[AnthonyMouse]

> Additional highways are at-best a stop-gap for day-to-day traffic, never a solution, due to induced demand

This continues to be wrong every time someone brings it up.

If you have insufficient road capacity, you have congestion, and congestion suppresses demand. If you increase capacity, some of the congestion goes away, and then some of the demand comes back.

What this looks like is that you currently have enough cars to require three lanes but have two lanes, so you build a third lane. The reduction in congestion causes you to have enough cars to require four lanes, leading to the fool's conclusion that adding enough lanes is impossible. But that's not it. It's that you needed four from the beginning to handle the amount of traffic that occurs there in the absence of congestion, but you only had three, or two.

Sometimes building a four (or five or six) lane highway isn't the best solution. Sometimes it's better to build more housing near the jobs so people have shorter commutes, or build mass transit etc.

Sometimes you just need a wider road. Pretending that's never the case is preposterous. If that was true then why do we keep multi-lane highways open instead of closing all but one of the lanes? Wouldn't that improve traffic, under this theory?

[azylman]

> Sometimes you just need a wider road. Pretending that's never the case is preposterous. If that was true then why do we keep multi-lane highways open instead of closing all but one of the lanes? Wouldn't that improve traffic, under this theory?

You're setting up this strawman where the argument is "improve roads" vs. "do nothing". That's obviously not the case. The argument is "improve roads" vs. "improve public transit". Demonstrably, improving roads is worse than improving public transit. You refer to this as a "fool's conclusion" yet this has been a well-known fact in the field for almost a century. The wikipedia article I linked has some good information on this if you'd like to learn more.

[AnthonyMouse]

> Your setting up this strawman where the argument is "improve roads" vs. "do nothing".

Your claim is this:

> Additional highways are at-best a stop-gap for day-to-day traffic, never a solution, due to induced demand

That claim is false and is not a straw man because you actually claim it.

Improving mass transit might work as an alternate solution, sometimes, in specific contexts.

That doesn't prove that adding more lanes wouldn't also work, and it's also not universally true.

A large fraction of the traffic on I-95 is trucks. How many semi truck drivers and their loads can you fit on a public bus?

Many highways are congested at a specific choke point. You could make a completely free thousand mile an hour bullet train to transport people from one side of the choke point to the other and solve nothing because people would get to the other side without a car and be unable to get the last ten miles to their destination. But once you get past the choke point, the traffic diverges in every direction and there is no longer enough density to justify a mass transit route.

Sometimes you just need a wider road.

[azylman]

Maybe try quoting the entirety of what I said?

> Additional highways are at-best a stop-gap for day-to-day traffic, never a solution, due to induced demand [1]. You really need large-scale investments in public transportation for this.

Clearly "improve roads" vs. "improve public transit"...

> How many semi truck drivers and their loads can you fit on a public bus?

You're again arguing against something no one ever said. No one suggested that we should just remove all semi-trucks and replace them with buses. Again, we're discussing where to allocate incremental improvements to existing systems. No one is suggesting doing nothing or, worse, shutting down existing systems.

Using your specific example of semi-trucks, moving more traffic (such as daily commute) to rail lines or buses can actually help semi-trucks as well, by freeing up road capacity for things that actually need it. And additionally, freight trains already make up a fairly large percentage of our freight network (~30%) so rail is actually a great alternative to semi-trucks in many cases.

[AnthonyMouse]

> Clearly "improve roads" vs. "improve public transit"...

You: Cars are never a solution because they can't go faster than 15 MPH. You really need horses for this.

Me: Cars can go faster than 15 MPH in many cases. Horses can't be used to transport industrial boilers and such.

You: Clearly you missed the part about the horses.

> No one suggested that we should just remove all semi-trucks and replace them with buses.

You have a two lane road that needs to be a four lane road to handle the amount of traffic it would have without congestion.

If more than half of the traffic that would occur without congestion is trucks, you physically cannot relieve the congestion with mass transit, because relieving the congestion would require removing more than 100% of the non-truck traffic.

> Moving more traffic to rail lines or buses can actually help semi-trucks as well, by freeing up road capacity for things that actually need it.

This the other stupidity with induced demand. It's not induced, it's suppressed by congestion, which means that any alternative means of relieving the congestion will also restore the demand.

Suppose you actually built mass transit and removed the equivalent of one lane worth of traffic from the road. Now you still need to add the other lane because the reduction in traffic congestion restored demand for the road and offset what was removed by the improved mass transit.

[asoneth]

Just to be clear, the "induced demand" that concerns me is not the latent demand of a few deferred trips being taken in the days/weeks/months/year after the road is widened, it's the long-term generated demand of people choosing to move further into the suburbs because they can commute more miles in the same number of minutes. The cumulative result is that property values on the periphery of the commuting zone increase and within a decade or so the highway traffic exceeds the optimal capacity again.

Some regions have concluded that adding a lane per decade is sustainable and already have highways more than a dozen lanes wide. I'm curious to see where the upper bound is.

(Personally I think dynamic pricing to maintain optimal highway capacity is a more sustainable approach.)

[AnthonyMouse]

> Just to be clear, the "induced demand" that concerns me is not the latent demand of a few deferred trips being taken in the days/weeks/months/year after the road is widened, it's the long-term generated demand of people choosing to move further into the suburbs because they can commute more miles in the same number of minutes. The cumulative result is that property values on the periphery of the commuting zone increase and within a decade or so the highway traffic exceeds the optimal capacity again.

Property values increasing there actually offsets the problem by making it less desirable to live there.

The real trouble is that people build more houses there. But the reason people build more houses there, and suffer a 30 minute commute (which more congestion might have turned into a 60 minute commute), is that they can't afford to live in the place with a 15 minute commute. Typically because zoning prohibits building more housing there.

Now let's see what our choices are here.

We can do nothing at all. Well, now people are screwed. They still need somewhere to live, the place that now has a 60 minute commute is the only place housing can be built, so the housing still gets built there, but now the commute is longer. That's just horrible and helps no one.

Second, we could widen the road and that's it. The new housing still gets built in the suburbs but at least now people waste less time in their cars.

Third, we could loosen the zoning so higher density housing can be built closer to the city, but not widen the road. This is pretty good, because now the people who live in the new housing get the 15 minute commute. But the people who already live in the suburbs are still stuck with the 60 minute commute.

Fourth, we could loosen the zoning and widen the road. Then new housing gets built in the city instead of the suburbs, because people prefer a 15 minute commute to a 30 minute commute, but the people who already live in the suburbs still get a 30 minute commute instead of a 60 minute commute because of the wider road. And it stays that way because the new housing is getting built in the city instead of the suburbs. This is pretty obviously the one that we want.

[asoneth]

> but the people who already live in the suburbs still get a 30 minute commute instead of a 60 minute commute because of the wider road

That is true for several years after the highway widening. Which is good for politicians who operate one election cycle at a time.

But the population doesn't just sit back and enjoy shorter commutes after a highway widening project. The towns that used to be two hours outside the city are now only an hour drive away and commuters looking for a deal start to move there. The towns that used to be an hour away are now only a half-hour drive to the city and are now hotter as well. Some of that is due to a game of musical chairs in which people expand out of the city to the suburbs, some of that is new residents choosing to live further out than they would have if they arrived before the highway widening. All of this increases the number of vehicle miles traveled. Eventually (over the span of decades) this results in a new equilibrium in which the highway is nearly as congested as it was before the widening and everyone is once again living at about the time-limit of how far they are willing to commute.

It's not all bad of course because the new highway helps grow the metro area and it increases property values, especially in the suburbs that were on the periphery of the commuting zones. But to permanently eliminate congestion you'd either have to keep widening the road every decade or two to handle the growth (Houston, Dallas), allow the congestion itself to act as a natural limit (NJ), or institute tolls/congestion pricing (Singapore, London).

[AnthonyMouse]

> But the population doesn't just sit back and enjoy shorter commutes after a highway widening project.

Notice that this is true of anything that relieves the highway congestion in any way whatsoever. As soon as the road is clear, no matter why, the commute is shorter and it becomes more attractive to live further away in distance.

> Eventually (over the span of decades) this results in a new equilibrium in which the highway is nearly as congested as it was before the widening and everyone is once again living at about the time-limit of how far they are willing to commute.

The way out of this is to make some alternative to increasing sprawl more attractive than increasing sprawl. Relaxing zoning rules to allow higher density is an obvious one, because people would rather have a 15 minute commute than a 30 minute commute, so they'll prefer to build housing where it's a 15 minute commute unless that's prohibited by law. When it is, they have to build further away and you get more sprawl.

But also notice that you're assuming population growth.

If the population was stable, and you built a road which is sufficient now, it'll probably stay sufficient as long as the population remains stable, because who is building enough new housing to move the needle on traffic in a city that isn't growing?

If the city's population is growing, continued population growth will require you to expand highways and such over time in proportion to the population. What else would you expect? The only alternative is intentional scarcity.

[asoneth]

> Notice that this is true of anything that relieves the highway congestion in any way whatsoever.

Is that true if we charge an appropriate price for road use? Isn't that what we do to manage availability of every other scarce resource to ensure it is being used optimally?

If you increase the price of using the road at the busiest times then eventually you will arrive at a price that maintains the optimum flow rate. You may need to adjust the price occasionally to track shifts in inflation and population, it has equity issues, and (like most sustainable solutions) it is politically difficult. But I don't understand what would cause it to stop working. In the places it has been implemented (e.g. Singapore, London, Stockholm) it is generally unpopular at first and then extremely popular after a year or so. And administering it should become cheaper as technology improves which may make it feasible for smaller cities as well.

> If the population was stable, and you built a road which is sufficient now, it'll probably stay sufficient as long as the population remains stable, because who is building enough new housing to move the needle on traffic in a city that isn't growing?

Even in metro areas that have stable populations, heavily subsidizing transportation infrastructure (whether transit or highways) between the city and the suburbs often has the effect of slowly shifting the current metro population outward to those suburbs. By subsidizing suburban commuters you are making living in the suburbs more attractive than it would be otherwise. As you make it more attractive, more people who currently live in the city will rationally choose to move to the suburbs.

I'm not saying it's a bad idea to build more lanes, just that on it's own it is not a sustainable approach. As long as rent/land prices are largely set by the market and roadways are free or heavily subsidized then it's not surprising that people take advantage of that and you end up with a shortage of road space.

(I agree with your comments on zoning.)

[_dain_]

You're missing half the story; "demand" only makes sense with reference to a price. At the moment, that price is zero, but it needn't be. As we all know, market prices are an efficient mechanism to allocate scarce resources. People have a curious blindspot about this when it comes to roads.

[AnthonyMouse]

> As we all know, market prices are an efficient mechanism to allocate scarce resources. People have a curious blindspot about this when it comes to roads.

It's not a blind spot. It's a characteristic of services with a high fixed cost and trivial variable cost.

To ask if we should price roads is to ask if the price needed to deter usage enough to relieve congestion without expanding the road would generate less revenue than it would take to expand the road. But this is basically never the case because most of the expansion cost is one-time (e.g. buying the land) whereas the congestion charge would have to be collected forever to continue deterring usage.

The strongest case for not expanding the road is if there is a more efficient way to relieve congestion, e.g. by relaxing zoning restrictions to allow higher density housing and reduce travel distances.

But if people value using the road at more than the cost of expanding it, and there are no higher efficiency alternatives, that implies the road should be expanded. And once it has been and there is no congestion even at zero unit price, there is no benefit in charging a unit price to deter congestion that isn't there anyway, and a detriment in deterring use of a public resource for which the same fixed cost has to be paid whether you use it or not.

Charging road tolls is also especially inefficient because the collections process has a high administrative overhead and a high privacy cost. Every dollar spent collecting tolls -- toll tags, gantries, billing, maintenance, customer service -- is a deadweight economic loss not incurred by any alternative that doesn't require them. The privacy cost is the same.

[_dain_]

>To ask if we should price roads is to ask if the price needed to deter usage enough to relieve congestion without expanding the road would generate less revenue than it would take to expand the road. But this is basically never the case because most of the expansion cost is one-time (e.g. buying the land) whereas the congestion charge would have to be collected forever to continue deterring usage.

First of all, that just doesn't follow. You can take out a loan to pay for the up-front cost and pay it back using the revenue you collect from tolls. Second, roads cost a lot of money to maintain. There are the usual ongoing costs to fix potholes and so on, and then they have to be totally replaced after 25-30 years. This is far more than what local taxation can bear in many cases. Replacement costs are chronically underestimated.

(e.g. Winnipeg would have to raise taxes 95% to properly fund their existing road liabilities: https://www.strongtowns.org/journal/2021/1/3/death-of-a-car-... )

>But if people value using the road at more than the cost of expanding it, and there are no higher efficiency alternatives, that implies the road should be expanded.

This is wrong, they are not valuing it properly because it is paid for by taxation and debt. A motorist pays exactly the same amount directly as a non-motorist: zero. The true costs are diffuse, invisible, and incomplete.

>And once it has been and there is no congestion even at zero unit price, there is no benefit in charging a unit price to deter congestion that isn't there anyway, and a detriment in deterring use of a public resource for which the same fixed cost has to be paid whether you use it or not.

This situation hardly ever happens. There is almost always more congestion after expansions than predicted by planners. If you wanted to overpower induced demand and get rid of all congestion the roads would have to be utterly gargantuan.

>Charging road tolls is also especially inefficient because the collections process has a high administrative overhead and a high privacy cost. Every dollar spent collecting tolls -- toll tags, gantries, billing, maintenance, customer service -- is a deadweight economic loss not incurred by any alternative that doesn't require them. The privacy cost is the same.

Time spent stuck in traffic is also deadweight loss, and it creates pollution.

[AnthonyMouse]

> You can take out a loan to pay for the up-front cost and pay it back using the revenue you collect from tolls.

The fact that you can do this is the point. It means the value of expanding the road is more than the cost of expanding the road, which implies it should be done absent some better non-toll alternative like increasing housing density.

But once you have enough capacity that there is no congestion without congestion charges, it's inefficient to deter use of the sunk cost road.

> There are the usual ongoing costs to fix potholes and so on, and then they have to be totally replaced after 25-30 years.

These costs aren't linear in the number of lanes. Resurfacing a four lane highway doesn't cost four times more than resurfacing a one lane highway (or your contractors are ripping you off).

Moreover, the initial cost is typically the highest, because you have to acquire land and possibly rebuild bridges and overpasses the first time.

> This is far more than what local taxation can bear in many cases. Replacement costs are chronically underestimated.

This applies to roads to nowhere that are under-utilized but still have to be maintained. Anything with traffic congestion is seeing more use than its cost.

You also have problems with corruption in government contracting inflating the cost of everything, but that's a separate problem and applies equally to mass transit etc.

> This is wrong, they are not valuing it properly because it is paid for by taxation and debt. A motorist pays exactly the same amount directly as a non-motorist: zero.

It's not a question of what they're actually paying, it's a question of what they would be willing to pay, i.e. the value they assign to the road. If you put a toll somewhere there is congestion, would the toll pay enough to expand the road? The answer is almost always yes, which implies that that the road should be expanded when congestion exists, unless there is some more efficient alternative to relieve the congestion.

It shows that expanding the road is more valuable than deterring usage with tolls. The possibility remains that some non-toll method of relieving congestion is still better than expanding the road, but it shows that expanding the road is better than deterring usage with tolls.

And whether to expand the road is a separate question from whether to actually fund the expansion from the tolls, because toll collection is inefficient and privacy invasive the deterrence function is undesired when the congestion can be relieved without it. The toll being able to fund the road proves that people value the road at more than its cost, but it's not the most efficient way to fund it.

> If you wanted to overpower induced demand and get rid of all congestion the roads would have to be utterly gargantuan.

This is only the case if you for some reason insist on using road expansion as the only solution to relieve congestion.

Mass transit can relieve congestion in higher density areas. Zoning that allows higher density housing to be built near jobs relieves congestion by both reducing the distances people have to travel and making mass transit more efficient.

Expanding roads works where those don't, or in combination with them. For example, if you have a growing city with restrictive zoning and a congested road between a large suburb and the city center, what you want to do is cause new housing to be built closer to the city instead of further expanding the suburbs, but the road to the suburbs still needs to be expanded because it is already too small for the existing traffic from housing which is already there and is not about to be removed.

This is politics. The most effective solution involves allowing higher density housing near urban areas, but this is not the solution desired by existing land owners, so they push inefficient alternatives like tolls. Because that increases rather than decreases property values closer to the city center by making it more expensive to live further away without increasing the supply of housing that isn't further away, so existing land owners collect higher rents to the detriment of residents.

> Time spent stuck in traffic is also deadweight loss, and it creates pollution.

Which is why we should relieve congestion by using the alternatives with fewer deadweight losses, like building higher density housing where that works and expanding roads where that works.

[tshaddox]

But there is some limit to the number of lanes you can add, even theoretically from a topological perspective, but more imminently from a practical standpoint of limited budgets and ability to tear down existing non-road infrastructure.

[AnthonyMouse]

The theoretical limit is irrelevant. It's like saying you can't always improve emergency response time because of the speed of light. Nobody is really up against the theoretical limit.

The practical limits are all trade offs. How much does it cost to add two lanes? How much does it cost to maintain low ridership bus service to low density suburbs? There are circumstances in which adding more lanes is the best available alternative.