4. What will the supporting infrastructure look like in 2045?

• Will the vehicles run on tracks, on the road, or on something else?

• Will the vehicles be steered or guided?

• Will the vehicles be segregated from other forms of transport?

• Will the system be manually controlled, or automated?

• If automated, will it be controlled through a centralised system or will each vehicle be independent
• What else will be different about the infrastructure supporting public transport?

I’m assuming surface transportation is mostly going to be on-road, and the only vehicles that will be on tracks will be high frequency passenger trains, and freight trains.

I’m also guessing that freight train traffic will go down over time – the US in particular because a lot of that traffic is bulk commodities, especially coal and oil, and that is just going to disappear with changes in the energy sector as renewables become increasingly dominant. Tracks will still make sense for other goods shipped in containers and some bulk commodities . Still trucks will be able to do a lot of this, and trucks will start to get the advantages over trains because they just have a wider spatial coverage (they can go everywhere there is a road, not just where there are tracks), and will benefit relatively more from automation.  To be clear there are still some energy advantages of trains, and the costs for longer distance markets which connect places that already have trucks for the last mile (or last ten miles), trains will remain economic. But the threshold for the truck and train trade-off will change.

Private transportation will primarily be on wheels, but we should start to see the beginnings of ‘flying cars’, for lack of a better term. By 2045 we are 30 years out, small helicopters, gyrocopters, and  multicopters are not only possible but will be on the verge of being feasible for daily transportation.  I don’t think they will be widespread by that point, but I think we will be in that period then roughly where we are today with automated vehicles. People will be doing tests, there will be some pilot cases, but it won’t be widespread deployment. The energy cost will still be relatively high and the range will be lower unless there are remarkable battery breakthroughs. There will remain safety issues with ‘flying cars’ as people on the ground will be nervous.

On ground, will vehicles be steered or guided by computer? They will be autonomous, I think this is critical, that there will be no central controls, there might be connected vehicles but vehicles will all be individually, not centrally, controlled. I don’t think people would trust a central vehicle controller.

In terms of centrally controlled vehicles, there’s the risk of failure in communications, hacking into communications. I’m not terribly optimistic about the effects that connected vehicles will have. CVs might be useful in terms of ‘here I am’ type of message, vehicle-to-vehicle connectivity may be useful negotiating between two vehicles on a rural road, on a rural crossing, ‘you slow down and I’ll speed up’ type of thing. But I don’t think that will be important, the use case that they say, that they show in the demos is ‘well, there’s an icy patch ahead.’ I’m sorry, I live in Minnesota, there’s always an icy patch ahead, that’s not really informative. So I think the use case for the connected vehicles is pretty weak.  I also think the expectation that local communities will upgrade their infrastructure to make infrastructure to vehicle interaction important, is also weak.

Over the course of decades, communities will upgrade infrastructure, and one assumes they will install whatever the state-of-the-art is at the point. At that point the infrastructure can tell where cars are.  I think the more likely case is, that cities and towns start to take out traffic lights and replace them with roundabouts where they can, replace them with yield signs where they can, and leave traffic signals for places (downtowns) where intersections are oversaturated for at least part of the day.

Just as cars broadcast ‘here I am’ messages, new traffic lights will broadcast ‘I am red’ and ‘I am green’ and even when the light is going to change. Cars can speed up or slow down based on when the light is going to change, if the light ahead is about to turn red, the car slows down enough now, it can catch the next green, or if the light’s about to change to red, maybe the car speeds up a little bit to catch the tail of the green. This can squeeze a little bit more optimization out of signals. Still it’s better not to have signals at all.

My philosophy on this is that the infrastructure is going to be dumb, just because of who manages it and their incentives, and it’s the vehicles that have to have the intelligence. The roads are not going to have intelligence about steering cars.

This is not to say there will be no changes. The lanes will be painted differently in different places, because with enough narrow vehicles, roads can have half lanes.

I don’t think the road will have a lot of intelligence to it, I don’t think traffic signals will get a whole lots smarter, the main thing will be increased information, so instead of the traffic signal trying to anticipate cars and be adaptive and actuated, it will move towards a fixed cycle, but broadcasts the fixed cycle. This cycle might change in response to traffic in big cities which can afford it. Traffic signals won’t be terribly dynamic in most places. There’s enough infrastructure out there in low technology places in rural towns and so on, that’s just not going to get upgraded even in the next 30 years.

And I think sort of sadly that trains, at least in the US, will be automated after cars. It just boggles the mind as to why that would be, it seems so irrational, as trains on tracks should be easier to automate, but I think that’s how it’s going to be because of institutional reasons.

Will vehicles be segregated? I can imagine a special lane for automated cars for a period of time, so let’s say between 2025 and 2045 there might be some HOT lanes that are converted to automated vehicle only lanes, in order to get advantages of closer following distances than you would in mixed traffic. Other lanes may be converted eventually, before the whole system switches over. But this will generally not be adopted too quickly because there will still be a constituency for manually driven vehicles in the early years.

3. What will public and passenger transport vehicles look like in 2045, and how might they be configured?

• What size will the vehicles be? Will they be designed to carry large volumes of passengers, or for small groups/individuals?
• How will they be powered
• Will the vehicles be automated, and if so, to what extent?

In The End of Traffic and the Future of Transport we talk about  an explosion of new vehicle types. This is a really important, and has been largely unrecognised. Everybody assumes that vehicles will look basically the same as today, and will just be automated and electric powered. They may look like Teslas instead of Chevrolets, but to a Martian, they all look the same, that’s a distinction without a difference. We have the opportunity for vehicles to suit the new technologies, but we’re not really understanding how this is going to work. There are concept cars out there, but no one is taking them seriously.

On the one hand if I want to get from point A to point B and it’s just me, I don’t need to carry around extra chairs, so I can get into a much smaller car. And that of course affects road capacity, that means that where previously a lane had to be 12ft/2.6 meters wide, it only has to be 1.3 meters wide, which will double capacity. Of course if they are automated, they can follow more closely together as well, and the lanes can probably be narrowed up again, because they are automated. So that gets more capacity out of that smaller vehicle, more energy efficiency out of a smaller vehicle, greater range for the battery and so on.

People who own these single passenger vehicles might have one for every member of the family, and just platoon for multi-person trips. There’s no reason that everyone has to be in the same car, or facing forward – that’s the traditional car.

On the other hand, if someone wants to have a meeting while in motion for whatever reason, one could think of conference cars, where it’s a larger vehicle, with 4-8 people arranged in a circle, and they are meeting while in motion. Now I’m not sure why people would need to meet while in motion, that part escapes me, I think it will be unusual, but I already said, if there is a use case for it,  we could imagine it happening. We could imagine the family wants to go on a trip and they want to be in the same vehicle, so they would use a minivan type vehicle, but the chairs can face in any direction.

If the car is doing the driving, the driver doesn’t need to have the steering wheel, or pay attention to any of that, and everybody is a passenger. So at the top end of the market there are these ‘executive jet’ types of cars, where people do whatever they are doing, they have an office type set up in their car, they have a conference type of set up inside of their car. And given that anyone can order cars on demand, specialised kinds of cars may still need a bit of scheduling, but with the ability to basically order cars on demand, no one will  need to own specialized cars, and they can just get one when needed.

This also means that people who buy an extreme car for the type of activity they engage in once a year or once a month, like a truck for the suburbanite, won’t need to own the truck because they will be able to rent that conveniently when they need it. They will have concluded that it will be cheaper to own the small vehicle, and then rent the larger vehicle.  I think that will be the most common arrangement.

Now in urban areas, I think the market will shift much more towards rental, because people aren’t going to want to pay for parking, but in suburban areas, low density areas, the ownership model will still stick, because parking is essentially free, and it will be cheap enough to own, people don’t want to pay for rent, they don’t want to wait for the car to come to them. In low density areas it will take longer for the car to get to them. I think those will be important considerations.

2. What sort of public and passenger transport will be available in 2045?

• Do you see a range of different public and passenger transport service types or a convergence to a narrower range?
• Can you comment on how the transition might occur (i.e. what the transport system might look like in 2025, and 2035?)

In terms of public transport services, as I said, large cities will probably have high frequency services, commuter rail, metro, light rail, bus transit types of services with frequencies no worse than 10 minutes, probably on the order of 5 minutes. But if frequencies are going to be longer, people who have a choice won’t accept that because, why wait when they can get a taxi?

Taxis will be relatively inexpensive. Because they are automated, no need to pay  for labour. Because the vehicles are smaller, they are less expensive. Because, if they are managed well, they are in service a lot of the time, they are not wasting capital just sitting around.

So I’m thinking that while public transport will be available it will not be a public transport service as we think of it. It certainly won’t be publicly owned or managed, it will still be regulated by the public sector.

I think by 2045 we will be there as automated vehicles become standard. My timeline is that a customer could buy high-end automated vehicles about 2020 and automation will be mandatory in new vehicles by 2030, and  by 2040, non-automated vehicles will be phased out. [By automation I mean self-driving in many circumstances most of the time, and ‘many’ and ‘most’ asymptotically converge to ‘every’ and ‘all’ over this time period as software (and hardware) get better]

This will happen faster for taxis and freight, and other industries with high fleet turnover, as they are motivated. So I can see maybe by 2025 we start to see fleets of automated taxis, by 2035 all taxi services will be automated and that change may have played out. Clearly this is not an instantaneous transition, it will take several years, it will happen sooner in some places than others, it depends on the price of labour and the regulatory framework.

An island like New Zealand has the advantage of not having so many vehicles, there it’s very easy to regulate out of country vehicles, so they could convert faster than say the United States, or continental Europe, because it’s an island, and I think that’s something that is important. Other island countries have that advantage of fast fleet conversion as well. But the benefits of converting over are so great, in terms of safety, in terms of lower labour costs.

Carbuyers are going to convert to automated vehicles at the same time as they are converting to electric vehicles in all likelihood, because when a carbuyers is in the market for a new car, most will have both of those things bundled, they don’t have to be bundled, but they will be.

The environmental advantages of EVs will lead regulators to require new vehicles to be electric. Similarly, I don’t see why regulators will continue to allow non-automated vehicles beyond a certain point. The traditional modes, walking and biking will still exist, I don’t think that’s going to go away. Biking will probably be more popular, given the biking environment will be nicer, as the vehicles with which they are competing for roadspace will all be automated and less likely hit them. Similarly with walking, but especially with biking. I’d expect bike share will start to pick up in urban areas that have been more car dominated historically. In suburban areas distances are still pretty large.

This was based on an interview I did for Marcus Enoch and the interesting New Zealand project PT2045. As it was an oral interview, it was a bit informal. I have edited, revised, and extended my transcribed remarks. Since their reports seem to be out, I will just report my interview, which I hope helped feed the report. This is an 11 part series.

1. What do you think the local transport system will look like in 2045? (think about the whole transport system, including private vehicles, active modes, as well as public and passenger transport)

• What different types of local transport will be available in 2045?
• How will attitudes towards vehicle ownership differ in 2045
• Will there be new forms of transport or new ways of delivering transport services?

The most important change, will be the automation of vehicles – both private and public transport vehicles.

The second most important change is the conversion of the vehicle fleet to being primarily electric.

The third change is more tenuous, and focused on urban areas, is the rise of  shared mobility,  mobility-as-a-service,  where people won’t be owning cars, but instead will subscribe to a service, or buy the services on demand.

What types of transport will be available in 2045? I think private cars will still dominate most places. But  mobility-as-a-service, that is taxi types of services will be more common than today. Urban areas will have high frequency transit services in selected corridors. But in corridors that can’t support high-frequency public transport services, today’s low-frequency services will be replaced by mobility-as-a-service; instead of having a bus that comes once an hour, people will be using taxis – maybe single passenger taxis, maybe shared ride taxis – it’s hard to say. There will probably be a mix of those.

I’m guessing, there will be a lot of single passenger vehicles. The best example I have is the Toyota iRoad  vehicle, which is basically an enclosed motorcycle that’s safe and stabilised and, will, well before 2045, be automated. And it is safe because, not only is the vehicle designed well, with a roll bar and all that, but because all the other vehicles are also automated.

With automation, the bias western countries have against using motorcycles (and other powered two-wheeled vehicles) for transportation (as opposed to recreation)  will go away.

In terms of freight, I think that there will be a substitution of logistic services for a lot of shopping trips, and we need to think about that. In urban areas the best model I have seen was a demo of using a small robot for delivering goods from stores to people’s homes (see Starship Technologies), it only goes out when people say they are home, travels at walking speed. And it knows how to cross the street, and just waits in front of people’s houses until they remove their package, and it’s secure. So something like that as a delivery service in urban areas, I think, is highly likely.

I’m a little bit more sceptical of drone type of transportation, not that it won’t exist, but I don’t think that there will be a huge market for it, one could imagine specialised drivers or something like that. They used it in Minnesota for delivering beer to the ice fishermen on the lakes, but I don’t think that there’s a large market for drones, even if they are carrying 5kg goods, just because people aren’t going to want drones buzzing up and down over private property and it’s still going to be more expensive than doing this by ground, but we will see how that plays out.

There is nothing more important in human society, so it would seem, than relative status. It doesn’t matter how great your wealth is, if someone else is wealthier. It doesn’t matter how great your university is, if someone else’s is ranked higher. But how do we actually rank universities? How can we compare such dissimilar items?

I believe university rankings follow from sports rankings. Sure, West Dakota A&M may be ranked number one in Ice Lacrosse, but we’re smarter. How do you prove that.

If it were a business, we might look at sales or profitability or market share or customer satisfaction ratings. For universities, we need other metrics. And there are a lot.

Ranking Universities

Having been in universities a long time, I have seen rankings come and go. Wikipedia has an amazing list. Comparing the lists, on most, year after year, the same set of schools, more or less, are at the top. But their positions vary widely.  West Dakotah A&M rarely makes it. There are a variety of ranking systems profferred.

In the US, the most famous is the US News and World Report College Rankings. USNWR is a defunct weekly newsmagazine, the RC Cola to Time’s Coke and Newsweek’s Pepsi. But their College Rankings live on. They use an arbitrary weighting of a variety of factors, a weighting which changes regularly, to develop a ranking.

Now if this were just about status, it wouldn’t matter except for alumni egos. It does however have a self-fulfilling element. If you are an international student and are offered admission to 3 US universities, with otherwise equal offers, you will tend to choose the one with the highest ranking. If you are a national employer, you will recruit from the highest ranked university for what are perceived as the best self-selected students.

However it is important to note the US college system is not nearly as an effective filter as those in Asian countries. While in the US, the best student at West Dakotah is almost assuredly better than the bottom two-thirds of Minnesota, in some countries which take entrance exams very seriously (Iran, China, India, e.g.), the worst student at the best school might very well be better, or at least smarter, than the best student at the next ranked school. (Well probably I exaggerate, but you get the point).

Now perhaps you aren’t looking at USNWR. Maybe it’s Shanghai Jiaotong (which incidentally means transport in Chinese) University’s ARWU rankings. Or maybe it’s the Times Higher Education Survey, or one of two-dozen  others. It doesn’t matter much. It’s good there is so much competition, many schools can claim to be number 1 or top 10 or some such.

My University of Minnesota once wanted to the be among the top 3 public universities in the US. It didn’t say which list though. I assume whichever put them at number 3. (I always assumed this would involve treating the different University of California campuses as one university). This goal was abandoned quietly it appears, though it still is mentioned in this 2011 document, which says top 3 public research university “in the world in a decade”.

These numbers are so meaningless, I wish people would stop talking about them. But no.

Ranking Programs

Rankings also give status to individual programs. This is even harder. Whose Civil Engineering program is the best? The one with the most money per faculty member? The one with the most graduates placed? The one with the most publications?  Caltech, a fine school which doesn’t even have an undergraduate civil engineering program, was highly ranked by one of the surveys one year. (Sorry can’t find the link).  Reputation is so strong it can rank non-existant programs.

The granularity of this is getting deeper. Joseph Chow has ranked  Transport Engineering programs. (Based on H-index). The University of Minnesota did well. Obviously H-Index is a function of size, so larger programs will tend to score better. Sydney (ITLS) also scored really high. My alma mater, UC Berkeley topped out though. While I don’t agree with every detail, and it is biased to larger programs, it seems truthy. But really, what’s the difference between an h-index 29 and 30? Three more citations. (The 29th ranked paper needs one more and the 30th ranked paper needs two more).

Ranking People

Individual academics are presumably the limit of this. Tom Sanchez has ranked US planning academics. Despite being an engineer, I show up on this list (I do have a Humphrey School affiliation, and I do some planning stuff). Still, a lot of my work is more engineering-ish than planning-ish, so I am not sure about what the right methodology would be. Google Scholar tends to be noisy, so I am reluctant to rely on that as the be-all, end-all, but ok, I’ll brag about it.

The always fantastic Etymology Online writes:

guild (n.) also gild, early 13c., yilde (spelling later influenced by Old Norse gildi “guild, brotherhood”), a semantic fusion of Old English gegield“guild, brotherhood,” and gield “service, offering; payment, tribute; compensation,” from Proto-Germanic *geldjam “payment, contribution” (source also of Old Frisian geld “money,” Old Saxon geld “payment, sacrifice, reward,” Old High German gelt “payment, tribute;” see yield (v.)).

The connecting sense is of a contribution or payment to join a protective or trade society. But some look to the alternative prehistoric sense of “sacrifice,” as if in worship, and see the word as meaning a combination for religious purposes, either Christian or pagan. The Anglo-Saxon guilds had a strong religious component; they were burial societies that paid for Masses for the souls of deceased members as well as paying fines in cases of justified crime. Continental guilds of merchants, incorporated in each town or city and holding exclusive rights of doing business there, arrived after the Conquest. In many cases they became the governing body of a town (compare Guildhall, which came to be the London city hall). Trade guilds arose 14c., as craftsmen united to protect their common interest.

So what guilds did was try to establish monopolies in their various domains. They aimed  for a cartel to control prices and entry, capture a greater share of economic rent and more profits (even if at the cost of the economy as a whole). For instance, you couldn’t have carpentry done without hiring someone who was a member of the  Carpenter’s Guild, who would charge the same rate as everyone else in the Guild. Anyone discovered working as a Carpenter outside the Guild would have likely have violence done to them until they repented.  As subspecialties formed, disputes over which guild was responsible for which work arose.

As I have written before, economic actors don’t want a competitive market for themselves, they want monopolies, so the rise of guilds or business associations, corporations, labor unions,  and the like is natural. Actors do want everyone else to be competitive though. So in short, members of a guild logically dislike (and are poorly served by) every  other guild (but their own).

What we have is a prisoner’s dilemma. I want my protections, you want your protections. We would both be better off if neither of us had protections, but there is no incentive to move in that direction, so we wind up with everyone operating as a cartel. A constitutional arrangement between the parties could break this up, as can changing not only the rules of the game, but also the players.

Eventually of course the guild system, which was stronger in cities and towns, and weak in rural areas where specialization was harder to come by, dissipated, as new competitive industries opened up, as the technico-legal innovation of corporations routed around guilds, as trade increased and brought in products from out-of-town, and so on.

We still have guilds today though, but they aren’t called that. To name a few:

• Police
• Lawyers
• Doctors
• Teachers
• Professors
• Civil Engineers
• Traffic Engineers

We could further look at intersections, like Civil Engineering professors, (can’t be one without a PhD from an accredited university in Civil Engineering or a similar field), or subdivisions like Transport Engineering professors (likewise).

Some fields are more open, but the trend is try to close things down with occupational licensing, (can’t decorate nails without 3200 hours of apprenticeship training nowadays, that’s longer than a Master’s Degree in Civil Engineering, by the way) which is in many ways replacing unions. At the organizational level, accreditation is a similar process by which entrenched interests (say universities), try to keep out new entrants (say bogus universities founded by famous real estate developers). This is all supposed to be done on the behalf of the public interest of course, and there are many bad players out there who need to be differentiated from the real. I am not clear why people cannot figure this out on their own, but apparently some people cannot.

Think about who stagnation benefits and who change benefits. This explains why we are in the situation we are in. The guilds are ensconced, and growing, and breaking this prisoner’s dilemma is increasingly hard. I suspect the police (and military) will be the last to go, but I think guilds like education are ripe for the taking in many non-guilded fields. Knowledge is essentially free now, higher education’s monopoly is that it provides accreditation. That might not be the simplest thing to provide, but surely it is providable with recommendation systems. It is a question of establishing networks of trust. The difficulty is that education feeds existing guilds (say Civil Engineering departments training Civil Engineers), and the guilded professionals will be loathe to accept new-fangled accreditation, as it could be perceived to devalue their own degrees. Since Civil Engineers are legally established (through the Professional Engineering licensure, granted by the state) it will be hard to erode this. On the other hand, consider Electrical Engineering, or especially Computer Science, much more fluid fields, where the degree counts for much less and ‘professional’ status is not considered as critical. Certification from an outside player needs to be accepted by private companies, not by the state, and motivations change from the downside risk-avoidance nature of government (fear) to the upside risk-seeking nature of capitalism (greed). Hence we see entrants like Udacity playing in the Electrical Engineering and Computer Science space, but not the Civil Engineering space.

But just as this system has lasted nearly a millennium (the University of Bologna dates from 1088), it won’t be unbuilt in a day. Universities have multiple functions, they are conglomerates, not all of which are the same monopoly position as accreditation. They function as dating pools for the upper middle class. They are large research labs and hospitals. They provide sports entertainment. Those are already competitive functions.

A South Korean magazine, Economy Chosun, asked my opinion of the recent Trump plan.  Update: This has now appeared in print (in Korean). My answers (in English) below.