A broken algorithm that affected American transportation
vice.com
vice.com
Background: my first job in the UK Civil Service was as an admin assistant (I did the filing) in the Highways Contracts division, Dept of Transport. My team's job was to let the contracts for developing plans for new/improved national roads across England.
The Algorithm: roads cost a lot of money to build; every proposed route had to demonstrate significant value for money through a cost-benefit analysis for each project, considering a number of different detailed options for the new or improved road. One of the main costs was how much money needed to be spent on buying and securing the land, and compensating the owners for loss of business, etc. In cities and towns, this favoured putting bypasses through the poorer parts of the (sub)urban environment. In more rural areas it favoured routing roads through less productive farmland. The least productive land was land that wasn't being farmed or mined. Routes put through these areas would be given an acquisition cost of £0/meter[1]
... Which is why in 1993 the Government settled on developing an improvement scheme for the A303 which put a new double-lane carriageway straight through the middle of Stonehenge.
The proposal didn't go down too well. 30 years - and several administrations - later, the Government is still trying to get the proposal through planning[2]
[1] - Though in fact highway lengths were still measured in 'chains' and 'links' back in the early 1990s: https://en.wikipedia.org/wiki/Chain_(unit)
[2] - Current proposal for the project: https://assets.highwaysengland.co.uk/roads/road-projects/A30...
However, the largest problem I see which the article briefly touches on is the failure to even consider building something besides a freeway. A `correct` model should be screaming to build public transport.
However, consider the following circumstances:
1. As it is many people are forced to have a car even if public transport can serve their main commute but often public transport is more expensive than the marginal cost of driving a vehicle (i.e gas + upkeep) so they choose to drive unless the congestion is extreme (i.e one of the worst commutes in the US).
2. Even if public transport is pretty decent (commute + groceries), they just happen to already have a car because their previous location did not have public transport. So we are back at #1. Or I don't know if I'll move in a year or two so might as well keep the car.
3. OR now that I have a car , I would probably not want to have to pay for expensive parking (& housing) in an area with public transport plus now I'll have to pay a mechanic a premium to even change the oil. So instead I live further away thus reinforcing the need to keep the car.
4 & etc. I need to travel >50 miles (skiing, hiking), renting is a pain and long trip ride sharing is expensive. Groceries are easier to do with a car especially when buying in bulk (Although, if/when I can get Costco delivered same or 2-day with perishables at <=10% premium, then this is a moot point).
Cars are pretty (at least marginally or per mile) cheap in America and we do a very poor job making the alternatives attractive if not impossible. Anyone who has a car seems to drive it even if public transport could be used (again outside of super specific commutes in a few cities). Until we can either make car ownership/use more expensive or greatly increase the practicality of not having a car at all, anticipate having a car and being stuck in traffic at least part of the time.
But then society will have to vote for sacrificing luxury today for the benefit of future generations, which typically doesn’t go well.
It's a reasonable position and one of the options before those we have chosen to lead us. Then again, the argument "but muh munny" is a strong counterargument among those who own them.
As it will be for you, should you manage to accumulate any wealth.
However, it has nothing to do with the separate problem of creating a hostile environment (the physical world) for future generations which can only be avoided by reducing fossil fuel usage which can only be accomplished one way - increasing the cost of using it.
As far as how much personal wealth I’m willing to depart with, I’m willing to do whatever it takes to get people healthcare, minimum vacation time, sick leave, higher education, and parental leave.
Reducing population also only buys time. I think you already know both of these things.
If your time is worth little, public transportation will work. But if you want to get from home to work and back, the car will be quickest. And if you want to go to the grocery store on the way home, no public transportation helps with that.
There are a few modifications of that.
If you have to get to work and you're not in a rush and you have something to do along the way like listen to a podcast or read a book, then that might be slightly different, because you're living a bit. (although podcasts work in the car if you are not against driving + listening - some people are)
Public transportation works really well when the roads are full full full. Getting from one side of Mexico City to the other quickly during rush hour is possible with the metro, not possible in a car.
The corporate performance of railroads in the 19th and 20th century are distinguished by comical poor performance, abusive practices and unrivaled greed. For all the faults, the industry around cars is highly competitive at almost every level.
This line in the article is the crux of the issue.
> The question is not whether the predictions of how they will behave are accurate, but what kind of behavior we want to have more of.
How we build determines what kind of environment we have not the other way around. People WILL sell their cars if they have good public transport. They will live in dense areas if they can. This has been the case in the US and all around the world.
P.S. I really cringe at you saying that cars are cheap while talking about your ski trips. You obviously have never been around poor areas of the United States. Go hop a ride on a bus on the south side of Chicago some time. Go tell the people who commute an hour a day by bus that they should just buy a car because its so cheap.
The immense popularity of the quiet suburb seems to disprove this statement.
I get what you are saying, but there is so much more going on that makes them "popular", particular in the United States.
Firstly, is the cost of living in the suburbs is massively subsidized. In the coming decades this will become readily apparent as most of suburban America will not be able to afford its infrastructure. It is already starting to happen.
Secondly, the suburbs in America are often the only legal way to build. We've outlawed density in most of the country. Many of the most urban parts of our major cities are less dense than they were a few decades ago because of zoning changes.
Thirdly, we simply do not have good public transit. You may think that New York or Chicago have great public transit but they have substantially less per-capita ridership compared to many other equivalent cities.
I'd suggest looking to other parts of the world as examples of immensely popular dense development. In Asia, Tokyo, Singapore, many cities in China and across Europe are great examples of this. The U.S. is an aberration - not the rule.
(high) cost, noisy neighbors (and poor insulation), zero design for ventilation or privacy. This isn't directly addressed in the suburbs either, but the lower density in the suburbs decreases the effects.
I've never lived in a shared dwelling that isn't an apartment though. Maybe the equally unaffordable condos and such in cities don't suck; however I'm inclined to believe that they share these faults and all of the distractions of others, particularly in cities with mostly US citizens.
Haven't we just learned that being in close proximity to people in unclean settings is a big risk to public health? Public transit advocates are going to have to address and overcome that before anyone should listen to them.
And that's not even considering the massive decrease in commuting needs from distributed/WFH workforces.
Plenty of non-US cities denser than NYC did fine. and NYC is also doing pretty well by now. Consider basic stuff like how much is spread before CDC switched to promoting mask usage.
Respiratory Pandemics are also relatively rare events. Sorry, but it's true. Meanwhile suburbanization and car use causes tons of air pollution which sickens people constantly. Integrate and I am sure the suburbs come off as causing more unhealthiness.
Recent wave in suburban eras (Forida) + whatever doom we'll face in the winter will also make the April NYC a fainter memory.
------
The subway in particular has excellent air filtration systems, and there have been no super spreading incidents linked to it, so as long as it's less than rush our super-packed, it's fine. See https://www.nytimes.com/interactive/2020/08/10/nyregion/nyc-...
You don't seem to have addressed the needs of anyone who doesn't live in a dense city. How about suburbia? The midwest? Try walking a couple blocks in Phoenix, AZ in summer.
I really doubt that our government, state or national, is capable of maintaining a decent infrastructure. California is certainly a good example of the propensity towards failure. (Maybe if we try again, with some extra rules, they'll do better next time?)
> we do a very poor job making the alternatives attractive...until we can either make car ownership/use more expensive
I really hope that the solution to poor transportation infrastructure won't be making the part that works even worse.
For reference, the four stages go like this: you start with "trip supply and demand" by zone, then inter-zone trips are generated by matching those up, then mode choice is computed per trip based on cost and travel time priors, then assignment to the network which results in actual trips and computed travel times based on route selection and congestion.
The travel times can then be fed back into mode-choice and iterated on: people will make the transit-vs-car choice differently based on the relative costs (in time and money and inconvenience) so iterating the last two steps can go a ways towards modelling that equilibrium.
Folks can exert some amount of choice around which trips they make (as mentioned in the article) and where they work, based on how easy it is to get there, so feeding the travel costs back into trip-generation can account for some of that, and over a longer horizon, both workers and employers can exert some control over where they live/site their facilities, so all 4 stages can be iterated as well.
None of this is perfect, and it's not always done, and the rest of the article's points stand, but I just wanted to say the modelling is not all that primitive :)
Edit: I studied this stuff in university 15 years ago with Professor Eric Miller, who wrote this book on how to integrate all this stuff in an iterative way... https://link.springer.com/article/10.1007/s11067-005-2630-5
All models are based on assumptions. Revisiting those assumptions should be a continuous process that forms a tight feedback loop, especially given the long timelines and large capital costs involved with civil infrastructure projects.
If population and land-use assumptions are the key factors then the important questions should be how these assumptions will be validated and how often. These questions seem to be missing from the article.
> City planner Jeff Speck has called induced demand "the great intellectual black hole in city planning, the one professional certainty that everyone thoughtful seems to acknowledge, yet almost no one is willing to act upon."
to me, as a user of public transit, they just seem somehow more pleasant, in a way I can't really make sense of. I strongly dislike riding in buses and try to avoid it, whether part of transit systems, Greyhound/Peter Pan, or even high-end charter buses.
maybe if you had electric (no exhaust, quiet) buses with frequent reliable service, dedicated lanes almost everywhere, nice bus stops without hostile architecture, I wouldn't feel such a strong preference, but I'm not sure.
I'm wondering whether it's not just that I subconsciously dislike getting jostled around and feeling slightly carsick.
Trains are excellent for dense urban environments and better yet when integrated with another modes. I take Berlin as an example: You have fare zones and S/U-Bahn everywhere in the city center extending for to Potsdam, etc., trams in the former east, buses on the less dense parts of the city. Also trivial access to regional and long-distance trains. And it also disproves the typical excuse "You can't do that in an old city": some U-Bahn lines exists since 1902, the city got devastated by war and then there was the DDR. The reunification needed a lot of work and yet the system survived. Took me a long time to notice there was no Ubers, maybe because it was not that advantageous for people to use it and Uber Inc. just gave up instead of fighting city regulations, as they did everywhere).
I compare with the systems from my native Brazil, where public transportation ranges from pathetic inefficient on big cities to non-existent in small towns. Hell, I really miss Berlin.
Second, in many cities, trains actually predate the highways.
Third, trains have super high capacity, as do subways and light rail. You need more buses to keep up. However, buses are cheap, and there are things like extra-long articulating buses and double-deckers that can be used on high-volume or long-haul routes. I think buses actually have a practical edge here: instead of trains coming every hour, you could have a big bus come every 10 minutes, which smoothes out load at end stations. And if one breaks down, the other buses can just drive around it. If a subway breaks down in New York, it can cause delays that ripple out to the whole system.
If you're good about it, you can drag the project out with delays but with milestones you can talk about politically.
Then by the time the project has been discovered to have been unnecessary you're out of power and it's no longer a problem.
And your investors have a good return to invest in the next guy to give them the next project.
But it has none of the anti-bodies to the kind of cynical corruption that many other countries find endemic. And so when it came here, America had no defence. Because the host is still so healthy, it still doesn't realize it is carrying this parasite.
It'll be interesting to see what happens when the host becomes unhealthy.
It feels very similar in spirit to that story about Google reducing data center energy use by 40% via ML optimization: https://deepmind.com/blog/article/deepmind-ai-reduces-google...
More broadly I've started wondering if there's a huge opportunity to apply software to civil improvement. Everything from detecting patterns of corruption, to aggregating data that can inform voters, to identifying sources of inefficient spending. I don't put much stock in things like digital voting/governance, but in the area of "data crunching and presentation" I think there's a lot that could be done.
I'm not sure this is very helpful to predict traffic patterns over the next 20-30 years in a policy-sensitive way... All the historical data in the world wouldn't have predicted the rise of Uber, the pandemic, the maybe-eventual rise of self-driving cars etc.
Ultimately what is the correct metric though? Well being? https://www.mdpi.com/2071-1050/12/8/3180/htm#
Americans have been basically brainwashed into thinking that patterns of development from that era are right and good and so forth and then we put those models on steroids and do more of that same.
Our suburbs began as houses of around 1200 sq. ft. Last I saw stats, the average new home in the US was over 2400 sq. ft. (though a recent private conversation with a professional urban planner suggests those figures may finally be dropping).
We need an entirely new set of mental models, something I think of as "the butterfly economy" based on a comment I made on HN once.[2]
There is a potential path forward, but we aren't going to take it because that would involve listening to a woman and this will not happen. This is a hill our current world is eager to die upon, so we will remain haunted by the ghosts of Christmas past -- prisoners of mental models that served our parents and grandparents extremely well, so they got burned into our collective psyche as the only good way to do anything and never mind that the world has changed and all of that is very much out of date.
I'm short of sleep, tired of getting zero respect, underfed (as is the norm late in the month), in a generally bad mood over how my crappy existence remains crappy largely due to sexist BS and no amount of pointing out the sexist BS causes it to really budge. So I think that's all I'm going to say here since I fully expect to get kicked in the teeth for saying anything and I've had quite enough abuse for one life.
And concludes it's politics.
Not sure if I read the same article as everyone else?
The future is autonomous minibusses with algorithmically-derived live routes, served on a subscription basis by private companies to private individuals. This way people of similar socioeconomic status can rub shoulders with eachother and feel more comfortable, and receive a greater standard of service.
It might very well be necessary to modify roads and road rules to facilitate autonomous vehicles - that's the kind of public investment we should be making.
You end up with a system where there is even more socioeconomic exclusion (poorer people living in areas where they need to use cars, but cars are more expensive and won’t get into the future cities).
Or - at worst - it won’t work at all, because you will need a car anyway for those 5-10% long tail destinations that private transport won’t cover.
I think this is the reason European countries have public transportation - which is forced to tackle less profitable routes, but can earn back money on the profitable ones. (Similarly post office)
Moreover, when they went public in 2019 they said that "Uber may never turn a profit". And yet here we are.
That's a lot of work just so you can avoid being near poor people.
The notion that increased road infrastructure will magically hold congestion steady or worsen it is laughable. It just means demand hasn’t been satisfied yet. Clearly a 1000 lane highway would not be congested, so there is some finite and lower number where demand balances out with supply. This is no different than supply and demand in any other situation, which is why the label of “induced demand” is without meaning and simply a rhetorical device.
It's different because highway usage is free. Demand is induced because you're giving away something for free. Demand plateaus when the costs (i.e. traffic) rise to the previous levels.
Image if SpaceX started giving away free rides to Mars. Millions of people might go. But what if they charged a price equal to the actual costs--manufacture of the spaceship, fuel, etc? The cost would be ridiculously high and only a tiny handful of the richest people would go. Those handful of people represent actual demand where it meets the supply curve. The difference between those willing and able to pay, and all those willing to go if free, is the induced demand.
None of this means that new infrastructure can't create a richer, more productive environment.[1] Obviously the real world is more complex. But the pathological phenomena of induced demand in especially clear and substantial for highway construction because of the process for choosing, designing, and budgeting projects. And to the extent the actual costs are born by the individuals voting for and using a new highway, the linkage is particularly weak and distant and so the costs don't moderate the demand as should normally happen in well-functioning markets. When you try to bring them closer, say through tolling, people become apoplectic; "That's what my taxes are paying for!", etc, ignorant of the point that costs don't moderate consumption very well unless the relationship is clear and obvious--usually the more immediate, the better.
[1] Though that's why the article criticizes reliance on modeling, because at the end of the day the choice is less about planning for an inevitable future than picking & choosing among potential futures, each with costs & benefits that are difficult to quantify independent of policy goals--equity, sustainability, mobility, etc.
No it isn't. The theory is used for the proposition that you can't satisfy the whole demand because adding capacity creates more demand. But that's plainly false. Even at a very low price there isn't unlimited demand, so there is an amount of capacity that can satisfy the demand.
Moreover, fuel taxes do actually pay for a large fraction of road costs, generally something like half. That means you get more demand than you would with a pro rata allocation, but that's what you want, because an even larger portion of the road cost is fixed cost, and charging more than incremental cost results in inefficient resource utilization. You paid for the road, it's not at full capacity and yet you're discouraging productive use of it.
Comparing physical construction and maintenance costs to fuel taxes is excluding 99% of the costs and benefits; it's misleading comparison.
People don't want to drive, they want to get to places. That's the actual demand that needs to be satisfied. Cars don't scale very well to do so. You can't realistically build roads in cities that can move the same number of people as a subway line. If you did you wouldn't have a city anymore, you'd have a parking lot with some buildings sprinkled across it.
But if this is your concern then the thing that really moves the things people actually want to do farther apart is low density zoning requirements. Which is also the thing that keeps mass transit from being viable, because it needs enough population density to fill the vehicles while still offering frequent enough service to be competitive.
It is incidentally clear that low density zoning requirements are not the singular driving factor for sprawl simply because the sprawl occurs even in places without those rules (e.g. houston).
Furthermore, effects like these have complex feedback loops. Zoning laws are entirely politically motivated - so the public's support is critical. You can be sure that public support would be affected if it became clear they led to crippling congestion that would not be solved by more road infrastructure; and conversely sprawl-encouraging zoning laws also encourage sprawl-enabling transport investments. Other factors to consider are parking availability (and rules concerning that), and probably more.
The correlation between these effects means that you should be careful blaming just one factor, and certainly cannot conclude anything meaningful if you intentionally and artificially exclude related factors from your analysis.
Otherwise it would be just one huge mess.
In the 1000 lane situation, people can get where they want, but where is it that they're going to if the fraction of space is mostly roads?
You can see the practical limits of pushing for more, wider roads at scale clearly in Los Angeles. A massive grid of wide freeways. Long commutes. Bad traffic even outside rush hour. Auto dependency. Air pollution. High household transportation costs.