurgence and cooperative areas differently. Within the former, it would maximize unweighted access, and monitor the rate of household car ownership. A network designed and evaluated according to existing cooperative policies would serve the rest.

Insurgence Isn’t Ridership

Jarrett Walker’s discussion of transit networks that favor ridership over coverage suggests that they reduce car use. Ridership networks allocate transit service to areas that have the densest population and the greatest number of desired destinations. By concentrating service in these areas, transit agencies can run service that is more frequent than if it were spread out. The frequency makes transit available whenever potential riders need it. The density also means that parking space is in short supply. These factors combine to generate ridership. Given this, is orienting a transit network around insurgence necessary, or is that subsumed by designing for ridership? Networks oriented around ridership and insurgence certainly have commonalities, in both the measurements that underly their design and their desired outcomes, but they are not identical.

If an insurgence area is exactly the places within a transit agency’s service area with the greatest density of people and destinations, than designing for ridership would be the same as designing for insurgence. Those areas are unlikely to be contiguous and coherent, though, so people wouldn’t give them a single name. There would be a mismatch between what a person thinks of as the place in which they live, and what is most accessible to them with public transit. They couldn’t say “where I live, people don’t need a car” because the less dense parts of their home are not prioritized for transit service when ridership alone is the focus.

Returning to the example of Seattle, the city is served by King County Metro and Sound Transit, which have service areas of a large county and region that comprises parts of three counties, respectively. While Seattle on the whole is denser with population and desired destinations than any other municipality in these regions, there are areas within Seattle that are less dense than the densest areas outside of Seattle. An insurgence-oriented network would nevertheless concentrate service evenly throughout Seattle, because it is the most viable insurgence area in the region.

The success of insurgence and ridership network orientations is assessed differently. Ridership is trip-focused, while insurgence is person-focused. The former is based on the number of times people board transit, not the number of people who eschew a car because transit is meeting their needs. A high-ridership public transit system may still only be useful for a limited variety of trips. Every rider may also own a car to make up for the network’s narrow focus. A ridership-oriented network is considered successful in this state, though the deleterious aspects of car ownership remain.

Furthermore, these indicators of success align differently with how other government divisions’ initiatives impact urban life. If a city changes its zoning rules to limit where grocery stores can be located, someone who could once walk to do their grocery shopping may have to shift to transit. This would make ridership look better, even though the transit agency did nothing, and no one’s life has improved. A goal around ridership acts as though riding transit is virtuous for its own sake, but its benefit is stopping people from using a more harmful mode. This same zoning change would also increase car ownership. A transit agency that has a reason to track both ridership and car ownership would see the discordance between the two, and assess this situation correctly.

Insurgence’s emphasis on serving all points within the insurgence area should evoke a coverage focus alongside a ridership one. A coverage network allocates routes evenly throughout the service area, to minimize the number of people who are far from any transit service. Unlike coverage, though, insurgence does not reward spreading resources as thin as possible and covering every street with infrequent transit. By encouraging the maximization of unweighted access, it incentivizes finding a combination of frequency, span, and density of routes that increase connectivity throughout the region.

The spectrum between cooperation and insurgence is not entirely disjoint from that of ridership and coverage. That could be said for all of the four questions raised by Jarrett Walker, though. The slower, less consistent transit speeds inherent in compromised right of way threaten ridership just the same as the poor frequencies associated with designing for coverage does, for example. Some amount of overlap does not imply that one question subsumes the other. An emphasis on ridership alone will not bring about every advantage that designing for insurgence would, even if they bear a resemblance.

Can Transit Agencies be Insurgent?

Jarrett Walker’s four questions vary in how constrained transit agencies are in choosing their position. Choosing between direct service or connections is unconstrained; networks favoring either can be constructed with the same amount of resources. Choosing exclusive over compromised right of way is necessarily more expensive, so only some transit agencies can afford to build any. Cooperation versus insurgence is like this too; not every transit agency can entertain insurgence. Those who can, should.

A cooperative transit network, which assumes the widespread ownership and use of cars, can be selective in how it deploys its resources. It can allocate service only where there is an unmet social welfare need, or where those who typically drive find it inconvenient. Insurgence is comparatively inefficient. Creating places where people generally don’t own cars or use commercial ride hailing services starts by designing a public transit network that assumes these don’t exist. This necessitates that transit serves the least dense and least popular areas as well as any other place in the insurgence area, lest these areas become unreachable. Not every agency would have the resources to do this without starving the places outside the insurgence area of service.

Not every urban place lends itself to establishing an insurgence area either. Some cities are too sprawling for their transit investment to serve them at a level that would reduce car ownership, and don’t contain any sufficiently amenity-rich districts that people recognize as a coherent, nameable thing.

Seattle, though, provides a plausible setting for trialing a measured shift toward insurgence. I created a hypothetical redesigned network that serves the entire city with routes that have headways no worse than 15 minutes, 24 hours per day, seven days per week, without any more in-service time than is used today. This boosted the Seattle-average unweighted access—the measurement of how easy it is to get around Seattle no matter the time of day—by up to¹ 26.9% on weekdays and 32.9% on weekends. Local transit service outside the insurgence area was largely left intact, though commuter routes into Seattle were connected to frequent routes within the city rather than running directly to the central business district. A greater shift toward insurgence would raid the suburbs for service, but this network’s small step toward it only required eliminating extra peak trips from routes within the city and increasing reliance on connections over direct routes to activity centers like downtown and the University District. Not every city is like Seattle in terms of transit investment, but it is hardly singular. Insurgence-leaning redesigns are likely plausible in many other cities.

The proposed network for Seattle—and perhaps any insurgent network—alone is likely insufficient for decreasing car ownership to the point where it is truly rare. Instead, insurgence makes a direct contribution to the effort, and also creates a space where other government agencies have cover to implement car-hostile policies. Tolls for operating a vehicle inside the insurgence area and mandatory minimum charges for all municipal and privately-operated parking within it could serve as sticks alongside the carrot of exceptional transit. Nevertheless, favoring insurgence over cooperation does force transit agencies to take on the risk of being the first mover, rather than adapting their service to the existing built environment.

Even for transit agencies that could shift toward insurgence, taking on risk may be an uncomfortable proposition. Many public transit agencies are facing a “fiscal cliff” that threatens their ability to run just their existing service. A less efficient approach to allocating resources seems at odds with the cost savings that agencies are seeking. Nevertheless, an insurgence emphasis may be beneficial as traditional funding sources become less secure. A public transit agency that positions itself as stridently opposing the most common cause of death in the United States is a public health and public safety agency. Perhaps this fosters new perspectives on the role of transit agencies, and new funding streams, like those that allow police departments to grow their budgets even in times of economic turmoil.

The four questions that Jarrett Walker asks in Human Transit allow for diversity in how transit agencies can plan their service, enabling them to meet the variable needs of the communities that they serve. Throughout the book, though, public transit service is uniformly viewed as something that does, and will continue to, operate cooperatively in environments of widespread car ownership and use. While this is a pragmatic choice, reflective of the present conditions in the countries in which Walker thinks the book is most applicable, it is overly constraining. Posing the question of cooperation or insurgence gives public transit agencies a structured way to reconsider their mission, if only within a small part of their service area at first. It could beget a more secure future for the communities that they serve, and for themselves.