Over the past few weeks, the Times and the Journal have both run pieces on how autonomous vehicles will change both the design of cars and how people will spend their time in them. Both pieces (and the countless similar ones) are breathless with anticipation; even their headlines (“Envisioning the Car of the Future as a Living Room on Wheels” and “Your Next Car May Be a Living Room on Wheels“) betray how little unique thought is being brought to the conversation surrounding autonomous and driverless vehicles. Inevitably, however, these conversations turn to the inevitable question of What Will Autonomous Cars Mean? Since most of the people who read this blog aren’t necessarily folks in the urban planning space, I figured that in light of a spate of articles showing how cool the cars will be (and let’s face it, they’ll be cool) I’d spend a little bit of time talking about what the most likely scenarios are for a driverless future.

(Also, since most people who read this blog know me in Real Life and might even be interested in what I do day-to-day, I figured I’d adapt something I wrote for work recently. This is mostly drawn from a piece I wrote for my agency recently.)

Most individuals thinking seriously about what a proliferation of autonomous vehicles will mean for cities and municipal regions in the United States are in broad agreement on what the best and worst case scenarios might look like.

The best case scenario is largely dependent on whether or not individuals will continue to own their own vehicles. In the best case scenario, autonomous vehicles will largely be owned by companies running large fleets which riders use for transportation on-demand.  This is likely to look similar to the current Lyft model, in which riders use their phones to summon a vehicle which will take them where they need to go. The benefits of this shared model of ownership are due in large part to the higher marginal cost of each trip. For an individual who owns her own car, any individual trip is very cheap; as humans, we’re really bad at understanding the relationship between the fixed, upfront cost of a car and the cost of every individual trip. If often feels like every trip costs only how much gasoline we’re consuming, but in reality we should be spreading out the cost of the car itself over each trip we take.

However, when individuals are forced to pay on a per-minute or per-mile basis, they become much more aware of the cost of driving and how it relates to other transportation modes.  It allows us to understand in a more concrete way how a car trip, for instance, stacks up in terms of cost against a bus or a subway fair. The car trip is almost always more expensive, though that’s a lot harder to intuit when the only immediate cost for a car owner is the price of gasoline. Because shared ownership makes users more aware of the total cost of a car trip, and not just the incremental cost, this pricing structure is likely to lead travelers to consider car usage as one way among many to get around their city. Moreover, because users don’t own a car in this scenario, they’re less inclined to use a car to make their purchase “worth it.”

There is reason to believe that this best-case scenario is actually the scenario most likely to occur under market operations. In the United States, the average car sits idle for roughly 23 hours a day, or just shy of 96% of the time. Because autonomous vehicles could theoretically operate for many more hours out of the day, they will be vastly more efficient which will substantially lower the cost-per-trip. In fact, according to a Deloitte study quoted by the Journal, these efficiencies could cause the cost per mile to drop by as much as two-thirds. Because they will be able to operate at such high levels, the cost to the end user of an on-demand model will be well below the cost of individual ownership which will push people to opt for shared rather than individually owned cars. Indeed, we currently see many car manufacturers and transportation providers positioning themselves to provide transportation-as-service rather than transportation-as-good in the near future. Ford, for instance, describes their autonomous vehicle project as follows: “The vehicle will operate… as part of a ride sharing or ride hailing experience.” Whether they operate their own fleet remains to be seen, but is largely beside the point. The important piece is that they, along with Uber, Lyft, and others, recognize that the “shared” business model will be more profitable. This is a good thing.

What’s more, as people start paying for transportation on a per-mile basis, they’ll naturally want to travel less. This desire for less travel will encourage demand for denser development in order to reduce the cost of moving around. We’ve talked about parking minimums here before, and we’ll no doubt talk about them again, but there’s an interesting model beginning to spring up already in some cities. In some areas, residential developers are being allowed to build fewer parking spaces per apartment if they’re willing to give their residents Lyft and Uber credits. As more and more people give up individual car ownership, we’ll be able to build fewer parking spaces, leading to higher density and less expensive housing.

So that’s the heaven scenario. And the hell scenario? That’s the one where current ownership models don’t change at all, and instead of two cars in every garage we see… Two autonomous cars in every garage. This is the model that Tesla is aiming to see, and one that will destroy our cities (no, seriously. Not an exaggeration.) Why would this model be so bad? A few reasons, but the first of them is how cool the cars would be:

Would you mind a commute that lasted an extra half hour if instead of focusing behind the wheel you could read, watch a movie, or get work done? Probably not. With driving, as with most things, there are financial and non-financial costs. Autonomous vehicles would obliterate the non-financial costs of driving as your car became your second living room. This increased tolerance of driving would lead to even longer commutes, meaning more community-destroying sprawl, more resource use, and more isolation as drivers spend more time alone in their cars, even if they are entertained.

The second major problem with individualized ownership of autonomous cars?

https://twitter.com/BrooklynSpoke/status/867835383064788992

TRAFFIC. This is in part because of the reduced sensitivity to driving; if you’re watching House of Cards in the back seat, do you really care whether or not you’re stuck in this traffic jam? Since people don’t mind traffic, they’ll be less likely to drive at off-peak times. The bigger contributor to traffic, though, will be empty cars driving themselves all over the place. For the folks who read this blog and are unlucky enough to be forced to drive to work each day: how much do you pay for parking? If you drive into Manhattan (I don’t anyone making that decision is smart enough to read, so we’re safe there), you’re paying roughly $60/hour plus that sweet 18% tax. If you think drivers won’t tell their cars to circle the block while they’re working or to drive home and pick them up later, you’re a literal crazy person. All those extra empty trips will add up FAST.

So what do we need to do? We need to find creative ways to encourage the first, shared ownership model and discourage private ownership of autonomous vehicles. This will largely take place through the tax system: above all, we’ll need to find ways to make letting your car drive around empty so ungodly expensive that no one ever attempts it. To treat the streets as your moving parking spot will immediately make them unable to move people or goods. We may also need to artificially increase the financial cost of driving; as the non-financial costs go down, the public sector will need to make sure that the pain of driving is not so low that people don’t mind doubling or tripling their commutes.

If, however, the first model can be brought about, it could be a boon for close-in suburbs and help make commuter rail more feasible. Autonomous vehicles won’t replace transit, walking, or biking in dense areas; the laws of geometry will still apply, and autonomous vehicles will still take up large amounts of space per person. But if we can move users’ experience of transportation cost away from high fixed cost/low marginal cost to all marginal cost, they’ll inevitably use less of it, enjoy a safer transportation network, and support denser development.