mini profile photo

Scott J. Forman
@sjforman
facebook icon twitter icon instagram icon linkedin icon

Autonomous Cars and the Future of Traffic

09 Mar 2016

As a I child of LA, I remember being outraged by traffic. Staring out the window at the endless river of concrete and steel and glass. Why couldn’t we all just go faster? Years later, my first economics class was a revelation, because I felt I finally understood. It had almost nothing to do with the way people drive, and everything to do with prices. We’ve set the price at zero, so the demand vastly exceeds the supply.

More recently, thanks to Jeff Speck (Walkable City, which is an absolutely essential read if you’re interested in how to…er…make our cities great again), I encountered the idea of induced demand, which explains why endless road-widening projects never alleviate congestion. Because these projects boost the supply of space on the roads. But they do nothing…worse than nothing, actually…about the demand.

Autonomous cars to the rescue, right? When they become commonplace, there should be far fewer crashes. And they will be able to drive very fast, smoothly, and close together:

autonomous intersection

Voilà - the end of traffic.

It’s a mirage. Because increasing the carrying capacity of our roads is, in effect, exactly the same as widening them. And we’ve seen where that leads. No question, autonomy will increase the capacity of our roads - I would guess it may even double or triple it.

Will that be enough to compensate for the increase in travel generated by the greater convenience of being driven without driving? And the latent demand that will surface when the roads are temporarily clear? No way. There will be a tsunami of new traffic that will overwhelm the new capacity, and we’ll be right back where we started.

Ask 100 economists what we should do about it, and they’ll all tell you the same thing: set a price. It really works. Look at the variably-priced toll section of I95 in Florida, and the London congestion charge. But it’s always a battle. Evidently, tolls are the only thing we hate more than traffic.

Which is entirely understandable. We’re accustomed to the idea that the use of the roads is “free.” Even if their construction and maintenance are anything but free. And even though we do pay in wasted time and fuel, in health consequences, and in damage to the climate.

So are we doomed? Maybe. Maybe not. Two weeks ago I described a possible virtuous circle in urban areas from reclaiming space currently dedicated to parking. Last week I articulated a second reason for optimism: autonomous vehicles will tilt the world towards sharing, and sharing makes the price of driving more visible. But this is a weak effect at best.

To head off disaster, we’re going to have to do something more. We’re going to have to put a price on use of the roads. All of them, all the time.

Now, before you scoff “fat chance,” let me say that I’m not naive about this. I know it’s a radical notion, and I know it’s not going to happen fast. But I think it’s worth contemplating, because we’re going to have to do something to avoid a traffic death spiral. It might as well be the right thing.

Here’s what I propose:

  • Cars already “know” how far they drive - and they’re fast becoming connected to the internet.
  • States - and maybe eventually the Federal government, replacing the gas tax - should administer programs whereby, to keep a car or motorcycle registration current, you have to pay for the miles you travel.1
  • If you own a vehicle, it would report back how far it has traveled, and you’d receive and pay a bill. If you rely on shared vehicles, the fee would be paid by the owner of the fleet and incorporated into the price you pay for each ride.

There’s actually a pretty good precedent for this kind of thing - it’s called Metromile - pay-per-mile insurance. A friend of mine has it. Here’s what he sees when he opens the app on his phone:

metromile dashboard

Where does the revenue go? It should be dedicated to mobility. Not solely to road projects, but also to public transit, bicycle and pedestrian infrastructure, and the like - the alternatives that make it possible for people to choose not to drive (or be driven) if they choose not to.

So we’ve got a road-pricing scheme where motor vehicles report their miles traveled and their owners pay a fee for each one. We’re dedicating the revenue to mobility projects aimed at partly at improving2 the roads themselves, but also to alternatives to driving.

It’s only two short hops further to address our traffic problems…and do a lot of good for the climate while we’re at it:

  • Prices should vary by vehicle. The more efficient the vehicle, the less the per-mile cost.3 This would encourage people to purchase (and hire) the lightest vehicles that serve their purposes. Which in turn encourages sharing, because you’re not going to want to pay the SUV price for your solo daily commute, but you’re also not going to want to own all the different size permutations of vehicle that you might need.
  • Prices should vary by time and place, with higher prices for the use of heavily-trafficked segments. A congestion charge, set at whatever price is necessary to keep traffic moving freely. Implemented in software, with no physical infrastructure needed. Intelligent routing systems could trade-off between time and cost. If you’re in a hurry, you’ll take the fastest route. If you have time to spare, you’ll take the cheapest one.

What about the privacy implications? It’s a legitimate question. But it’s worth noting that we already live in a world where private companies like Uber and Lyft know exactly who you are and where you go. Still, the concern is (properly) heightened with government agencies are in the loop. I think it’s a surmountable problem. I suspect a system could be designed where congestion fees are paid, but without the locations of specific vehicles ever being revealed.4

How would the price be set? It should internalize the full social cost of each mile driven, including the public health and climate effects of emissions, and the motorist’s fair-share of the cost of road construction and maintenance. This would be a large number. Maybe we’ll get there eventually. But we don’t need to start there, and even a fraction of it would begin to tip the scales against the traffic explosion. And perhaps even begin to reverse the catastrophic sprawl that cars have wrought.

By implementing a system like this, we can address traffic, climate change, and infrastructure funding all at once. There’s a real risk that autonomous driving will make all three of those problems worse. But it doesn’t have to. If we seize the opportunity and make wise policy decisions, we can turn the advent of this radical new technology into a catalyst to finally solve them.

  1. The long-standing Federal restriction states applying tolls to the interstates will need to be lifted. 

  2. “Improving the roads,” in many cases, should actually mean narrowing them. Here, I’m arguing that we should price the use of the roads to discourage ever more driving and avoid a congestion nightmare. But we can do better - we can reclaim much of that space for bicycle paths, wider sidewalks, parks, trees, etc. 

  3. I would propose that prices vary as a function of three factors: weight, size, and emissions. 

  4. Perhaps there’s a blockchain-based solution. If you think that sounds interesting and you’re blockchain-savvy, send me a note - I’d love to chat.