The Science Is Clear: More Highways Equals More Traffic. Why Are DOTs Still Ignoring It?

Photo: B137/Wikimedia Commons
Photo: B137/Wikimedia Commons

Numerous studies have documented the phenomenon known as induced demand in transportation: Basically, if you build highway lanes, more drivers will come. And yet, transportation agencies rarely account for this effect when planning road projects.

In a recent paper published by the Transportation Research Record, author Ronald Milam and his research team reviewed the various studies documenting the induced demand effect. They found that for every 1 percent increase in highway capacity, traffic increases 0.29 to 1.1 percent in the long term (about five years out), and up to 0.68 percent in the short term (one or two years). One recent study found a one-to-one relationship between new highway lane capacity and traffic increases.

However, highway planners are failing to incorporate this effect into their models. Milam told Streetsblog that “it is rare to find an induced travel analysis in most transportation infrastructure design or environmental impact analysis.” That means transportation agencies are green-lighting money for highway expansions that are destined to become congested again only a short time later.

While some transportation agencies may lack the sophistication to model induced demand accurately, they should not disregard the phenomenon, Milam and his team write.

Highway planners should assume traffic increases will result from highway expansion projects in line with what previous research has demonstrated, the authors conclude.

More recommended reading today: nextSTL publishes a poem about the epic struggle to convince the Missouri Department of Transportation to design a walkable Gravois Street. ATL Urbanist reports MARTA will take over operations of the struggling Atlanta Streetcar. And looks at the relative dearth of retail stores in urban locations in the Twin Cities.

46 thoughts on The Science Is Clear: More Highways Equals More Traffic. Why Are DOTs Still Ignoring It?

  1. I definitely think induced demand is a real thing but how do you prove that the highway actually induced demand as opposed to the highway just taking traffic off of side streets?

  2. In Denver, CO, the widening of I-25 south of Broadway into the Denver Tech Center that was completed about 9 years ago could be a poster child for the induced demand phenomenon. It’s congested almost all the time and it’s truly horrible at rush hour. We had a couple of good years immediately after completion but that also coincided with the Great Recession.

  3. Science? Since they don’t build lanes just to sit there looking pretty, common sense tells me it will induce demand. That’s the whole point isn’t it? Many times the demand is already there just waiting; other times continued growth provides the demand.

  4. Induced demand “theory” should apply to housing and other subjects, and not just directed towards roads. Substantial funds are raised specifically for road transportation, so to not use them in that manner conflicts with the funding premise.
    And, just because congestion may be unavoidable with the movement towards densification – it doesn’t follow that road improvements shouldn’t occur. It’s a huge part of the infrastructure that operates the nation.

  5. Route shifting — for instance, from crowded side streets to a widened highway — is indeed one of the ways that travel is induced. Route shifting is usually accounted for in standard traffic models. One study (Duranton and Turner, 2009) found that route shifting accounts for 0-10 percent of the induced travel effect in a region.

    There are other ways that travel is induced, and some of them aren’t included in standard traffic models. Those include new trips that would otherwise not be made, and in in the long term, changes in land development patterns and population growth caused by the new roadway lanes.

    Governments often spend money on roadways to reduce congestion. But sometimes they’ll spend money and get less bang for the buck than they expected. Sometimes they get no congestion reduction at all. That’s why it’s important to understand induced travel.

    In some cases, the evidence will show that new roads or road widening are poor options compared to alternatives. For example, a huge study by RAND titled “Moving Los Angeles” (2008) found that adding roadway capacity was fundamentally unable to reduce peak congestion. Instead, it recommended roadway pricing strategies, along with improved transit, carpooling, biking and walking, and smarter use of the existing thoroughfare network.

  6. But that doesn’t mean that it doesn’t induce new traffic too though. And it doesn’t justify spending hundreds of millions of dollars to move traffic.

  7. Wait… wouldn’t traffic coming from other roads be new traffic to the freeway? New traffic could also come from growth ofc.

  8. Saying that “demand is induced” is like trying restate “growth following investment in infrastructure” in the most negative way possible. I understand that some people are anti growth but growth does in fact make infrastructure less costly per capita, which is why densification is correlated with availability of frequency of public transit. Why wouldn’t we want new infrastructure we built to get used? It’s better than building a highway out in the middle of nowhere that only gets used by a few vehicles a day.

  9. Yes science. Economics. I think your argument is a bit misleading- demand and congestion is always going to be there. When we build more car capacity, it will always reach capacity (within 4 years or less) because, as induced demand explains, adding capacity creates traffic. If we don’t build it, we don’t create more traffic. The point is that if we don’t build a car lane, but instead build a train line, we reduce traffic congestion. trains, buses, walking, and biking have the ability to carry more passengers.

    When we stop adding more lanes, it means less people are willing to drive less distances and more density can become a reality. If we add more lanes, we continue to sprawl.

  10. In Los Angeles, they added a lane to the 405 at a cost of more than $1 billion, and the day it opened, the road was 1 minute slower than before the project. The city and metro tried to argue “just think if we had not built it”. Yeah- we’d be one minute faster and we’d have $1 billion to spend on a subway over the same pass.

  11. Yes, new traffic to the freeway but not new traffic in general. But expanding that freeway may also induce people that had a shorter commute to entertain the idea of going to places farther away because the traffic on the freeway isn’t as bad as it was before. Or someone might be thinking about moving here, tests out the new freeway during rush hour and says to themselves, “that’s not that bad, I’ll make the move” etc.

  12. The question isn’t “why wouldn’t you want new infrastructure to get used”, it’s “why build new infrastructure”. Especially when it’s not sustainable and that building that new infrastructure requires that you continue to build new infrastructure, in a vicious never ending cycle.

  13. It’s not the infrastructure that’s unsustainable, it’s the gas powered vehicles that are on them. I don’t consider building infrastructure to be vicious. It might be never ending but it is serving a human need.

  14. Some people are pro-growth, some people are anti-growth, but sensible people are more selective and want growth of things that make life better but not of things that make life worse.

  15. Robert Johnston, prof. emeritus at UC Davis, has developed a model that takes induced demand into account when modeling the effectiveness of new freeway capacity.

    For example, see and for more examples, google “Robert Johnston Induced Demand.”

    His model might be a good subject for a streetsblog article.

  16. Actually, building housing closer to transit and job centers should not produce demand. Well-located densification mitigation in CA includes a list of projects which will reduce the VMT of new tenants. AB-743.

  17. One way to stem car growth when new housing is built is to reduce parking dramatically, and make car-owning residents pay for it separately from their unit. People who don’t have a place to park a car are less likely to buy one.

  18. I have read that some companies will do test drives from a few locations to get a sense of time spent getting to their prospective site and while that may have helped Phoenix be #1 in population growth last year it hasn’t seem to hurt Denver which clearly has more congestion than Phoenix.

  19. It seems that your supposition might indeed induce demand, for housing, transit use, schools, etc. Any increase in population would correlate to increased usage of infrastructure. Being closer to jobs & transit might induce less demand than in exurbs, but it would induce demand nonetheless.

  20. Someone already noted how Denver is making major mistakes. So it is no wonder Denver still has issues. Building infrastructure capacity that focuses on transit/bike/ped is just one part of the larger smart growth approach. Another poster brought up parking and that is very important. Induced demand’s basic premise is that demand will always be slightly higher than supply. By adding supply you generate more demand. Decreasing supply decreases demand because people move to alternatives. If there are no alternatives that can match demand, then you have a problem.

    Also, can you provide evidence that 1) Denver has really focused on transit and is less about driving; 2) has Denver increased housing at the job centers? If so, how many units and how many more are needed?

  21. My community’s problem highlights this discussion. I live in a valley that physically separates us from the urban core. We have a north-south freeway (405) to the west that is far too west from that core and so many choose to drive through my community instead of taking the freeway.

    They added a lane to the 405 each way and promised it would save everyone 10 minutes. Induced demand predicted traffic would be slower, and it was. Worse yet, it took 6 years to complete the project and a large number of people moved further out, just north of me, because of the promise that the 405 lane would save them.

    The project completed, the traffic moved 1 minute slower, and now more people drive through my community over the canyons. We get 80,000 cars going over the canyon daily, those roads were designed to take 10,000 daily. Anyone who understand induced demand knows we need to reduce car capacity, not expand it. But add transit and multi model opportunities.

  22. Yes, and for highway to city driving, certify, build, share, lease, sell, buy, and drive single-width, highway-capable electric vehicles. They will eventually induce traffic, however, like pedaled and motored cycles, best width for cars is also thin.

  23. For starters… Denver RTD is in the process of completing $6 billion and six (for a total of eight) mostly light and commuter rail lines.With respect to job centers downtown has cranes everywhere and completed a $450 billion transit center a couple of years ago.

  24. The key difference between roads and housing is that housing is generally priced by the market while road use is either underpriced or (usually) completely free. Blocking new housing construction in an in-demand area merely shoots housing prices up and up.

  25. Agreed. In that vein, blocking road improvements shoots the cost/price of driving up. I know that for some, that’s the goal, but honestly, the majority of citizens need & use automobiles.

  26. This is why we need some congestion pricing. Free travel induces the demand. When properly priced, a transportation system will lead to better land-use patterns. An new office project that adds 500 additional cars will not be approved because it will directly increase the price to use the roads surrounding the project. Instead, new projects will only be approved when they do not add additional parking spots to the neighborhood. (Either by building no parking, or moving parking from another property.)

  27. Building anything ‘induces’ demand for that thing to be used, assuming it’s something people want to use.
    By this theory we shouldn’t build anything ever, because people will USE it.

  28. I think the point is that freeways are a poor investment, when PT can do a much better job for lower overall cost.

  29. This is a special case, because the use of highways is related with congestion and environmental externalities, generating market failures which are not commonly observed in other issues but urban transport.

  30. Yo should be careful. Congestion price theory establishes that there is a social benefit, but the only agent that obtains benefits is the government or the tax reciever. Every other agent or driver experiment losses. Said that, the only way to apply congestion pricing succesfully is with an effective revenue recycling.

  31. Not really Jorge. Studies of the use of Congestion Pricing applies to fixed systems confirm commuters on general purpose lanes experience faster drive times from an overall drop in traffic on the GP lanes and in some cases (mostly European) higher use of public transportation. The compelling nature of express lanes using congestion pricing is their ability to “self correct” for demand in almost real-time.

    In one article I read at least a year ago or more, the author questioned what other shopping experiences would be like if congestion pricing was used. Items having price inelastic demand are strong candidates. Gas prices drop and rise quickly, even among competitors, based upon demand as much as supply. We often hear about supply issues effecting gas prices, however, demand side pricing is more active than many people realize.

  32. Why do people ignore the fact that more traffic is a good thing and induced demand is a good thing? It means people who couldn’t go somewhere they wanted before now can.

  33. This social engineering stuff is guaranteed to blow up in your face if you push it that hard. People will start running for local government positions on pro-car platforms, and they will get elected because excess always generates pushback.
    You won’t like what happens to transportation policies when that happens. Better to take more middle-ground positions and accept not getting everything you want.

  34. The inefficiency is that the highway is free at the point of use. The vehicle is a sunk cost and the fuel expense not material.

  35. I appreciate the informed & civil discussion here. Wish that were more common.

    Regarding “…for every 1 percent increase in highway capacity, traffic increases 0.29 to 1.1 percent in the long term” – is the increased traffic limited by continued congestion? Or demand? Or what? Increasing capacity with a smaller rise in traffic could be interpreted as a good thing. That is adding more capacity (say 1%) and only somewhat more traffic (say .5%) seems like flow would indeed be better. What have I got wrong here?

    Also, couldn’t it be argued that getting that .29% increase in traffic means that more people are getting where they are going? & that perhaps the cost of increasing capacity 1% is well worth that. And that if demand requires, say, a 5% increase in traffic, that building, say 15% more capacity is money well spent?

  36. Meeting latent demand can be an economic boon but
    1. It’s seldom the official justification for such projects (the official justification is almost invariably congestion relief, which new roads don’t actually provide)
    2.The economic boon provided seldom matches the loss of resources which expanded sprawl and greater vehicle use engender.

  37. Eventually it hits a point where there is no more demand or population. This is some real liberal thinking right here.

  38. Induced Demand is not a theory anymore than any other component of Supply and Demand.

  39. This article is idiotic. It would be like saying, “The science is clear, more food equals more people, but many agricultural planners are failing to take into consideration the futility of growing more food and the obvious evil of there being more human beings.”

  40. This is the dumbest post I have seen in a while. If capacity increases by 1% and use increases by .68% short term then I would say the statistics are on the side of the DOT in this case. The long term statistic is also pretty dumb as we can spend 2 seconds on google to see that the United States has a annual population growth rate of roughly .7%. More people means more traffic. Not only are you cherry picking statistics, you chose statistics that don’t even support your point. I really wish they taught logic in schools.

Leave a Reply

Your email address will not be published. Required fields are marked *


Sober Non-Partisan Analysis: America Wastes a Ton of Money on Highways

A good deal of the $46 billion the federal government pours into highway spending each year is going to waste, according to a new Congressional Budget Office report [PDF]. The conclusion won’t surprise regular Streetsblog readers, but it’s the source that’s interesting. The CBO is not an advocacy group or an ideologically-minded think tank. It’s a non-partisan budget watchdog charged […]