Self-Driving Cars Are Going to Make Congestion Worse — But Autonomous Minibuses May Save the Day

Autonomous vehicles have the potential to dramatically change how we get around. However, there is currently a vision that self-driving cars are going to solve our traffic woes and reshape where we live in the process. I think that vision is mostly wrong. Self-driving cars will definitely influence our cities, just not in the way a lot of people think. If anything, rather than solving our traffic woes, these cars will make things worse, accelerate trends already underway, and lead to cars losing both influence and real estate on the road over time.

The real game changer arguably isn’t the car that can drive itself, it is the minibus that is much cheaper to operate (because it can drive itself). How does this play out? Both as a transit solution and as an inter-city travel solution.

The Paradox of Congestion

At first blush, the idea that autonomous cars will eliminate congestion and reshape the boundaries of cities sounds simple and attractive: with autonomous cars, the geographic gaps between areas will effectively be decreased because people can be productive or relax in the cars. People can live wherever and eat, work, and sleep in their cars as they are whisked to their destination. This vision has a lot in common with electricity that is too cheap to meter and flying cars: theoretically awesome, and not likely to occur at large scale any time soon.

The main reason why people are not going to shift aggressively to living in further-flung areas and having cars whisk them to work is congestion. Roads are already congested; as cars become cheaper to maintain (through electrification) and operate (through autonomy) this congestion will increase. Autonomous cars make being stuck in traffic less unpleasant, so people may be willing to hang out in traffic longer than when they are driving themselves, but I don’t think this will significantly impact most folks’ calculus. It doesn’t matter if you can sleep or answer emails — most people still won’t want to commute two hours each way every day.

After all, if it was as cheap as public transit to take an autonomous vehicle to work and even more convenient, wouldn’t everyone want to do that? Yes, generally, they would, and it would be a mess. For example, Chicago averages about 1.7 million bus and el rides a day — call that about 500,000 people (round-trip, with some transfers). If you tried to put most of those people in their own cars, we’d have a pretty bad time.

In the urban and suburban areas where most people in the United States live, the gift of autonomous cars to these areas is likely to significantly increase congestion and travel times in the nearer term. It will be cheaper and simpler to get from one place to another, but because of this cheap simplicity, people will take more trips and increase congestion on our roads. As a result, living 40 miles away from your office is unlikely to become a great deal more appealing. In general, I believe the increased congestion from the proliferation of cars will more than offset any growth in long-distance commuting.

The Game Changer: Autonomous Minibuses

I think the real game changer here isn’t self-driving cars themselves — it’s the really cheap and frequent shuttle service with autonomous minibusses (in dedicated lanes) that autonomy enables, with heftier versions of those already being produced by Navya and Baidu today. We have a limited amount of road space; I believe it will be allocated towards solutions that can move people efficiently — probably autonomous minibusses, but maybe some regular old city transit buses too (and in the further future, perhaps single-occupancy vehicles as well). Think of it as a High Occupancy Vehicle lane on steroids. Indeed, this does not necessarily sound so different from shifts taking place today, where traffic lanes used by cars are being switched over to bike (and scooter) lanes, dedicated bus lanes, and transit stops, and creating more walkable and livable cities in the process

In addition to bolstering local transit, autonomous minibusses will also improve travel between cities as inter-city shuttles, especially those not served by public transit. Trying to get from South Bend to Cincinnati, or for that matter traveling between large cities like Chicago and Indianapolis? These trips are a pain without a car today, but with lower operating costs, it will be more practical to run fleets of vehicles to provide transportation between cities that are right-sized for the customer demand. Plus, with the proliferation of autonomous vehicles and last-mile solutions, you won’t need a car once you get there, making the shuttle between cities much more practical.

I think all of these use cases will actually push in one direction: they will decrease the amount of real estate on the road for cars (whether driven by humans or computers) and increase the amount of space for other solutions that enable greater density of travelers (whether they are autonomous solutions or not). By leveraging the road infrastructure that exists with autonomous vehicles that are right-sized for the ridership and cheaper to operate, we will be able to move people around much more efficiently. Indeed, in order to prevent crippling gridlock, autonomous minibusses in dedicated lanes may be the solution that autonomous vehicles practically require.

This won’t happen on its own of course — we will need forward-looking and active city planning and changes to how we think about our road infrastructure to avoid autonomous gridlock hell. Dedicated lanes and congestion pricing are good places to start.

The Future

Personally, I’m looking forward to autonomous vehicles. I think they will be great for safety and for public transit. And, I think they will help cities continue to get denser and operate more efficiently. At the end of the day, autonomous vehicles will encourage, rather than discourage, the growth and densification of urban areas — a trend that is already underway today, with or without autonomous vehicles.

So, while I don’t think we’re going to have a huge proliferation of self-driving cars, I do think we are going to have a lot of autonomous minibusses that will change how we live in cities.

Thanks to my colleague Ben Gaddy for his helpful recommendations on this post.