Hacker News Reader

Comment by adamjc | Hacker News Reader

I thought the current thinking was that building more roads means that there will be more roads to use, and therefore more cars on the road.

https://en.wikipedia.org/wiki/Induced_demand

View on HN

Induced demand often happens, but certainly not always. It just depends on latent demand and network effects.

More importantly, the appearance of induced demand is not a general argument against building roads! People get value out of getting to places. If we build road B intending to relieve congestion on road A, but road A stays just as congested, we nevertheless have enabled people to get places with road B that they would not otherwise have been able to reach. That increases utility.

All of this applies to mass transit as well. The BART is currently at capacity during rush hour; if a parallel transit line was built, this would probably fill up with people without reducing congestion on BART, but it would still be creating value!

Do you happen to have any examples of where induced demand has not happened? My understanding is that it is pretty much guaranteed, especially in larger cities. I think New York's bridges and tunnels are the canonical examples of this. I say pretty much because I assume there must be cases where it didn't happen, but I am unaware of them.

Maybe induced demand has always happened, in the sense of demand increasing when the product is more available. But it's plenty common for the amount that demand increases to be less than the amount of extra capacity available. Furthermore, is induced demand that bad? If you double the capacity and it again gets filled to the point of gridlock, well then at least twice as many people are now able to be in gridlock, which they apparently prefer to the alternatives.

The example of when induced demand is bad is when it causes increased pollution. In many parts of the world, there is a deliberate movement towards avoiding building too many roads, in order to decrease road use and therefore pollution levels.

However, a car sitting in traffic will produce much more pollution per mile than one travelling at 70mph on a clear road, so the thinking behind this may be flawed.

It's possible to change the incentives.

If the roads are free they fill up until gridlock makes them more painful than the alternatives. If you charge money then the expense becomes part of the cost/benefit calculation. It favors car pooling, rich people (who don't care about the price) and other people who urgently need to get somewhere during rush hour (who can pay the price that one time).

You can also just do car pool lanes, which has some similar benefits.

> Furthermore, is induced demand that bad?

It's good in cases where you can eventually meet that demand. Say induced demand for broadband leading to more fiber optic cables being laid down.

However there isn't enough space in cities to ever meet demand. What needs to happen is a shift away from the single-car culture in America and towards mass transit.

To be fair, for traffic, it's not necessarily the number of cars that's increasing. If everyone doubles the number of miles they drive, that will have a roughly equivalent effect. Car-miles is really the metric to watch here.

City's tend to have a lot of latent demand. There are some highway segments outside of city's that get minimal usage after they where built but which then grew over time instead of the sudden surge due to latent demand ex: 81 in southern VA which now sees heavy use, but was fairly empty right after construction.

However, in the city case new roads with high latent demand you may see a shorted rush hour which is a very useful side effect.

It seems to me that an urban car tunnel would be the most likely thing to induce demand: it's useless for local traffic, and is so expensive that it's only going to built in places with lots of latent demand.

Build one and I agree with you, build 10 and that's no longer true. So, the question becomes can you make it 10 times cheaper.

Remember, congestion lowers throughput and surface traffic scales really poorly.

Airports. Especially in Europe, many small cities built their own airports, hoping they'd get passenger traffic. Many of those are now closed or have to be sold off for a nominal amount, as flights concentrated on major airports instead of spreading out.

How many people commute by plane? It's not a good comparison.

Well, considering that time to get through security equates to a large number of commutes, and you haven't even gone anywhere yet...

The total number miles driven on roads will almost always go up when you build more roads, just like increasing the supply of apples will lead to more apples being eaten. (The rare exceptions would be if there are a fixed number of people taking a long circuitous route from A to B, and a shortcut road is built; then the fixed number of drivers would get there more directly, reducing the total number of miles driven.) What normally happens with apples is that the price of apples goes down when there is new supply, roughly analogous to a decrease in congestion (i.e., the "price", in their time, that drivers pay to use a road).

When people talk about induced demand as an argument for building roads, they are claiming that induced demand will mean that congestion on the existing roads will increase (or at least not decrease). Here you have to take into account that most places where you build roads are growing for exogenous reason, so it doesn't make sense to compare post-construction congestion to pre-construction congestion. You have to compare post-construction congestion to the counterfactual where no construction was done. When you do that, you find (my emphasis)

> for every 100 percent increase in capacity there’d be an eighty percent increase in travel, reflecting increased travel speeds and land use shifts along improved corridors. However, only around half the increases in speed and growth in building permits was due to the added capacity. Factors like employment and income growth accounted for the other half. Accordingly, the traffic gains that one can attribute to the added capacity is actually around half of eighty percent, or forty percent. This is substantially less than reported by past induced-demand studies.

http://www.uctc.net/access/22/Access%2022%20-%2004%20-%20Ind...

The answer to all of this is well known: dynamically price road usage to keep roads at max capacity without noticeable congestion. Then each new road built is pure benefit.

We have one of those dynamic price roads around here. It starts out overpriced and goes up from there. It's basically the road for the 1%. Usage has been a quite a bit lower than the estimates so the taxpayers have been on the hook for lost profits.

http://www.washingtontimes.com/news/2013/nov/17/not-so-fast-...

I heard once that there was a government-managed construction project not featuring dynamic pricing, and that it cost much more than expected so the taxpayers were on the hook for the cost.

> My understanding is that it is pretty much guaranteed, especially in larger cities.

Induced demand can only happen when there is a non-trivial amount of people who have and use some other option. (NYC, for example, where there is a large subway network. A new freeway might convince some subway-users to switch, inducing demand)

For a city with no alternatives (like no meaningful public transit), it's impossible to induce traffic demand, because demand is already at 100%. Traffic still grows as population grows, but a person can only physically drive one car at a time, it's currently impossible to induce any higher demand than that.

> Do you happen to have any examples of where induced demand has not happened

https://en.wikipedia.org/wiki/M-6_(Michigan_highway)

"A year after the freeway opened, traffic volumes along parallel roads like 44th, 56th and 68th streets dropped 40–50%"

That's pretty typical for proper freeway construction/expansion. It's just that we stopped doing that 20-30 years ago, but population growth hasn't stopped in the meantime, so we're way behind on our urban transit infrastructure (all of it, freeways, rails, and public transit).

> For a city with no alternatives (like no meaningful public transit), it's impossible to induce traffic demand, because demand is already at 100%. Traffic still grows as population grows, but a person can only physically drive one car at a time, it's currently impossible to induce any higher demand than that.

This is not true. An example is Houston building a 26 lane highway only to lead to ~40 increased commute times [1]. Initially commute times decreased, but then the population felt they could move farther out into the suburbs, get a more affordable house, and still have the same commute time. Once enough people did this traffic times got worse.

So, demand increased.

[1] http://www.citylab.com/commute/2016/01/highway-boondoggle-us...

It'd seem in that case it's not that demand increased but that people moved further way from their final destination so their commute time was increased.

It may very well be that the same people are still driving or wanting to drive, but simply by living further away the commute times have gone up.

If people travel further then demand has increased. It is strange how people think about congestion in terms of number of people rather than miles driven. A single car that hits several choke points on a 50 mile commute is causing far more congestion than a vehicle travelling 5 miles. Yet the users are treated exactly the same and never get any social pressure to reduce their demand.

There's no such thing as demand being at 100%, certainly not in this context. Demand is a curve. If the cost of driving decreases--i.e. less congestion, easer traveling--people will drive further and more often. It won't happen instantaneously, of course, but demand will adjust even if the population stays the same.

That doesn't mean the new road will fill up completely. There may be other constraining factors. But if a route was highly congested before expansion, you can bet demand will go up after expansion.

  > "A year after the freeway opened, traffic volumes along parallel roads like 44th, 56th and 68th streets dropped 40–50%"

That doesn't tell you anything about the total volume of traffic. And in any event it's precisely what you'd expect to see. If a line at the checkout counter has 10 people in it and new counter opens up, expect about 5 people to immediately move over. But _also_ expect a few more people to join one or the other line now that they'll clear quicker.

Spot on with the "still creating value" stuff. With most things, creating more so more people can have it is seen as good. With roads, it's somehow seen as bad.

It makes sense if the roads have negative externalities, like noise, or ugliness, or pollution, or destroying pedestrian access. But none of that applies to tunnels used by cars running on clean energy.

> somehow seen as bad

I've noticed a certain utopian streak in some urban planning enthusiast communities --- one that's not compatible with the automobile as the primary mode of personal transportation. The general thinking seems to be, "$PROPOSAL reduces the number of cars on the road? Good!"

Back when I lived in Buffalo, NY, there was a group of people who'd write editorials in The Stranger (the local alternative weekly) and various blogs demanding that the city or state tear down the "Skyway" (a large bridge for NY 5), rip up expressways, narrow roads, and so on. Buffalo is not a wealthy city. The city is lucky to have inherited this infrastructure from a more prosperous era; it'll need this infrastructure if it's to become prosperous again.

It would be ridiculous to spend millions destroying this infrastructure when the city can't even afford to remove condemned houses in abandoned sections of the city. Yet week after week, people would publish articles demanding that the city do just that. It was incredible.

The only explanation I have for this impulse is that these advocates of infrastructure removal were really motivated by an aesthetic romanticism, an impractical idea of what a city should be, and a general dispositional aversion to modernity generally and to internal combustion engines specifically.

This mentality drove me up a wall. The automobile is an incredibly useful invention. It's allowed billions of people a degree of personal mobility that our ancestors could only have imagined. That's why we use the things so much. We can talk about how to make automobiles more efficient. We can debate the best schemes for allocating space in limited public infrastructure to the most important uses. (I'm a fan of congestion charges.) But just railing against the existence of automobile traffic is ridiculous.

I share the aversion to internal combustion engines. Carcinogenic exhaust should be a deal breaker when buying a car. I find peoples romantic attachment to legacy technology to be perplexing.

Gah. Buffalo's alternative weekly is Artvoice, of course. The Stranger is Seattle's. Mental wires crossed.

> More importantly, the appearance of induced demand is not a general argument against building roads! People get value out of getting to places. If we build road B intending to relieve congestion on road A, but road A stays just as congested, we nevertheless have enabled people to get places with road B that they would not otherwise have been able to reach. That increases utility.

There are cases where this doesn't apply, though. For instance, building motorways into the centre of cities. Yes, it means more people can get to the centre, but now there are too many cars in the centre for anyone to move around, including much more efficient forms of transport like buses. In that case, the city is going to function better if the road was never built, and people adapted their behaviour, either to move out of suburbs closer to a higher density core, or to work outside of the central area, or to use public transport. Unless you can have some other form of regulation, for instance to enforce a minimum number of people per vehicle, then an influx of inefficient vehicles in terms of road space to utility, will clearly damage the efficiency of the transport system, and the ability for people to move around.

Right, so what we need to do is take into account value of space that roads will take up (that could otherwise be used for houses, businesses, parks...), and only build more roads if the value they produce outweighs the value they consume. That's not an induced demand issue though.

I would say that's a pretty clear example of induced demand and the problems arising from it. You build the motorway, people build low density suburbs outside the city to drive and take the motorway to commute into the centre, that delivers hundreds of thousands of cars into the centre which crashes the road-space efficiency of the central area, which damages not just the people driving into the centre, but also the people who live close in, who are using much more efficient and scalable forms of transport. If the motorway was never built, people would live in apartments instead, they'd use buses or trains, and walk, and you could have more people working and living in the city while maintaining a functional transport system, and higher utility as a result.

It's not just a question of the motorway filling up, and maxing out an increase in utility, the motorway filling up means that more efficient uses of the roads in the centre are prevented. There is benefit balanced against detriment, and it's perfectly possible (depending on the exact circumstances) that the detriment will be higher than the benefit.

> which crashes the road-space efficiency of the central area, which damages not just the people driving into the centre, but also the people who live close in, who are using much more efficient and scalable forms of transport.

More traffic on the central roads is induced demand, but as jessriedel said, that's likely a sign of added value. Cars from outside displacing cars from the centre should only happen when it's a net gain - the limited road space is being used by those who gain the most value from it. People who live nearby and use trains etc. are unaffected by traffic on the central roads and can live exactly as they would if the motorway hadn't been built. Cars displacing more efficient buses is a real issue, but a general one not really related to building motorways (an increase in local car traffic would cause the same issue), and can be solved with e.g. bus lanes, congestion charges...

Certainly people get value out of getting to places. If you walk a mile in Paris, there are like 50 places worth going to. If you walk a mile somewhere around the middle ring road surrounding Houston, there's nowhere worth going to. Scale back car use to something reasonable, make the places smaller, move them closer together, and that way we can have nice things again.

I think this is indeed the solution, and European cities figured it out a long time ago.

However, there is a usability problem on the municipal level: if you're starting from scratch, which aspects of the European city do you build first, in what order, so as to build utility slowly without causing planning problems and without spending a bajillion dollars up front?

Surely the total amount of resources dedicated to transport has increased, and potentially the number of people wasting time commuting. There needs to be some practical limit to how much "infrastructure" we build of all types. Especially in localities that are already completely dominated by roads and cars.

The problem is that people live increasingly elongated lives that lack a grounded sense of place or community. Whole neighbourhoods are just a nuisance in the way between place A and place B. What about place C that is stuck in the middle with the pollution, noise, and congestion?

That's the good thinking, but because it isn't intuitive it often gets kicked to the curb by bad thinking. More roads = less traffic, right? right? No?

One could also use tunnels to build subways and presumably Hyperloops. I wonder, though, about why big infrastructure projects so horrendously expensive, in America in particular. There are developers, planners, consultants, politicians, unions, contractors, sub contractors, sub-sub contractors, and they're all out for all they can get. There are lawyers every step of the way. Nobody communicates well. Small hang-ups at the wrong place send the entire operation grinding to a halt. Can this be fixed? Is a big dumb tunnel digging operation the point of entry for infrastructure construction disruption in America? That would be nice. We're good at moving the bits around. We suck at atoms.

Or is Musk setting himself up to be the next Robert Moses? Someone with the capability to move mountains, but not terribly discriminating about whether or not a given mountain should be moved. Musk doesn't talk like someone who's ever had a discussion with an Urban planner about the real reasons as to why our transportation problems are problems.

US infrastructure construction costs are among the highest in the world. There appear to be multiple root causes including: extensive legal protections for private property owners, environment reviews and lawsuits, need to work across multiple government boundaries, labor laws, union work rules.

https://www.bloomberg.com/view/articles/2016-04-08/why-u-s-i...

http://www.citylab.com/work/2014/04/7-reasons-us-infrastruct...

Tunnels will be required on Mars to shield from radiation. Getting a head start on the technology while on Earth only makes sense. Pretty much everything he plans is in perspective of having a dual use: will it be needed on Mars - SpaceX's true goal? I hate to say the word but, it's all about synergies.

If a restaurant started giving out free food, you wouldn't be surprised when more people than usual went to eat there.

Induced demand only happens when the road is underpriced. Any road can be made immune to induced demand by tolling at a price that reflects demand. Too much traffic? Raise the price. Looking a bit empty? Lower the price.

Car dependence is only common in the US because drivers do not pay the true cost of driving. Many people could not afford to drive if they had to pay market price to use roads and parking. By disconnecting the cost of transportation from its use, we have evolved a society that is overly dependent on transportation, and that is overly sensitive to increases in transportation costs. This applies to all modes - not just cars.

Then why build roads at all? If N + 1 means more traffic than surely N - 1 means less? So 0 roads, no traffic! Seriously though isn't this a related rates problem? Adding roads adds capacity and also induces demand. Are the two in lockstep? Can the capacity added be greater than the demand induced?

Not only that, but the more hyperloops we have, the more overburdened they will be.

</s>

A tunnel doesn't have to specifically be for cars. Could run busses, trains, etc.. More electric cars on the road though wouldn't output more pollution around the road.

Electric cars are a good match for tunnels: ventilating the tunnel enough for a whole lot of combustion engines to use is a real cost.

And if the tunnel was limited to electric cars (likely), then that would boost Tesla.

The idea that we shouldn't build infrastructure because it will get used is one of the more silly talking points out there right now.

Infrastructure projects should respond to the density and context of the neighborhood. Don't build a low-throughput highway between high density Manhattan and high density western Queens or it'll immediately become congested. Build a train instead.

A bias toward highways in cities, which are inherently low throughput, has been the norm and also the problem over the last 80 years of urban infrastructure projects.

Perhaps not a problem for Musk who also builds cars.

Induced demand is one of those things that always struck me as an abused concept. Yes, it happens, but it's impossible for it to be always true. New roads will attract more traffic, but only up to a point. There exist roads in the United States that aren't experiencing traffic jams at rush hour. If Induced Demand were a law, that wouldn't be true. So you can't just say that building new roads won't alleviate traffic, period. There are cases where it is true, and there are cases were adding new roads actually make traffic worse, but there are also cases where adding roads (or tunnels) will alleviate traffic.

There has to be a ceiling to how much demand can be induced.

One could view the US Interstate Highway System and its legacy of suburban sprawl as a massive, decades-long effort to reach that ceiling, but we don't appear to have run out of headroom yet.

Presumably the tunnels could also be used for transit, or maybe even walking/biking in some cases.

In economics terms, demand is a relationship between price and quantity demanded.

Increasing the supply doesn't affect demand; it just changes the realized price and quantity to another place in the same demand relationship.

If you know 285 has an extra lane now, you'll go ahead and run that errand at a certain time that perhaps you wouldn't have otherwise. Fascinating stuff.

If Musk had talked with any expert in transportation at all he'd know that building a tunnel highway won't solve traffic problems in the long term.

Either: A) He hasn't talked with any experts at all. B) He doesn't believe the experts. or C) He doesn't think inducing demand is a problem because he owns a car company.

None of these possibilities are very good looks.

> C) He doesn't think inducing demand is a problem because he owns a car company.

Just think how bad the smog would be on a car tunnel. Unless they limit it to zero emission vehicles....

If you move into a bigger house, your hallways will not fill up with more traffic. Currently humans are required at least 1 per car. Thus the housing and employment of those humans is more of a limiting factor on road demand, or perhaps car ownership or desire to drive somewhere, more so than availability of roads.

Your hallways may not fill up with people, but all your new closets, cupboards, wardrobes, drawers and whatnot will fill up with "stuff"

If roads are clearer, then people tend to move farther from work, which increases the number of cars on the road at any given time. Then the roads clog up again. Increased highway construction has been enabling people to move farther and farther out into exurbs (which people think will be nice places to live, and cheaper, but that turn out create rather shabby quality of life compared to walkable cities).

"(which people think will be nice places to live, and cheaper, but that turn out create rather shabby quality of life compared to walkable cities)."

"Shabby quality of life" by whose standards? The people who actually live there? I think not.

If life in the suburbs actually sucked, people simply wouldn't stay there. It's not like they're given a one-way exit visa when they leave the city. They could move back any time they wanted. But, as a group, they never do. The reason they don't is because they find the amenities found in the suburbs (in specific, safety, cleanliness, room for their possessions, privacy, and better schools) preferable to those found in the cities. "Preferable" here means preferable by their standards, which are not the same as those of a single 22 year old programmer, or a single 30-something art connoisseur, or...

Reply on news.ycombinator.com