Will Autonomous Cars Destroy Cities As We Know Them?

Recently, I had a debate with a group of friends about the future health of suburbs when autonomous cars arrive (let’s set aside when exactly that might be). The general consensus has been that the suburbs are going to grow rapidly, since commutes into the city (or just going out for a night on the town) will be far safer, efficient, and convenient than today’s status quo of driving a car and having to find a place to park.

I disagree with this view quite strongly.

The decision on where to live isn’t made in a vacuum. In fact, quite the opposite - people spend enormous time choosing where to live and the mix of amenities, convenience, and price they are willing to bear. Apartment searches, along with job searches, are among the canonical examples of search costs in the economics literature, and for good reason.

First, a lot of the analysis I have read on this topic assumes that people will choose suburbs with autonomous cars instead of cities, but they somehow miss the fact that cities will get these vehicles as well. Autonomous cars will usher in incredible conveniences for everyone, regardless of where they live, and I don’t think cities or suburbs are frankly going to beat the other on this point.

Instead, the key factor is going to be economics. There is this assumption that a) traffic will be better with autonomous cars, and therefore b) people will live even further away from cities, because the decay function of convenience with regards to distance is going to degrade much more slowly than it does today.

But driving cars is ridiculously expensive per mile. Take the IRS milage reimbursement rate for businesses, which is based on gas costs, auto mileage efficiency, depreciation, repairs, insurance, among other factors. The rate for 2017 is 53.5 cents. So for a 40 mile one-way trip to the city, the IRS quotes a cost of $21.40. Round-trip will be $42.80.

Now, maybe those costs will come down significantly. Insurance could get cheaper if there are fewer accidents, autonomous cars may be more reliable and need fewer repairs, and lack of ownership will mean that cars will get more efficient since the buyers will be companies instead of individuals. So let’s say costs go down massively — 50%. That means that it will still cost more than $20 to ride into the city. Suddenly, a regular commute into the city doesn’t look so cheap after all.

Even if we do shared rides, we have to assume that there is enough density in rural areasfor automatic carpooling to be effective. I have a hard time seeing that today, and even if it exists, there is a cost in time and money in picking up additional scattered passengers.

The more likely scenario I see is that autonomous cars will completely sever the relationship between people and car ownership, and that this change will fundamentally drive interest in urban living. Today, an enormous percentage of Americans believe that they cannot give up their car, and so urban living arrangements are just not feasible. But if ownership is no longer a concern, then there will suddenly be a large swath of the population that could consider city living for the first time.

More interestingly, I would guess that this dynamic most helps cities like Houston and perhaps Los Angeles which are large urban agglomerations but lack high-quality public transportation. The revolution will be making the urban sprawl more livable against the suburban sprawl it currently competes with.

All of this might be a bit counterintuitive until we start focusing on fundamental human decision-making rules. One of my on-going annoyances has been the absolutely breathless “analysis” of the future of autonomous cars and how they will change cities. I have already complained somewhat bitterly about this before in regards to traffic in cities (short summary: traffic will still exist because pedestrians exist and have to cross city streets, so traffic lights are going to be a permanent fixture of cities). It has never stopped amazing me how shallow most of the analysis on this topic has been and how obviously wrong a lot of it is. But sometimes you get to a more nuanced analytical challenge that’s worth diving into a little more deeply.

I would caution readers that we need to be vigilant about overhyping technologies coming out of Silicon Valley. I get that we want to sell a compelling future, but at some point, we all have to realize that Ethereum and ICOs are not going to cure cancer (I’m waiting for the commenter to tell me I am wrong). Autonomous cars are going to mean huge things for cities and suburbs, but we need to be careful lest we overestimate a revolution in urban planning that doesn’t seem to be in the making.

Photo by J. Triepke used under Creative Commons


Have a comment? Feel free to leave a reply using Discourse. Please try to keep comments civil, or they may have to be edited. ~Danny