Ignoring Induced Demand is Engineering Malpractice

Leesburg, VA. (Source: Wikimedia Commons.)

The standard approach for engineering projects is well-established. A team of technical experts determines that a project needs to happen. They get whatever authorization they need to move forward in evaluating options. They put together options, most frequently three of them. The first is the “do nothing” approach, the hated status quo that most everyone agrees is intolerable. The second is some lavish, over-the-top approach that seems ridiculous. And the third approach is the baby bear; not too big and not too small, in contrast with the others it will seem to be just right.

This is all part of a sales job, including the anchoring effect of juxtapositioning the professional’s preferred alternative between two unappealing options. I have called these “dumb, dumber, and dumbest” on occasion. Engineers and project planners use the public comment process to go through engagement theater and simulate that they actually care about the feedback they are receiving, earnestly considering it instead of what they are really doing: working through the process to overcome opposition so they can move on to the building phase of the project (because that is the favorite phase).

I’m not saying anything deeply insightful. Every major organization or department has an expert, or team of experts, who are really good at shepherding projects through this kind of process to deliver results. 

Recently, a Strong Towns member shared the professional responses to citizen inquiries prepared as part of a highway building project near Leesburg, a town in Loudoun County, Virginia. (Note that the mayor of Leesburg and at least one councilman are highly skeptical of the widening project, but they have little say in this countywide proposal.) 

View the full pdf here.

This was a lengthy document that was painful to read, but one answer stood out as being particularly awful:

Question: Explain how induced demand is relevant or not relevant to the potential widening of US Route 15, as proposed by Concept B.

Staff Response: The concept of induced demand suggests that adding capacity to a roadway creates or “induces” new demand for a roadway where demand did not previously exist. However, travel demand is not generated by the addition of network capacity, rather travel demand is created by existing and future jobs and household population produced trips. The distribution of trips produced by the jobs and population is then distributed on the available network based on a gravity model where travel time via various routes drives the distribution of trips. Certainly, the addition of capacity on a roadway could result in increased volumes on that roadway, but it also may reduce volumes on parallel roadways. Population and employment growth will also drive increased travel volumes.

When I read this response, I was kind of dumbfounded. To say this is antiquated thinking is being generous. It’s lacking in thought, introspection, and rigor—unless they don’t believe it, which is very possible. And disturbing.

Let’s break it down line by line:

The concept of induced demand suggests that adding capacity to a roadway creates or “induces” new demand for a roadway where demand did not previously exist.

This is not what induced demand suggests. It’s not that people or cars magically appear where none existed before, all other things remaining static. It means that trips that otherwise would not have occurred are now going to happen solely as a result of the added capacity. That’s the “induced” part of it.

Let me give a simple example: I’m cooking dinner and I’m getting low on milk. Prior to the new capacity, it’s a twenty-minute round trip and, because the kids are expecting dinner, we drink water instead of milk. I get more milk the next time I’m passing by the grocery.

Now, the planners and engineers add more capacity to the road network and suddenly it’s only a ten-minute round trip to get milk. Now I decide to go because I can get there and back in time to serve dinner. Nothing has changed in the overall environment, but I’m taking more trips than I otherwise would because trips are now more convenient to take (at least for a while). 

Adding capacity induces a trip that otherwise would not have occurred. That’s induced demand.

This example is the simplest form of inducing traffic, but just a tiny bit of brainpower will produce an infinite number of examples of people making decisions to take trips when doing so becomes easier and, conversely, reducing the number of trips they take when doing so becomes more difficult. 

However, travel demand is not generated by the addition of network capacity…

The concept of “travel demand” is where traffic engineers have stunted their own intellectual development more than perhaps anywhere else. And they’ve done so for two reasons. First, it makes their models easier to run. It’s really difficult (impossible, really) to create models that factor in the behavioral responses of humans. Better to just assume a static level of demand, even though that assumption is a farce (remember, traffic models are all about justifying projects, not actually modeling what is going on in the world).

Second, it allows traffic planners and engineers to position themselves and their craft as responding to demand, not creating it. That’s an important distinction because it allows them to be confident in what they do without having to struggle with the underlying reasons that things aren’t working. 

I’m not belittling them with that observation. Simplifying complexity is a reality that can be observed in many practices that, in the modern economy, have become simply about production. Economics, education, health care, etc.

Engineering in the auto age is about building—build, build, build—and not about optimizing or managing systems. When your ethos is merely to build more stuff, you develop myths and models that support that ethos. That’s what you’re seeing in the patently absurd assertion that additional capacity does not generate more trips.

View north along U.S. Route 15 in Leesburg, Loudoun County, VA. (Source: Wikimedia Commons.)

…rather travel demand is created by existing and future jobs and household population produced trips.

This is so incoherent that it can be read in two different ways, both absurd. 

The first is that they are suggesting it is the choices people make that impact demand, not the added capacity the engineer is building. In other words, the engineer builds the highway but nobody forces people to drive on it more; that is their choice and we engineers can’t control it. If this is what they mean here, we basically have second-grade logic being applied to traffic analysis. I hope that is not what they are suggesting.

More likely, the engineers here are suggesting that additional traffic in the future is going to be caused by new homes and businesses opening up, which they would see as a zoning or land use decision that has no relation to their project. Of course, any highway expansion project does influence land use—good grief, visit any American city or look at any zoning map in North America to see how self evident that is—but, for the sake of discussion, ponder a scenario with no changes in land use. There is still induced demand.

I’m deciding whether to go to the local hardware store six blocks from my house or the Home Depot out on the edge of town. I can get to the local store easy enough, but it’s a little bit of a drive to get to the big box. Now, expand that roadway and make that distant trip easier and more of my hardware trips will be out to the edge. Is this scenario, and an endless number of scenarios just like it, so difficult to imagine?

That’s in a static land use model. What isn’t acknowledged is how these projects induce changes to that model, even when nothing else is added. Just follow the hardware store example. With the expanded highway, it’s now easier for Home Depot to bring their supplies in on a massive truck, using the subsidy of continued highway expansion to grow their retail network and drive down costs. In and of itself, that shifts demand from short, neighborhood trips to the local hardware store to distant, regional trips to the big box.

Engineers and others might argue that outcome (i.e., the big box store winning a competitive battle over the local hardware store) is a net good. Okay, but that’s not the technical response they are making here. The professional assertion they are making is that this kind of induced response doesn’t happen, that nobody already in the system alters or changes their behavior in response to the new construction. That’s nonsense.

Now, to continue with this example, on top of these obvious shifts in driving behavior, add all the chain restaurants and strip mall stores that thrive in the same environment of a Home Depot and the result is a radical shift in traffic demand within the existing static model, all before anyone has been induced to relocate their business or put in a new subdivision on the commuter outskirts of town. 

The distribution of trips produced by the jobs and population is then distributed on the available network based on a gravity model where travel time via various routes drives the distribution of trips.

Transportation is getting worse and costing more. We have to do better.

Learn how with the Strong Towns Academy course, “Aligning Transportation with a Strong Towns Approach” (worth 9 AICP continuing education credits).

According to this response, any change in trip behavior that occurs after the project can only be a result of new homes or businesses opening up, not on any prior traveler in the system increasing or even changing the route that they drive. That statement is either fraud or a case of intentional ignorance; I’m not sure which, but professionals with either mentality should not be allowed to plan and direct projects like this.

They also suggest that, where and when those trips happen, they will be based on a “gravity model” distribution, an approach to traffic prediction that is statistically coarse and questionably applicable for large metropolitan areas (it assumes an auto-centric and linear relationship between distance and trips that doesn’t reflect reality in complex, urban environments). For the area around a town the size of Leesburg, Virginia, the coarse and unrefined outputs of the gravity model is, again, an absurdity. Using it in this context reinforces the conclusion that these professionals either (a) don’t know what they are talking about, or (b) believe everyone else doesn’t know enough to question them.

Certainly, the addition of capacity on a roadway could result in increased volumes on that roadway, but it also may reduce volumes on parallel roadways.

Let me rephrase that last sentence: “Okay, forget everything we just said about there being no induced demand from adding additional roadway capacity. There certainly is, but it also might remove trips from other roads. Or not. It’s tricky.”

Throughout this piece, I’ve been hard on the group of professionals that put this together, and rightly so. This is shoddy work, either tragically misinformed or deviously packaged. Either way, we have to do better.

With the massive backlog of routine maintenance and the number of lane miles already in poor condition, the bar to build any new highway capacity should be extremely high. Even without digging deeply into this project, it is clear that the bar is not being met here, especially if the seriousness of these responses is any indication of the thoroughness of the project evaluation.

In 2022, denying how highway expansions induce people to drive more should be considered professional malpractice.

---

---