Saturday, July 12, 2014

Is induced demand really about road capacity?

Induced demand is a common warning from urbanists whenever a new road is proposed or an existing one is widened. The basic idea is that adding road capacity will generate new traffic that will recreate the congestion that justified the additional road capacity anyway. Some have concluded that any additional road or street will always induce demand.

I've seen a case of this recently when Laval, a big suburb of Montréal (and very car-centric) proposed to build a boulevard across a small wooden area. Many people argued against it, saying that this was proposed only by car-obsessed planners and that it would just result in induced demand. Here is an image of Laval's street/road/highway grid, with the proposed section in red.
The proposed new street in context
I think this is a mistake. People have taken specific cases where increasing capacity increases traffic and assumed it is an universal truth.

I think that, essentially, induced demand is not about capacity per se, but about speed.

Speed, watershed and congestion

The point is that speed if perhaps the single most important factor that determines what transport mode dominates. You can have a free transport mode that is supremely comfortable and pleasant, but if it crawls around at 1 km/h, no one would take it. That's why highways are so disruptive in urban areas, they go at speeds that no other mode can approach. Yes, highways are effectively the fastest urban transport mode, being able to maintain speeds of 70 to 100 km/h (45 to 60 mph) without stopping for kilometers? Walking is 5 km/h, bikes are 15 km/h, buses vary between 10 and 20 km/h on average, cars on local streets vary between 20 and 30 km/h. Even rapid transit averages 30 to 40 km/h.

People also have a certain tolerance for distances, but not in kilometers or miles, but in minutes. So if people, say, tolerate a 15-minute travel distance to reach a destination, depending on the average speed of the street they drive on, they will want to inhabit a certain area. To make an analogy, the following figures are the "watershed" of the roads, just like water from a watershed flowing to a river by gravity.

Watersheds of an urban street, a suburban stroad and an highway
In the previous figure, you can see the "watershed" of different kinds of streets and roads, one an urban street, max speed 40-50 km/h (25-30 mph) with many stops or traffic lights, the other a suburban stroad (mix between street and road), max speed 70 km/h (45 mph) with few intersections spread apart, the last an highway (as in limited-access motorway/expressway) with a max speed of 100 km/h (60 mph). With the highway, people will be willing to live much, much farther away, in places where the only viable mode of transport is a car on the highway.

From experience, I can tell you that "induced demand" look a lot like "induced development" in lands once thought too "far", but that are brought closer to points of interest (jobs, shopping malls, etc...) through a faster road opening.

Where road capacity enters the picture is when congestion rears its head. Look at the watershed of the highway versus the urban street, it's ginormous. If the population density is the same in the watershed area, the highway could have 3-4 times the demand of the urban street, so it requires way more lanes. When it doesn't, well, congestion occurs, which slows down traffic significantly. Roads that normally allow cars to zoom around at 100 km/h suddenly are full of cars in stop-and-go traffic (or at least, the entry points of the highways are). So the effective speed during peak hours crumbles down, reducing the watershed. People know this and farther areas along the highway become less desirable.

The presence of a chokepoint during peak hour congestion reduces the desirable lands around the highway
In this situation, developments farther than the chokepoint are slowed, if not downright stalled, because the speed of the highway collapses past it. This limits traffic by making developments less likely, or if they occur, they will be "edge city"-type developments, meant to create new points of attraction outside the zone surrounded by chokepoints.

In such a situation, adding a new highway to "relieve" the current highway, or widening the highway may eliminate the chokepoint (for a while) making lands that used to be undesirable desirable again, and restarting developments as lots of land come on the market at one time, making land cheap and thus, housing cheap.

So road capacity isn't really what induces demand, speed is. It's simply that when capacity is too low, congestion happens, reducing the effective speed of roadways. Increasing capacity only induces demand in that case when it restores speed to previous levels.

Why does it matter?

This distinction matters because it means that we don't need to be afraid of capacity in itself or of building new streets with low speeds. Yes, highways are highly disruptive, but a street grid isn't an issue. In fact, many cities with extremely high capacity street grids do not really sprawl far but are instead great multi-modal cities. Vancouver for instance has a very high-capacity street grid but without any highway within the limits of the city itself. Though streets still have unused capacity, recently car use has fallen and other modes of transport have increased. Why? Because on Vancouver's tightly packed street grid, pedestrians, bikes and bus riders aren't that much slower than cars.

The same can be said about Manhattan or Portland.

Return to Sapporo

I've spoken of Sapporo recently, a Japanese cities built on the same street grid of wide streets that Dallas, Austin or Houston were initially built on, yet that has a pretty good record in creating a multi-modal city, with only 45% of all trips in the city and its suburbs taken place in cars. When you look at Sapporo and see the grid of 4-lane streets everywhere, you can be amazed at the road capacity of it all:
Yet, despite this capacity, Sapporo does better than all North American cities for sustainability, save, maybe, New York. What gives? Well, here is an image from Google Maps of Sapporo, with different road categories in different colors:
Sapporo's urban area
The A pin is the downtown area. The highways are in orange, the yellow roads are prefectural (I think) roads, but they're not anything we'd call highways here. They are almost indistinguishable from regular streets, with max speeds of 50, even 40, km/h (30 and 25 mph) and regular intersections rather than interchanges. This is an example of one of them:
One of the roads marked in yellow in the previous image
The only real highway goes around the urban core, making it nearly useless for commuters. Worse, it's tolled, HEAVILY tolled. About 25-30 cents per kilometer driven (40-50 cents per mile). So the remoteness of the highway and the tolls make it much, much less disruptive than our own highways in North America.

The point is that Sapporo's streets have incredibly high capacity, but that doesn't seem to induce demand too much, because speed, not capacity, is what really induces demand. Since Sapporo's streets are low-speed due to low speed limits and frequent intersections, they don't make walking, biking or transit uncompetitive. The only significantly faster road is a heavily-tolled remote highway made for long-distance trips only. The result is pretty good in keeping car mode share down by keeping other modes of transport viable.

Coming back to Laval...

OK, so let's back up to Laval for a while. Let's zoom in on the area in question:
Proposed new street connection in red

Now, the point is, there are already lots of high-speed stroads and highways in Laval. The damage is done. The streets are often poorly connected and uses are kept separate, but even there, they've started building higher density residential areas, especially since the subway was opened (the orange line on the image). The proposed street would connect two existing streets to make one unbroken street at reasonable speeds (50 km/h). So it's not an high-speed road, the danger of induced demand is not really there, especially that there are plenty of high-speed roads around, so, a bit too late for that.

The main effect I see is that it would end a detour of 1,2 km by the south (0,75 mile). This has little impact for car drivers, but for pedestrians, cyclists and transit riders, this is a significant deal. This detour to go to the schools to the left, or the shopping center, costs them 15 minutes (for pedestrians), 5 minutes (for cyclists) and 3-5 minutes (for buses). The transit operator will now be able to run buses along the entire street. So it's a massive deal for active transport, much bigger than for cars. It's a good step towards finishing a street grid rather than relying on canalizing all the traffic on a few stroads and highways, and a very good idea, especially if they build a parallel bike lane while they're at it.


So induced demand, it's real, but it's about speed, not capacity per se. It appears to be about capacity only because high-speed roads attract a lot of users, more than they can support, the result is congestion which effectively reduces their speed. So we shouldn't worry about highly porous street grids, even if they have extremely high capacities, as long as the streets themselves are calmed and slow.


  1. Do you think it would get more local support if the new connection was for transit and cyclists and pedestrian only?

    I guess you could argue that with drivers not having to take the detour to Boul de la Concorde, that might mean less congestion slowing things down and drivers from further away that previously weren't using it before starting to use Boul. de la Concorde. That assumes congestion was bad enough to slow things down in the first place.

    These stroads tend not to have congestion in the same way highways do though, any problem they do have is related to too many cars queued up at red lights. Either too many queued up to make it through one light cycle (especially for left turns) or too many queued up to fit between the first light and the one behind it. Is that an issue in Laval?

  2. The important point here is about better connecting and completing the street grid. It reduces strain on surrounding streets and major thoroughfares, which could have a minimal impact on induced demand I suppose, but it's not like this is some mega arterial that the whole metro area drives on. Better connectivity for everyone, especially pedestrians, cyclists, and transit users is a laudable goal. Citing induced demand in a situation like this just sounds like NIMBYs grasping for straws. Of course if they plan to build this connector 5 lanes wide with 12 foot lanes, then I'd probably be fighting against it too.