What We're Really Talking About When We Talk About "Induced Demand"

A week or so ago, I tweeted out the results of an ABC News poll finding that most Americans support investing in transit, rather than roads, as a tool to “reduce traffic congestion.” Almost immediately, a few wiseguys on Twitter piped up to say that neither transit nor highways “reduce traffic congestion” since shifting drivers to transit or adding new highway lanes simply creates space on the newly uncongested highway for more drivers, causing congestion to return. Induced demand!

(An excellent primer on induced demand from, who else, Todd Litman can be found here.)

Now, if I were conducting the poll, I would have phrased the question differently, not just because of the induced demand issue, but also because there are many better reasons to invest in transit than reducing traffic congestion. But, coupled with this story in Vox about the failure of a recent Los Angeles highway expansion to reduce travel times, it seems as though the relationship between transportation investment and congestion is having its moment in the sun. So let’s take a moment, while the sun is still shining, to tease out a couple of issues related to induced demand.

1.      Induced demand has important political ramifications.

David Levinson at the Transportationist posted a reader response today to an earlier post by David A. King that dove into the long history of academic study of induced demand. The reader wrote, “I hate the idea that induced demand is something bad; something to be avoided.” Indeed, when we make large public investments in infrastructure, we usually think it’s a good thing – not a problem – when that infrastructure is used to its maximum capacity.

But there is one group of people for whom increased usage of a particular transportation route is unequivocally a bad thing: incumbent users. And it just so happens that those incumbent users – whether they are truckers or armies of fed-up commuters – have historically provided much of the political juice (or at least political cover) for transportation infrastructure projects.

If I am a suburban commuter, I am not going to be motivated to endure months or years of construction-related delays and higher gas taxes or tolls in order to make someone else’s commute faster, put more trucks on the road, make a suburban real estate developer rich, or allow for the construction of a shopping mall with a restaurant that I might one day drive to with my family. I want my own commute to get better, and quick.

By and large, this is how highway expansion projects are sold to the public: as medium-term fixes for hellacious commutes. And that is why revelations, like the one we made in our recent Highway Boondoggles report about Dallas’ proposed Trinity Parkway, that billion dollar-plus highway projects won’t actually reduce congestion much, tend to leave project proponents speechless. They understand that it is much easier to sell the public on investing money for perceived short-term gain than it is to talk about the myriad costs and benefits of proposed transportation investments.

2. We only care about induced demand when it comes to cars. It’s important to understand why.

Induced demand is not just a phenomenon with highways. Any transportation investment that reduces the monetary, travel time or others costs of travel, or that provides improved access to a new portion of a metropolitan area, thereby opening it for development, has the potential to induce demand. This is why it usually takes a while for ridership of a new transit service to develop – individuals, businesses and developers need time to adjust their behavior to the existence of a new transportation option. Induced demand is also pretty much the entire raison d’etre of investments in streetcars, which are sold almost entirely as tools to attract investment and spur economic development.

Again, when it comes to investments in transit, like most investments in public infrastructure, we usually see high levels of demand as a good thing, at least up to a certain point. We mainly measure whether a transit line is succeeding or failing – rightly or wrongly – by whether it is meeting our expectations for ridership.

So why does induced demand bother us when it comes to highways but not when it comes to transit? In most cases,[1] the marginal cost of serving an additional transit rider is near-zero, while the costs imposed by an additional vehicle – in congestion, pollution, crash risk, etc. – are significant. Failing to incorporate those additional costs into our benefit-cost assessment of highway expansion projects skews the outcome, leading us to think that highway expansion projects are more beneficial, or less harmful, than they actually are.

If highway expansion projects don’t reduce congestion for incumbent users, and if the marginal societal costs of additional driving cut into or outweigh the benefits of improved mobility, then the debate around a given highway expansion project can be reduced to the following: do we want to enable more automobile-oriented development, and, if so, at what cost?

In many places, the answers to those questions might be, “hell yeah” and “just about any.” The important thing, however, is to realize that this latter conversation – about the type of communities we want to build and the economic development pathway we want to follow – is the one we really ought to be having. In Dallas, the debate over the Trinity Parkway seems to have turned in this direction, becoming more about the city’s future writ large than about the details of traffic management. And that’s a good thing.  

The arguments we have about induced demand, in short, are usually arguments about other things: most importantly, about the kinds of transportation investments that will best bring economic prosperity, improved environmental and public health, and better quality of life in the future. Induced demand is an important phenomenon to understand, but a dangerous one to over-interpret. And, as with everything related to transportation these days, we need to continually reevaluate old assumptions to make sure they still make sense in a rapidly changing world.


[1] In cases where transit capacity is strained – see, increasingly, the core subway systems of cities like Boston, New York and Washington, D.C. – the marginal cost of adding new transit riders is non-zero. But we rarely hear calls for public policy interventions to “get people off the subway.”