A friend came over to my house the other day driving a brand new Tesla. It was a sudden reminder that all this automotive innovation that we keep hearing about is actually happening. Just a few days previous, I had been watching this demonstration of Tesla’s new “autopilot” feature:
I have to say that, as someone who hates driving, watching the video below got me pretty excited.
I can’t recall offhand any piece of technology with the prospect of more dramatically improving my quality of life.
On the other hand, driverless cars have the potential to exacerbate a whole lot of other problems. The first is that, unless we make some changes in the way that access to roads is regulated, they will massively increase congestion. The economics of it are pretty simple. You can pay for roads in one of two ways, either with money or with your time. If use of roads is not priced, then there will be overconsumption, which will manifest itself in the form of congestion. What limit is there to how congested roads become? It’s all a matter of how much time you have to waste to get somewhere, and perhaps more importantly, what the opportunity cost of that time is.
To give an example, the other day I was looking to buy a drone (don’t ask). I found a shop in town the sells them, unfortunately it was up in Vaughan, which from where I am requires a trip up the Don Valley Parkway (the DVP, or as it’s know locally, the Don Valley Parking Lot). That means at least an hour in the car each way. For me that’s a deal breaker, so I mail-ordered it from the U.S. instead. As a result, there was a bit less congestion on the DVP, because I decided not to take a trip.
Think now what happens when the driverless car appears. Right away, the opportunity cost of sitting in traffic declines precipitously. Looking at the Google car above, it’s obvious that you’d be able to bring your laptop along and get some work done while you’re in the car. I probably would have taken that trip up the DVP, if it hadn’t required giving up 2 hours of work time.
Also, while the first generation of driveless cars will undoubtedly still require there to be a driver in them, as more cars become driverless, and therefore traffic becomes less dangerous, it’s only a matter of time before people start sending the cars out empty. Why not have your car drive you to work, then turn around and go home again? Or if you have to run a quick errand, why not have your car circle the block until you’re done? Or why call a cab, just tell your car to come meet you at a specific location and time. All of this can be expected to again increase congestion – because the cost of having your car sitting empty in traffic is practically zero (especially if it’s electric, and so using practically no power in stop-and-start traffic).
The only way I can see of making this manageable is to have a completely comprehensive system of congestion charges. So it seems to me there’s no point debating whether or not to introduce road pricing – we will have no choice in the matter. The only question is how to implement it – which strikes me as something worth thinking about sooner, rather than later. For example, the system would have to be so comprehensive that it raises questions about whether it’s worthwhile investing in existing technology, which relies on entry points and toll cameras. The classic highway toll or perimeter toll is not going to come close to dealing with the problem. An alternative is something that the insurance industry has been talking about, which is to use the GPS in cars as a basis for imposing per-kilometer charges. (Insurers have been interested in the idea as a way of developing more sophisticated risk-classification of drivers, based on the actual number of kilometres they drive. The idea would be to have, instead of a yearly rate for insurance, a much lower flat rate to cover theft, plus a per-kilometre charge to cover accident risk.)
Thinking about it from a “nudge” perspective, the ideal would be to have a big dashboard display, just like a taxi meter, showing how much the drive is costing you. It could add up fuel consumption, insurance rates and congestion charges, calculate how much it’s costing you to drive as you move along, and give you a real-time display.
One fun alternative is the idea that all the cars could communicate with one another and a central planner, and then the central planner takes this data and constantly optimizes the routes the cars are taking to their destinations. And maybe this reduces congestion.
In any case, the effects of AI and how legislatures respond will be very exciting to see.
@James…I get your point but your use of the the term “Central Planner” in relationship to an all seeing, all knowing GPS tracking system sounds like something Soviet-era secret police agencies would have come up with had they lasted long enough…of course I’m sure the NSA and CSIS are salivating at the prospect. This likelihood is why at some point I will have to move off the grid, head for the bush with a cross bow and two metric tonnes of Lipton Cup-a-soup — and my tinfoil hat — and wait for the fireworks to start. Of course Joseph would be able to track me down with his mail order drone, so I guess I’m screwed either way.
Hi. This comment has nothing to due with road pricing. I just wanted to say that I follow you on academia.edu and have really appreciated the uploads from the last few weeks. CBA and liberal neutrality, and recent on “just wages”. I am also a big fan of your two puzzles of contractarianism. And the chapter you linked to on this blog about “ideology” as misdiagnosed collective action failures was very insightful. I hope this encourages you to upload more of the same. The more the better, even the older stuff!
An alternate perspective on driverless cars is small scale transit — particularly if the vehicles are owned by some collective rather than individually. Remember, the suburban dream was one of cheap energy, cheap land, unlimited open space. Two hours to get to work was just a side effect. One downside is that we are exchanging the high volume low maintenance capabilities of rail for the reverse on roads — roads take a larger slice of the country than rail. But its all moot if we turn the countryside into a big powerplant, continue to rely on distant lands for food and manufacturing. Question then becomes — why do we exist of we don’t do anything useful?
Congestion is one thing that proponents of other schemes seem to forget about, and I think there is a valid point even with driverless cars. This is why I have always found it amusing that bicycle proponents seem to think that everyone riding bikes is the answer to all core traffic issues, but fail to (or willfully ignore) the obvious, that bikes are still personal transportation and ratcheting up their use comes with all of the same problems that entails, like parking.