How will shared autonomous electric vehicles change our cities? A Troppo challenge

Driverless cars

Artist’s incorrect impression, from the film “Minority Report”. In the real future, these autonomous cars would be travelling much closer together, and there would be more of them.

This month’s print and online editions of The CEO Magazine include a piece by me on carmakers’ concerns about SAEVs – shared autonomous electric vehicles.

Short version: the carmakers all believe that SAEVs will happen, and perhaps quite soon, probably sometime in the 2020s. And they fear the arrival of SAEVs may essentially destroy them, by turning them into commodity producers of hardware for “mobility services firms”. In this world, you won’t own a car; you’ll just tap the mobility firm’s app whenever you need to go somewhere.

This will puts a huge hole in the business model of “The Ultimate Driving Machine” and all those other car brands:

Are the industry leaders scared? “Oh, yeah,” says Sam Abuelsamid, co-author of a major report on the evolution of SAEVs published in April by technology research firm Navigant Research. “The industry has been doing business in the same way for more than 100 years. This is a huge transformation … everybody’s terrified.”

Writing the piece was a lot of fun: you don’t often get to write up an industry contemplating its imminent extinction. These guys have been thinking for a long time about how SAEVs will change their world.

SAEVs and the city

Thinking about the full range of second-order effects of SAEVs is limited by the reality that we really don’t see any self-driving cars yet.

But one obvious key to all this is the Jevons effect, named after the world’s first great resource economist, William Stanley Jevons. He noted in 1865 that technological improvements that burnt coal more efficiently did not cut coal consumption. “The very contrary is the truth,” he declared; cheaper coal energy meant more coal-burning. He was, of course, exactly right. Our entire modern world is shaped (as they say in the pop history business) by the Jevons effect.

When we think about the effects of SAEVs, we can say that they do several things. But all of those things together add up to a huge fall in the monetary and time costs of travel:

  • SAEVs will be far cheaper than the car in your garage: they will have less moving parts, lower insurance premiums, lower energy costs and far, far higher utilisation rates.
  • They will put you in a private cocoon where you can work or relax, combining the best features of cars and public transport. My guess is that a lot of SAEVs will come with big screens and surround sound.
  • By enabling cars to be packed together in giant convoys, they will reduce congestion.
  • Because they will be so much safer at a given speed, speed limits will be raised – probably first on certain freeway lanes, where they will be separated from the human-driven traffic.

And the Jevons effect predicts that cheaper travel – easier on the hip pocket, quicker, and nicer for you – will mean more travel.

Or looking at it another way, SAEVs will do much the same thing that building freeways will do: attract more traffic onto the roads, except at a higher average speed than we’re used to today. A corollary of this is that people will take more, longer trips.

And particularly, people will be willing to accept living further away from their work, because the commute will still take no more time than they’re spending on it now.

What half an hour gets you

If you doubt this could happen, be aware that there’s a thing in the world of urban transport called Marchetti’s constant, which is the average amount of time people reserve for their daily commute. It’s about an hour, or half an hour each way. Marchetti’s original paper, which still seems to be held in some esteem, argues that this has dictated the shape of cities back to Athenian times.

If all of this is right, SAEVs could well give us cities sprawling on a scale we have never seen before. Melbourne’s boundaries could extend to Bendigo, and Sydney could become one continuous city from Newcastle to Nowra. After all, for most of the city’s history, most people have wanted more space.

This is just one possible and fairly obvious second-order effect of the move to SAEVs. There are dozens of others. My favorite so far is Benedict Evans’ suggestion that SAEVs could lower smoking rates: a substantial proportion of cigarettes are purchased at service stations as an impulse buy, and that won’t happen when no-one stops to get petrol any more.

I’m inviting Tropodillians’ speculation on possible holes in my scenario above, and on ideas for other second-order and even third-order effects. For instance, will public transport survive? I’m starting to think not. And if it does, how long can the unions hold out against the removal of train drivers?

Extra: A few reports.

About David Walker

David Walker runs editorial consultancy Shorewalker DMS (, editing and advising business and government on reports and other editorial content. Newsletter: . Among other roles, David has edited the award-winning Acuity and INTHEBLACK magazines, been chief operating officer of online publisher WorkDay Media, held senior policy and communications roles at the Committee for Economic Development of Australia and the Business Council of Australia and run the website for online finance start-up eChoice. He is a former economics writer for The Age and News Ltd. He has qualifications in law and corporate finance.
This entry was posted in Economics and public policy, Uncategorized. Bookmark the permalink.
Notify of

Newest Most Voted
Inline Feedbacks
View all comments
paul frijters
paul frijters
6 years ago

It’s hard to see SAEVs making much difference to travel times in most of the major cities we have. What you describe sounds awfully close to a metro/train. And surely many of the same things that slow traffic now will continue with SAEVs: roads that are dug up, other SAEVs breaking down, the need for ‘something’ to unload in particular places or ‘wait for another passenger’. If the SAEVs are going to retain the flexibility of cars and trucks, then the same disadvantages will limit the speed improvement, particularly in cities.

And if you are going to kill the key flexibility aspects of a car/truck, which is that it can pick up anywhere and deliver anywhere whilst allowing for the inability of anyone to be truly on time all the time, then you’re back in the world of trains and metros: particular stations where you hop on and hop off.

Now, outside of cities, what you describe doesnt sound quicker than bullet trains (which cost a lot of electricity to get up to speed! Which in turn means the costs of being flexible about where to stop would be costly) and certainly slower than airplanes, so I can’t see it have much effect on the urban sprawl.

And don’t forget that there is a transition period in which these SAEVs are going to have to share with other road users like bikes and pedestrians who will not keep to traffic rules, which means speed limits will be where they are now (if not lower) until these SAEVs have dedicated lanes (like train tracks).

My guess hence? Pod-like SAEVs for travel to and from the hubs located near work and home where space will have to be shared with walkers, bikes and others. Public transport between those hubs. Some SAEVs and other vehicles remaining on what we now think of as the main arteries in and out (something like a truck is going to remain. Those shops need loading in and out, kids needs to be dropped off, and garbage needs to be collected too). The pods will effectively lead to the demise of chauffeurs and car drivers near home and work (perhaps starting with exorbitant insurance costs if they are not SAEVs, leading to a legal ban).

Will it make a lot of difference? Nah, because the speed increase is minimal for the saevs and you in fact will need more of them since you need one at either end of the public transport hubs, some of which will be privately owned and some not (a bit like the shared bike system you have now).

To make a big difference in speed and hence urban design, one would have to take out the autonomous human element completely on the door-to-door infrastructure. No more walking , cycling, front-door supply, kids dropping, etc. That requires much more than just an saev-revolution though. Can’t see it happen in a hurry.

What is then the main effect? Fewer accidents, a few more coal fired power stations to generate the electricity, a few more large underground parking lots for these things.

Moz of Yarramulla
Moz of Yarramulla
6 years ago
Reply to  paul frijters

David, can I just point you to a famous image that I hope you’ve seen before. It shows a busload of people in a bus, next to the same number of people on bicycles, next to the same number of people in cars. Some of the car people are hard to see because they’re a long way away.

The point is that if you want more people on the road the solution is to put more people in each vehicle or make the vehicles smaller.

David Walker
6 years ago

Thanks Moz. As a veteran public transport user, I’ll simply share the sentence written by one of the commenters at the link:

“Perhaps they can also figure out a way of graphically displaying how it takes 2 hours to get by public transport to somewhere you can drive to in 17 minutes.”

paul frijters
paul frijters
6 years ago
Reply to  David Walker

I don’t see how these scenarios are going to have much impact on commute times.
“certain roads start to reserve their middle two lanes for SAEVs, with higher speed limits. This would happen first on the M1 and the M31, and then spread.”

this sounds like the German Autobahn which has multiple lanes with different (effective) speed limits. Trucks to the left, fast cars to the right, mediums in the middle. Its cool for people with a Ferrari to get a thrill now and then, but net time saving is minimal when it concerns commutes: the traffic jam leaves little spare capacity and the presence of cars that move between lanes really reduces how fast the fast lanes go. In and out lanes will have to remain to retain the flexibility of a unique starting point (home) with a unique end-point (work).

In fact, you see the opposite movement in London, New York, and some of these other very dense cities: the city centre is becoming relatively car-free. There is just not the space to park, so you have taxis and city-bikes ferrying people inbetween a transport hub and the final destination. Which is the model I describe above. As long as you want to retain the tripple flexibility (from any A to any B at any Time), you are going to park those saevs at home and at work, which is just not going to happen in the big cities because space is too expensive. So either the flexibility in space has to go (which gets you to the hub model) or the flexibility in time has to go (your personal vehicle arrives ‘in 10 minutes’). And that time flexibility is worth a lot, particularly since the other road users are going to mess up time-schedules.

So I just dont see that ‘any A to any B’ model for home-work transport getting that much faster.
The battery stuff is an irrelevance for this, though in transition they are a step backwards. These charging poles dont take up much space but charging takes too long, killing flexibility again.

Dedicated SAEV routes? Unless you are thinking of doubling the transportation networks, I dont see how that could work: those routes will intersect with other roads and road users unless they are truly stand-alone routes. And even with dedicated routes, you still have to park the things to be flexible.

Moz of Yarramulla
Moz of Yarramulla
6 years ago
Reply to  paul frijters

you see the opposite movement in London, New York, and some of these other very dense cities: the city centre is becoming relatively car-free.

… Sydney is heading that despite the best efforts of the traffic sewer people. Significant chunks of the inner city densification program are reflecting the lack of utility of cars in their design. The Chatswood tower-over-train block, for example, has relatively little private car parking and increasingly what there is is being repurposed to car share schemes. When they run the numbers the cost of car parking gets really ugly, if City of Sydney packs 50,000 extra residents into Redfern/Alexandria then sure, they can make developers pay for 30,000 car parking spaces underneath, but those cars have to be able to get in and out of the area if there’s to be any point in having them. Which means another big “X” of four lane motorways added to … a tunnel, I presume?

John R Walker
6 years ago

Re Dedicated SAEV routes?,
There is a similar problem about running a fast tilt train from Canberra to Sydney on the existing tracks. While it could do 160+ kph from Canberra to Mittagong , for the rest of the trip to Sydney it would be stuck in traffic , would be stuck behind all stations trains doing an average speed of around 60 kph .

Also feel that the problems of mixing autonomous vehicles and human controlled vehicles on the same roads are likely to be much greater and much more unpredictable than most seem to think.

derrida derider
derrida derider
6 years ago

Assuming the technology to make true SAEVs develops (we aint there yet – there’s still some way to go before production cars have no steering wheel), the future could well be as the car makers fear. But like all previous technological revolutions that future will be more transforming but much more delayed than people think. I don’t think car makers need worry about the “autonomous”‘ bit in the 2020s – the 2060s maybe. OTOH the “electric” bit is closer.

It takes an enormous amount of time before physical infrastructure and social organisation is adapted to the technology possibilities – and as often as not its the second of those that is the laggard. Robert Gordon has written at length on this – the extreme example is the fractional horsepower electric motor, which was developed a good fifty years before factories were fully reorganised to exploit it. And that was far less transforming than SAEVs.

6 years ago

This reminds me of the imaginings surrounding other technology that seems prepetually “just around the corner”. I agree with DD that the technological hurdles that remain are significant. The more probable scenarios is that there will be incremental safety improvements for new up-market cars and business as usual for everything else. SAEV’s aren’t revolutionary they are just another form or taxi.

Moz of Yarramulla
Moz of Yarramulla
6 years ago

What I’m seeing is more focus on non-car vehicles. Your approach seems to be a half-step, and some planners and engineers are looking at that and saying “where will we end up if we go that way” and wondering whether the incremental version is worth the effort. Like the leap from type 3 automatic driving to type 5 – lane keeping to fully autonomous, because “it drives but you have to pay attention” doesn’t seem feasible.

You have the “beeping noddy car” thing in London’s Docklands, where it’s somewhere between a pod and a trackless-tram. The autonomous “single car elevated railway” thing in Seattle. Plus a variety of similar things. But at the other end of the process you have the EU light and heavy quadricycle rules, which are being used in cities like Rome to enable car-share fleets of tiny electric two-seaters that fit three or four to a parking space and can be driven with a scooter license.

I expect that at some point we will cross from over-powered electric bikes illegally using bicycle lanes and “electric quadricycles” being used as micro-cars to a bike-lane type conversion process. Streets and zones will get dedicated lanes that are narrower than truck lanes, quite likely half-width so that trucks can still get down them (emergency vehicles, for example). Vehicle rules will change/expand so that smaller motor vehicles are practical (or in Australia, legal) and we will see a contest between smaller-lighter-cheaper vs bigger-shinier-restricted in the cities.

The questions of whether those will be autonomous or not, mass vs single ownership, single scale or single purpose and so on, I don’t think we have answers yet. I’m going to guess that we will see more and more “I don’t own a van, I rent one” as the cost of those keeps rising. It’s only $10k or $20k a year to won a big 4WD or van for the annual trip to the bush or house-moving when it’s a petrol vehicle, but if it’s electric I think it will be more expensive – not least because it’ll be 10 years before they’re available second hand in any numbers. I’d be more likely to buy an ex-DeutschPost electric van than a car, for example, because when I need a motorised vehicle it’s usually because I’m hauling things too big for the bike. But I never have, because even $500 to rent one for a day pales in comparison to the cost of ownership (my driver’s license costs me about $5/km)

Paul Frijters
Paul Frijters
6 years ago

like Moz, I see a lot of future for all kinds of hybrids (half bike, half motor).
I agree with John about the unknown interaction between self-drive and humans though. When I was thinking about that, I was envisioning the traffic lights of London with its swarms of bikes and pedestrians going through red lights and in front of cars. Imagine a self-driving car with software that tells it not to progress if there is something closer than a meter in front of it. It would be the last to leave any traffic situation and clog up everything behind it. Only because there is the hint that a madman is behind the wheel who would drive into those persky bikes, do these bikes not swarm in front of the cars all the time. Self-driving might turn out to be slower and lead to more congestion than not self-driving…..

I did understand what you meant with saev’s, David. I just don’t see it being much of a game changer. Not transformational at all. A bit inbetween the carriage of a metro and a car. Not all technology that appears revolutionary actually is revolutionary.

Moz of Yarramulla
Moz of Yarramulla
6 years ago
Reply to  Paul Frijters

Imagine a self-driving car with software that tells it not to progress if there is something closer than a meter in front of it.

Wasn’t it google who had to reprogram their cars to be “more assertive” because they found that when it obeyed the road rules it got into all sorts of trouble? Apparently motorists don’t just break the rules, the expect all the other motorists to do likewise… there’s also a terrifying moment in this month’s “Dashcam Australia” video where someone lane changes in front of a truck then hits the brakes. Arguably both drivers were obeying the law but it’s the car driver’s sheer good fortune that the truckie was able to change lanes while braking rather than shunting them through the red light.

There is definitely an interesting question “what if everyone obeyed the law” that we might actually be able to test using autonomous vehicles.

paul frijters
paul frijters
6 years ago

“There is definitely an interesting question “what if everyone obeyed the law” that we might actually be able to test using autonomous vehicles.”

Definitely, which is precisely why the issue of whether these saevs have to share the road (or something that replaces the road) with humans or human-driven devices is important. If they have to share, saevs really dont offer much beyond the metro/car packages we have now because the humans wont abide by the rules. If they don’t have to share, that is itself a huge change in urban design and regulation which would have far reaching implications: no more walking in the cities except in traffic-free places, no more outdoors activities that intersect with any form of saev traffic. That changes the nature of cities more than the saev would, I think.