For years, it looked like the next big thing in public transportation for the suburban city of Altamonte Springs, Florida, would be an innovative program called FlexBus. Instead of running on fixed routes, buses would respond to demand from kiosks located at specific activity centers. It was, city manager Frank Martz says, “the first demand-response transportation project ever developed in the United States.” Some even referred to it as an “Uber for transit.”

Unfortunately, it didn’t work out. The regional bus operator administering the plan lost key federal funding, and Altamonte Springs had to look for a new solution. “Rather than be mad,” Martz continues, “We decided to solve the problem. We still needed to serve our residents.” 

This time, officials went with Uber itself. This past spring, the Orlando suburb announced a straightforward partnership with the ride-sharing firm, subsidizing citizens who opted to use that service instead of their own cars—particularly for trips to regional rail stations that connect population centers around Seminole County. The pilot has proven popular enough that several municipalities in the area have already launched similar programs. 

Most of what we hear about the relationship between municipalities and ride-sharing startups involves contention. Right around the time Altamonte Springs started this pilot program, a standoff over regulatory details in Austin, Texas, led both Uber and its chief rival Lyft to stop doing business in the city. But Altamonte Springs is an example of how some cities, planners, and scholars are trying to find opportunities within the rise of ride-sharing’s prominence and popularity. MIT’s Senseable City Lab has worked with Uber; UC Berkeley’s Transportation Sustainability Research Center and others have been diving into ride-sharing data with an eye toward public-transportation impacts. And this past March, the American Public Transportation Association released a study assessing how new services can complement more familiar forms of “shared mobility,” and suggested ways that agencies can “promote useful cooperation between public and private mobility providers.” 

“What it’s going to boil down to is how this new system interacts with the existing, traditional system,” says Daniel Rodriguez, a Lincoln Institute fellow who teaches planning at University of North Carolina and has studied transportation innovation in Latin America and the United States. He expects more experiments as cities work to figure out how “to get Uber users to complement the existing infrastructure.” 

That almost exactly describes one of the prime motivations for Altamonte Springs’ Uber pilot: the service was, Martz points out, an existing option that required none of the time-and-money commitments associated with a typical transportation initiative. “The focus could not and should not be on infrastructure,” he said. “It needed to be on human behavior.” In other words, ride-sharing services already respond to demand that has been demonstrated by the market, so how could the city hitch a ride on that trend? 

The answer was to offer local users a subsidy: the city would pay 20 percent of the cost of any local ride, and 25 percent for rides to or from Sun Rail stations, the region’s commuter-rail system. Riders simply enter a code that works in concert with Uber’s “geofencing” technology to confirm location eligibility; their fee is lowered accordingly, and the city seamlessly makes up the difference. “It’s all about user convenience,” Martz says. But he’s getting at a bigger point than ease of payment. Instead of building systems that citizens respond to, maybe it’s worth trying a system that responds to where citizens actually are—and adjusts in real time as that changes. 

Whether this works out in the long run remains to be seen, but as an experiment the risks are pretty low. Martz has estimated the annual cost to the city at about $100,000—compared to $1.5 million for the earlier FlexBus plan. While the pilot is just a few months old, he says local Uber use has risen tenfold—which is why neighboring municipalities Longwood, Lake Mary, Sanford, and Maitland have all joined in or announced plans to do so. (“We’re creating a working group among our cities,” Martz adds, with a focus on managing traffic congestion and “how to connect our cities.”) 

As Rodriguez points out, the land-use implications alone, both short- and long-term, are compelling. On the day-to-day level, affordable ride-sharing as an option for, say, doctor visits or school appointments or similar errands lowers demand for parking spaces. On a higher level, it leverages options that already exist instead of devising more land-intensive projects that can take years to plan and complete.

In a sense, the experiment fits into a broader trend of seeking ad-hoc transportation innovations. Rodriguez has studied experiments from home-grown bus systems to aerial trams in Latin America that supplemented existing systems rather than building new ones. And while at first blush the concept of partnering with a ride-sharing service sounds like something that would work only in a smaller municipality that lacks a realistic mass-transit-system option, he points out that it could actually play a role in bigger cities. One example: Sao Paulo, Brazil, which offers what The Atlantic’s CityLab has called “the best plan yet for dealing with Uber”—essentially auctioning off credits, available to both existing taxi services and ride-sharing upstarts, to drive a certain number of miles in a set time period. The regulatory details (devised in part by former Lincoln fellow Ciro Biderman) aim to give the city options, while capturing and exploiting market demand rather than trying to shape it. 

That captures Martz’s broader attitude. “Why,” he asks, “should the public sector focus on infrastructure embraced by people who used it 40 years ago?” While he readily notes that this line of policy thinking is very much in step with the pro-free-enterprise attitudes in “a very Republican county,” he also insists that local political support for the plan crossed party lines. And more significantly, he stresses that this solution leaves the city much more easily positioned to adjust as technology changes. Carpooling scenarios seem like one logical possibility. And Uber and other technology companies are known to be working on driverless-car scenarios that could prove even more efficient. Martz doesn’t quite come out and say this, but if Uber gets “disrupted” by some more efficient solution, striking up a new partnership would be a lot easier than a do-over on a multiyear region-wide project. “Let market forces carry the day,” Martz says. 

Of course, as Rodriguez notes, all of this remains very experimental at this stage—and a full-on embrace of ride-sharing carries potential downsides. It obviously remains car-centric and not necessarily affordable to broad swaths of many city populations, even with the 20 percent discount. The ability to travel longer distances for lower costs has been a major factor in city sprawl. “This could be another step in that direction,” he observes. 

But the combination of uncertainty and potential is exactly why it’s worth attending to efforts that embrace ride-sharing upstarts instead of fighting them. “There’s no correct answer right now; it’s still an exploration,” Rodriguez cautions. But the likes of Uber do offer one attribute that’s hard to deny for those willing to experiment, he adds: “It’s tangible, and you know it works.” 

Rob Walker (robwalker.net) is a contributor to Design Observer and The New York Times.

In his 1937 essay “What is a City?,” Lewis Mumford described an evolutionary process through which the “badly organized mass city” would evolve into a new type of “poly-nucleated” city, “adequately spaced and bounded”: 

“Twenty such cities, in a region whose environment and whose resources were adequately planned, would have all the benefits of a metropolis that held a million people, without its ponderous disabilities: its capital frozen into unprofitable utilities, and its land values congealed at levels that stand in the way of effective adaptation to new needs.”

For Mumford, such cities, designed with strong public participation, would become the nuclei of new poly-nucleated metropolitan regions that result in:

“A more comprehensive life for the region, for this geographic area can, only now, for the first time be treated as an instantaneous whole for all the functions of social existence. Instead of trusting to the mere massing of populations to produce the necessary social concentration and social drama, we must now seek these results through deliberate local nucleation and a finer regional articulation.”

Unfortunately, since Mumford wrote these words, we have not achieved poly-nucleated cities or regions. Nor have we advanced a theory of urban evolution. Urban theorists have described cities, used basic pattern recognition to detect relationships among the potential components of urban evolution, or offered narrow prescriptions to fix one urban challenge while generating inevitable unintended consequences that pose new challenges. This is because we have never developed a real science of cities. 

For more than a century, planners, sociologists, historians, and economists have theorized about cities and their evolution by categorizing them, as noted by Laura Bliss in a well-documented 2014 CityLab article about the likelihood of an emerging evolutionary theory of cities. They generated multiple typologies of cities, from functional classifications to rudimentary taxonomies (see Harris, 1943, Functional Classification of Cities in the United StatesAtlas of Urban ExpansionAtlas of Cities). But they based these classifications on arbitrarily chosen categories and did little to inform our understanding of how cities became what they are or to presage what they might become. 

Even Jane Jacobs, in a foreword to her 1961 book, The Death and Life of Great American Cities, called for the development of an ecology of cities—a scientific exploration of the forces that shape cities—but provided only narrative accounts of what defined great cities, mostly with regard to design, as part of her ongoing assault on the orthodox planning profession. In some of her later work, Jacobs set out principles to define great cities, based mostly on form, but she never provided a framework to improve the science of urban theory.

Modern urban theory is plagued by several shortcomings. It is not analytic. It fails to provide a framework for generating hypotheses and the empirical analysis to test those theories. And the research, in general, focuses on big iconic cities, rather than a representative global selection of urban settlements that captures the differences between big and small cities, primary and secondary cities, industrial and commercial cities. Importantly, the research provides little guidance regarding how we might intervene to improve our future cities to support sustainable human habitation on the planet. 

The New Urban Agenda—to be announced in October at the third UN-Habitat conference, in Quito, Ecuador—will present consensual global objectives for sustainable urbanization. These objectives provide guidance for United Nations member states as they prepare for the gargantuan task of welcoming 2.5 billion new urbanites to the world’s cities over the next thirty years—culminating the 250-year process through which human settlement moved from almost entirely rural and agrarian to predominantly urban contexts. But before we attempt to implement the New Urban Agenda, we must confront the serious limitations in our understanding of cities and urban evolution. A new “science of cities” would buttress our efforts to get this last stage of urbanization right.

I do not intend to present a new science of cities in this message. Instead, I will suggest a way to frame one that borrows from evolutionary theory. The evolution of species is driven by four main forces, and it seems reasonable that corollary forces help to shape the evolution of cities. These forces are: natural selection, gene flow, mutation, and random drift. And they play out in predictable ways that shape cities—where city growth replaces reproductive success as an indicator of evolutionary success.

Natural selection is a process of impulse and response. It relates to how a city responds to changing external factors (impulses) that support or inhibit success. Impulses can be economic, environmental, or political, but they are, importantly, outside the control of the city. Economic restructuring, for example, might select against cities that depend on manufacturing, have inflexibly trained workforces, or extract or produce single commodities that face changes in demand in global markets. Climate change and sea-level rise will inhibit the success of coastal cities or those exposed to severe weather events. Political impulses might include regime changes, social uprisings, or war. Or they might be something as seemingly minor as a change in allocation formulae for national revenues. Every impulse will benefit some cities and harm others. A city’s ability to respond to different impulses might be a measure of its resilience, which is directly influenced by the three other evolutionary forces. 

Migration (gene flow) helps to diversify the economic, social, and age structures of cities through the exchange of people, resources, and technologies. Presumably, the in-migration of people, capital, and new technology improves a city’s ability to respond to external impulses. Out-migration, in general, would reduce this ability. 

Mutation, for cities, is an unpredictable change in technology or practice occurring within a city. It might be shorthanded as innovation or disruption. 

Random drift involves longer-term changes in cities that result from cultural or behavioral shifts. These might include decisions to maintain or preserve long-term assets, real or cultural. Drift describes the unpredictable ways that cities might change their character. 

As noted, I do not want to lay out a new theory of urban evolution here. I merely want to recommend this direction in order to invigorate our thinking around urban change more rigorously and systematically. A significant amount of work has already gone into quantifying elements of this framework. Risk theorists and insurers have quantified many of the external impulses that challenge cities. Demographers and population theorists have studied human migration, and macroeconomists have studied capital flows. A lot of attention has been paid to innovation and disruption in the last couple of decades. Random drift is a little less studied. But, as Bliss points out, big data and new technologies might help us to detect longer-terms drift. In any case, a larger framework that weaves these disparate areas of work together would advance our understanding of urban evolution. 

On a cautionary note, while an evolutionary theory of cities would be a signal advancement of urban theory, it is useful to remember that, unlike evolution, which is a mostly passive process—species enduring the external forces that act on them—cities, in theory at least, are driven by more purposive behavior: planning. But planners need better tools to drive their practices and to test their approaches. If we are to successfully implement the New Urban Agenda, a toolkit based on evolutionary science would be hugely helpful. As Mumford concluded in his 1937 essay:

“To embody these new possibilities in city life, which come to us not merely through better technical organization but through acuter sociological understanding, and to dramatize the activities themselves in appropriate individual and urban structures, forms the task of the coming generation.”

We at the Lincoln Institute of Land Policy stand ready to support coming generations in comprehensive and scientific analysis of urban evolution and the important role that effective land policies can play in driving it. Our urban future depends on it.