
Conventional wisdom seems to be that self-driving cars will do a lot to alleviate congestion. They will be able to interact with each other and the environment so that they can space themselves closer together, detect and reroute around congestion, and not cause crashes (which accounts for about a quarter of the existing congestion). However, there are some factors that suggest that they may not completely solve the problem.
First a comment about the “rental vs. ownership” model. There are a lot of predictions that self-driving cars will act more like taxicabs, where you don’t own one, but punch a button on your smartphone and one shows up at your door.
Personally I’m skeptical.
1) There’s the whole “pride of ownership” thing, and being attached to your car. Of course maybe this will eventually die out with me and the rest of Gen-X.
2)What if the person that rode right before me has Ebola? Or a bad cold? And what’s that mystery stain?
3)What if I buy a couple of 50 pound bags of dog food and want to leave them in the trunk while I’m at work? Or if you just keep your bicycle in the back, as I do.
You can laugh at these scenarios, but people do think that way. There’s no strong rational reason for “rail bias” either, but it exists. There are already alternatives to owning your own car: taxis, ride-sharing, car-sharing, car rental, and transit. And there are drawbacks to each of these. You’d think they’d be more popular if people really wanted to not own cars.
It’s also possible there might be a hybrid model, where a family owns the primary car but rents one as needed, or rents out their own car when they’re not using it.
But now lets look at the big reason why they might not fix congestion…
Self -Driving Cars will likely substantially increase Vehicular Miles Traveled.
People routinely mention that VMT is rising a lot slower in days gone by, or is even decreasing on a per capita basis. But that may not hold true in the future for a number of reasons. And with increased VMT (assuming highway expansion can’t keep up), comes increased congestion.
Here are some of the reasons:
1) Moving car storage around is going to add more miles
It seems a dream of urbanists to get rid of downtown parking garages. With self-driving cars, the cars could store themselves in an out-of-the-way location.
Fair enough, but that’s going to add miles traveled. And how close to downtown would massive car storage be palatable? They’re already trying to move the impound lot. Who’s going to want this facility in a developed area?
What’s more, suppose a downtown worker decides to send a car all the way back to Lake Elmo for a family member to use during the day? Or just because they doesn’t want to pay to park downtown? You’ve just doubled the miles traveled in a day.
It’s also possible you’d want to use your garage at home for a shop, rec room, or junk storage, and have your car park somewhere else.
2) People may choose to own more cars
Take a look at this 1950s Levittown ad. Space for one car. And then this typical Shakopee house. Notice how much more space for cars there is.


Right now in many cases we’ve reached a limit with one car per licensed driver. But what if you don’t need to be a licensed driver any more? Suppose Junior has a car to get to elementary school. Grandma has a car to get to her quilting club.
Alternatively, if the rental model comes to pass and people don’t own cars or own fewer cars, there’s still going to be a lot of these extra trips, as well as driving from one fare to another.
3) People may increasingly choose longer commutes
I know “Millennials are choosing the city,” but not all of them are doing so. And not everyone moving into the area is a Millennial, and it’s possible preferences will change over time.
There’s been a desire for more space by people in the country ever since the resistance to the first anti-growth boundary, the Proclamation of 1763. The 1950s freeways and affordable houses just made possible what had been a long-standing desire (previously limited to streetcar suburbs). A couple of studies for one demographic group do not convince me things are going to to change in the long term. And it’s been decades since we’ve built a substantial number of affordable houses. The house in Shakopee above sells for more than triple what the Levittown house, in inflation adjusted dollars, cost.
The reasons for this suburban desire are complex, and might be worth their own post: zoning issues, builders figuring out they make more money building mansions and stack & packs, and consumers thinking they have to have granite because they saw it on TV, or thinking it’s child abuse for kids to share a bedroom.
But with no new supply of affordable houses, and lots of people wanting them, eventually the price of existing houses is going to rise to unaffordable levels, as it has in San Francisco and other places. For people that want houses, the choices would then be a long commute or settling for multi-family housing.
But what if you could take a nap or surf the internet during the commute? A lot of us already spend several hours a day working at home, or on the computer, or on the phone. What if you could do that all during a commute?
The predicted 103,830 vehicles a day over the St. Croix Crossing in 2067 could actually happen. Maybe they even choose a school for their kids near work so they can have family time in the car. (Or what if even they choose a school a long ways away for their kids since it’s better and transportation is no longer an issue?)

4) There could be more delivery service
I think the idea is awesome that I could punch an order into my touchscreen in my car, and then pull up to Famous Dave’s and have food handed to me. (The kitchen would of course know when my car was due to arrive.) And then actually be able to eat something like ribs, even drinking a beer while in a car… it would be a welcome break from Taco Bell and Burger King.
But if driving is easier and cheaper, there may be more discretionary trips. Want a cheeseburger instead of ramen tonight? Order one, and the car will make a trip between Burger King and your house that you ordinary wouldn’t bother with.
5) The coming of Electric Cars
Even using gasoline, if energy costs remain the same or go up, with the above factors it’s likely VMT is going to go up. But something else is happening simultaneously…

Yes, the electric car. Right now a Leaf is impractical for anything but a second car for most families, and a Tesla is too expensive for most people.
But what if someday you could have an affordable car with a 200 mile range? I don’t know if it will happen or not, but there is a lot of motivation from different parties to try to make it happen. Electric cars are a lot cheaper than gasoline cars to operate, even more so that they don’t pay fuel taxes. Right now society has decided they’re fine with that. But at some point we’ll need to rethink transportation financing. And while we’re at it we might as well make users pay the full cost of the roads, most likely with a mileage based fee.
At $3 a gallon, a 30 mpg gasoline car costs 10 cents / mile in fuel costs, 1.6 cents of which is taxes. An electric car is typically 3.3 cents a mile, of which none is taxes. If you double the taxes to replace funding from general sources and add them to cost of an electric car, it’s still only 6.5 cents a mile, and the gasoline car goes to 11.6 cents. Going further, we could triple the tax and in turn eliminate the tab fees. This makes the heaviest users of the roads pay proportionately, but even then it’s only about 8 cents a mile. Probably still enough to induce a lot of extra travel And to be fair, we should have transit users pay the whole cost of their transportation choices too.
6) Shifting of Trips from Transit
But speaking of transit, remember the ridiculous proposal to just buy transit users cars? Well, that might not so ridiculous any more. Anyone that can physically ride in a bus can ride in a self-driving car. The energy cost would cheaper, so more people could afford it, or the cost or the government to provide it would be less. Or maybe, with a lot fewer reasons for people to take transit, we wouldn’t have to provide services like the Gold Line. We could limit buses and trains to areas inside the beltway, by subsidizing the cost of renting or owing self-driving cars for the poor.
But wouldn’t that be just another form of transit? The lines get blurry!
It’s also interesting to think about what self-driving cars would do to intercity transport. When I’m in Chicago I normally stay at a Holiday Inn Express in the northwestern suburbs. My house is in the southern suburbs here. To go by plane, I would drive to the airport, pay to park, then pay to rent a car in Chicago. What if I could leave at dusk, sleep, and then arrive in the morning with my own car? And with an electric car the cost being less? I might travel a lot more for pizza. But if I’m going to take regular trips of that distance I want something comfortable.
There May be a Focus on Comfort, Rather than Efficiency
We’ve all seen this cute Google Self-Driving car. You could fit a lot of those on I-94, right?

Well, you can buy a conventional car that looks like that today.
There seems to be a complete and utter lack of interest in the US. The Smart Fortwo sold 9264 in 2013, compared to 360,089 Honda Accords. You could make the case that if the rental model of self-driving cars comes to pass that people could take a Smart Fortwo to their job in the city, and use the Ford F-150 for towing the boat to the lake on the weekend. But right now there’s already that market for “2nd cars”. People could have a pickup for the weekends and one daily driver, and a city car for the second daily driver, but they don’t.
Instead, here’s another take on what a self-driving car could look like, the Mercedes concept:

Just a little bit bigger. If people increasingly choose two hour commutes to save on housing costs or to get more space, and no longer needed to drive, they’re going to want more space to be comfortable. So you might see recline-flat seats, big screen TVs, and maybe even RV-like amenities like lavatories and kitchenettes.
A car being too big for people to be comfortable driving is no longer going to be an issue. And the cars will undoubtedly be programmed to increase comfort by gentle acceleration and stops. Everyone who’s been behind a semi truck “gently accelerating” knows what it can do to traffic.
Ultimately I look forward to self-driving cars. I figure they’ll be ready for prime time when it’s about time for me to surrender my “manual car” license, and thus I won’t loose my freedom. But I don’t think they’ll won’t solve congestion. They’ll just make it matter less.
Great job Monte!
I think you’re right that autonomous cars will lead to more vehicle miles traveled, because people will be sending their vehicle off on their own, but I think it’s ambiguous as to what that will mean given the efficiency gains involved.
That said, I have a hard time imagining them leads to owning more cars. First, there’s the hole taxi-like thing, which might make it easier not to actually own one (while I agree, this will not be anywhere near universal). Then there’s the fact that your car won’t be stuck at the office with you, so it can go get grandma and Junior. Heck, even married couples might be able to stagger their commutes so they can both get by with one car. So, yeah, my guess is yet fewer cars per person, not more.
As to longer commutes, it really depends on whether we decide to continue to subsidize them.
You’d think I’d proof-read before hitting submit… Man.
Excellent points. I think you really nailed a huge problem with how much these might potentially or are likely to increase VMT.
BTW, really affordable might be a few years out but I do think we’ll see a half dozen cars with 200+ mile range and below $35k price tags by 2017; Nissan Leaf, BMW i3, Tesla Model 3, Smart, and others.
I agree he’s right on that they may increase VMT but I don’t think he considers it a problem.
I think you are on to something with an increase in vehicle miles. Travel will be much easier, in a sense “cheaper” and there will be more of it. So long as the vehicles are electric and charged with greenhouse free electricity, I’m not sure it much matters.
I do think there will be higher utilization of pool or fleet vehicles. You can get exactly the car you need for the trip, without the expense of owning. Kind of car timeshare. Just you going to work – request the single passenger economy car. Family of four going out to the movies? Get the economy for four vehicle. Vacation time? The family truckster has been reserved for two months and arrives at your driveway the night before for packing and a late night nap-time departure.
Hauling a bunch of crap? The F-150 shows up ready for loading and automatic departure.
The impact on mass transit will be interesting. People are the largest expense of Metro Transit, so mass transit will become a lot less expensive to operate.
Interesting times to come, so long assuming we don’t cook the planet so quickly that none of this ever happens anyway.
I’m not sure I can buy in to your skepticism regarding the ownership model of AVs. I don’t think people love their cars as much as they love the freedom their cars grant them. The freedom to live wherever they want and still get access to other places in a region – that’s liberating! The convenience of being taken from almost any point A to almost any point B is what cars currently offer, and the dangerous task of driving is merely a means to an end.
Currently if you want that service, you need to pay the price. Upfront cost, gas, maintenance, a place to park, and whether one can afford to give their undivided attention to the two tons of steel that you have sent careening down the road all factor into whether someone can afford the perceived freedom a car grants. When I got rid of my car over 10 years ago, it was because the cost of the car outweighed the benefits I was receiving from it. I had to move it every few days if I wasn’t using it or if there was a snow emergency, plus my job was on a bus line so I didn’t even want to drive since I could relax and read a book on the bus instead (and pay to park in a lot, of course).
Maybe just the idea of having an autonomous vehicle parked strikes me as absurd. The hybrid model of someone owning a car and renting it out as a “taxi” is more plausible, as there’s no incentive to just leave a high-value piece of capital just sitting unused on the curb.
Your middle paragraph undermines the first.
That is, we have been successfully marketed into a culture that views cars as freedom that’s pretty significantly at odds with the reality. Sure, we can get between any two points, but that’s only one aspect of convenience. When I drive, I fairly regularly decide against making a second stop because parking and navigating crowded parking lots seems like a pain.
Actual freedom to me is, “hey, yeah, I could just walk or ride over there and I don’t have to worry about what I’m going to do with my car.”
Anyway, sorry for the tangent.
We’ll have to see whether people want to share their high-value piece of capital with strangers. I’m with Monte on being skeptical.
I think that difference in the perception of “cars as freedom” has more to do with where someone is physically located. A car provides a significant amount of freedom for someone who lives in rural MN compared with someone who lives in the city. Cars do provide a certain degree of freedom that many who live in urban areas don’t need. Sometimes I want to get somewhere that’s 100 miles away and can’t use transit or a bike. I don’t think that means I need to own a car, just that I need some kind of service to get out there and buying a car isn’t an option (in this case, I’d probably just rent for a day).
The market for business models such as Car2Go and Uber are somewhat useful when speculating on the ownership potential for AVs. I’m not sure either of those businesses is capable of reaching the suburbs and providing the level of service that we expect in the city (or at least, portions of the city…). And I think car ownership will always be higher in less dense places, but that in dense places, it is clearly unnecessary. And people are moving away from ownership across American society anyway, often preferring to rent and let someone else deal with the burdens of property ownership (and keeping flexibility – another important form of independence).
Right. Cars mean freedom to get to places that are far away. Heck, cars and the infrastructure we’ve built for them make it easiest to get to stuff that’s far away. But what’s wrong with going to stuff that’s actually close to you?
You’re absolutely stranded if you’re in the suburbs or farther and there is nothing close to you.
Which is one reason I don’t particularly enjoy going to the lake anymore. Once you’re there, you can’t do anything or go anywhere without getting in the car (or boat).
“3)What if I buy a couple of 50 pound bags of dog food and want to leave them in the trunk while I’m at work? Or if you just keep your bicycle in the back, as I do.”
Well for purchasing large items you could do what I and plenty of others who don’t own cars do and just get it delivered. The price of getting stuff delivered to your door has been falling for a while to the point where you can get all sorts of things delivered for cheap or free in places that aren’t NYC these days. Amazon has some questionable business practices, but now that people are used to the idea of clicking on something online and having it show up in a reasonable amount of time for cheap or free, you can’t really put that genie back in the bottle.
And actually regarding your other item about trips for things like food, I’d much rather have food come to my home than eat it in a moving vehicle. These days you can get a ton of things delivered by third party services that handle delivery for places that wouldn’t normally do delivery (bite squad, door dash, etc). I imagine they might be able to streamline their business model by utilizing the same autonomous vehicles to the point where delivery on demand becomes even cheaper. At that point I don’t really see the advantage of getting in your own self-driving car to make the same trip when the cost difference is pretty negligible (since you’re still paying for your own gas and spending time away from the comforts of home).
I could even see something happening where you basically have a mobile automat that doesn’t even require a person other than to load it at the store/restaurant/whatever and delivers things to you at home or work or where ever.
My daughter participated in the community theater production this summer and at my house we spent much time fantasizing about self driving cars so we could avoid those pickups and drop offs. In my dream scenario, the time and place of the events would be sent to the self driving carpool van along with the addresses of the kids participating. The self driving van would arrange its own carpool route to pick up and drop off the kids. I am definitely not interested in the burden of owning another vehicle. And I definitely don’t want to spend more time in the car. For me, the benefit of self driving cars would be improved safety and less time spent driving (or riding) for me.
Very nice read, thank you. It isn’t often enough that I read something which gets this far into the unintended consequences of change.
Nice analysis. I don’t actually believe that self-driving cars will ever be practical in rural areas, due to what I call the “deer detection problem” (and people in less deer-infested areas calls the “child detection problem”).
In urban areas… well, they will make congestion worse. So…
If they can’t detect random obstructions or unpredictable movements, self-driving cars will never be viable.
But those actually sound like problems that can be solved, and things that the computer running the car will actually be better at handling than a person, eventually.