Most contemporary coverage of driverless vehicles is only about driverless cars. But analysis of a suburban automated car future is one step too far down the analytical pipeline. Driverless vehicles can develop as solo cars serving distant houses with limited stops, or as minibuses (and big buses!) serving denser neighborhoods with frequent stops. Booming cities may grow denser in the core even as they sprawl further on the fringe. The future balance of these two development models will be a function not only of what consumers want or what Google/Waymo cooks up next, but of our political land use and infrastructure choices.
These two divergent services of driverless transportation already have clear precedents in existing large cities: commuter rail versus rapid transit. Driverless tech writers are excited about the possibilities; you’ll be able to read, sleep or answer emails during hour-long trips from distant suburbs! New Yorkers will recognize that sounds like Metro-North Commuter Rail. You’ll be able to share a vehicle with other people going the same direction, avoid drunk driving and tweet while you cruise from Williamsburg to FiDi! Ok, that’s the J/Z train.
The point is that driverless tech is not categorically different from good transit and commuter rail. With appropriate policy, mid-size metros like Columbus, Ohio will gain a cheap, flexible microtransit and pooled taxi network as frequent and extensive as big-city traditional transit. Big dense cities will be able to cheaply designate high-capacity driverless Bus Rapid Transit (BRT) routes on every major route. That’s really awesome! But it is only different in degree, not in kind, from today’s best transport practices.
This leaves the question of whether, on net, we’ll see more driverless urban minibuses and driverless BRT, or more personal cars serving sprawly suburban markets.
Transportation and land use are complementary and evolve together, so each transportation scenario is inherently tied to a more compact or a more sprawly land use scenario. Today, cities like Columbus, Nashville, and Austin are trapped in a “chicken and egg” dilemma with infill development: They need transit to support pleasantly dense development but aren’t yet dense enough for frequent traditional transit across most of the metro area.
In the driverless minibus scenario, new microtransit could kickstart infill growth in those cities currently stuck between bad traditional transit and low density. The transportation and “walkability” amenity gaps between midsize urban cores and America’s superstar transit-rich cities would shrink further. Indeed car-light infill growth is already starting in these transit-poor cities, aided by a mix of ride-hailing services and cycling infrastructure to supplement their handful of frequent bus routes.
In the driverless personal car scenario, long-distance commuter cars could drain city centers into new sprawl. Imagine a rerun of postwar urban flight along the Interstate Highway System. What public policies make either the compact, driverless urban bus future or the sprawling, driverless suburban car future more dominant?
Land use choices are made not just by individual preferences in isolation. They’re made in response to constraints of policy. For example, mid-rise infill on defunct parking lots is impossible, even where people are willing to pay for it, when banned by typical land use regulations mandating single-family homes. Conversely, exurban sprawl can’t happen around cities like Portland, where Oregon’s unique land use regulations preserve farmland and forests with a “greenbelt.” Both vertical and horizontal land use restrictions make housing scarcer, but with different transport implications.
Commuting choices, even holding land use regulations constant, are similarly shaped by the interaction of personal preference with policy. Cities like Singapore use congestion pricing to price road space when it’s scarce. It gives drivers a financial stake in the social benefit of conserving road space during rush hour — taking UberPool instead of driving solo, or transit instead of UberPool — but drops the charge when roads are empty. Whether human or driverless, supply and demand for urban road space can only be permanently balanced by pricing.
Autonomous tech is shiny and new but its future will be dictated by the old politics of road pricing and land-use regulation. Most cities restrict dense infill, some cities ban sprawl, Californian cities ban both! The key is, unless most land use regulations are eliminated, the form of future urban development will remain a governmental choice and not a voluntary market phenomenon.
Both infill-supporting autonomous transit and sprawl-supporting personal cars will emerge, of course. But the balance of the two will be politically dictated by cities and states: what density and mixture of new real estate development land use laws allow, what infrastructure or services we build or permit to support it, and how we price and fund our roads.
This piece originally appeared at The Hill