There’s plenty of evidence that building roads doesn’t do much to relieve congestion. This fairly exhaustive literature review from the Victoria Transport Policy Institute shows that building new road space in an urban area tends to encourage drivers to take advantage of faster-moving traffic by making extra trips.
Estimates vary, but it seems that somewhere between most and all of any new road capacity is quickly occupied by new “induced” traffic. (See, for example, the chart to the right from the VTPI lit review.)
Reading this body of literature, Brookings Institution researcher Anthony Downs argues that traffic congestion has become an inescapable fact of urban life. In fact, he argues, the steps we take to fight congestion—such as building new roads—often carry the seeds of their own destruction:
Visualize a major commuting freeway so heavily congested each morning that traffic crawls for at least thirty minutes. If that freeway were magically doubled in capacity overnight, the next day traffic would flow rapidly because the same number of drivers would have twice as much road space.
But very soon word would get around that this road was uncongested. Drivers who had formerly traveled before or after the peak hour to avoid congestion would shift back into that peak period. Drivers who had been using alternative routes would shift onto this now convenient freeway. Some commuters who had been using transit would start driving on this road during peak periods.
Downs calls the shift of drivers from other times, routes, or modes the principle of “Triple Convergence”—a force that tends to keep traffic congestion at a rough equilibrium regardless of how much money a metro area throws at road construction.
But to me, that raises an interesting question: if a city can’t build it’s way out of traffic congestion by adding new roads, what about investments in transit? Transit advocates sometimes argue that bus or rail investments can help ease traffic, by getting people out of their cars.
Yet as far as I can tell, the evidence for this isn’t so good.
- This paper by researcher Antonio Bento and colleagues suggests that significant increases in bus service have essentially no effect on vehicle travel. Rail service increases do decrease vehicle travel, but by a surprisingly modest amount.
- This paper by researchers from the University of Toronto found—unsurprisingly—that increases in road capacity were quickly matched by increases in traffic volumes. But it also found that increases in transit service had no effect on traffic volumes. In the authors’ words: “these results fail to support the hypothesis that increase provision of public transit affects [vehicle miles traveled].”
- And this study from a University of California-Davis found that higher residential densities and greater land use mix did decrease vehicle travel—but found no statistically significant link between better transit service and less driving.
I’m sure that there’s more literature on this issue, some of which finds stronger connections between transit and vehicle travel. But in general, based on what I’ve found I have to align myself with Anthony Brooks and transit planner Jarrett Walker, who both argue that transit investments have little impact on how much driving goes on in a crowded urban area. To quote Walker:
To my knowledge…no transit project or service has ever been the clear direct cause of a substantial drop in traffic congestion. So claiming that a project you favor will reduce congestion is unwise; the data just don’t support that claim.
Transit is good for an awful lot of things. It helps move people to where they want to go; it gives people who prefer not to drive, or who can’t drive, a decent transportation option for many trips. It can reduce a region’s reliance on risky fossil fuels; and on and on. But for folks who hope that transit investments will offset the impacts of road expansions—well, sadly, I don’t think the evidence lines up that way.
John Newcomb
Take-away message from your article is that its better to do nothing in the face of traffic congestion – and I agree. There will be lots of political pressure to provide more roads, buses, and trains, but at the end of the day, those who commute should be prepared for congestion (drive a hybrid, put nice music on the car radio) and look at it as the price they pay for choosing to live in cheap suburbs or exburbs.
Eric Doherty
What this shows is that if you follow the logic and language of the road builders, and look at only the situations likely to produce marginal changes, you just get stuck in logical traffic jams.In fact, there is very good evidence that people on rapid transit (with dedicated transit lanes) avoid experiencing traffic congestion. Good quality transit gives people what they want – a way not to waste their time in congested traffic. It is people’s experiences we need to think, talk about, and measure – not cars’ experiences.If you look at the few ‘test tube’ studies available where big coordinated measures were taken to make transit more attractive you do see traffic volumes and congestion reduced radically. For example, when the U-pass program was put into place at the University of British Columbia, combined with parking price increases and staggered class times, traffic congestion disappeared from the access roads as traffic volumes dropped. But in any case, I have no ambition to eliminate traffic congestion. I enjoy watching cars stuck in traffic as I whiz by in a transit or bike lane.
Uncle Vinny
The Economist had a terrific article years ago about congestion that really stuck with me. They argued that you want major roads to be heavily congested during at least some periods of the day, or else you’ve over-built. The goal is to price major roads according to the demand so that you’re getting maximum utility out of the public investment in that road. If we encourage people to think of a certain level of congestion as a sign of efficiency (odd, isn’t it!?), and give them alternatives (transit in faster lanes, time-shifting commutes, safe walking/biking routes), maybe people will change how they feel about it.
Matt the Engineer
Anything we do to make it easier to commute long distances will induce sprawl. In the case of rail we can at least create dense nodes at stations, which is better than the pure sprawl created by cars. This in my mind is the best reason to support rail – if your choice is between building more roads or building rail, neither will reduce congestion but at least rail can create dense nodes and provide reliable and fast transportation.Now let’s go a step further. If you replace a traffic lane with a transit lane (rail or bus), would this increase congestion? Logic and the research listed above would suggest the answer is no.
Andrew
It seems to me that building a good transit service is no difference than building more lanes of traffic from the perspective of congestion relief. You will move some people out of their cars and into transit, but their spaces will simply be replaced by other drivers making more or longer trips.Investing in transit does however give you a more robust transportation system than simply investing only in roads. People do then have choices if fuel or parking prices get to high. But transit is not for everyone, just as well as driving is not for everyone.As has been said, and has been illustrated amply around the world, the only way to deal with congestion is through road pricing, or parking pricing or through fuel costs. (or more recently illustrated by eliminating the need to travel through job losses). Having a good transit system in place (and/or having well designed communities which reduce the need to travel) will make it practical and fair to introduce pricing schemes if we ever decide to get serious about dealing with the congestion issue. In the meantime were just shifting money from taxpayers pockets to road builders pockets with very little to show for it.
Chris Bradshaw
This is an important finding. Transit-systems subsidies have been justified for over 50 years on the assumption that expanding roads to the centres of cities is prohibitively expensive and that expanding/improving transit, while costing public money, is cheaper. This ‘truth’ has been promoted by the downtown land interests and the fringe land interests who know that lots of movement between their two realms is necessary their their pocketbooks. But the major factor that predicts transit use to workplaces is the cost of parking (which occupies more land than the worker). And it is why so little transit service exists to non-central worksites: parking is free. But this is not a problem to the local road engineers, who point out that there is much more ‘expandability’ to the roads that do NOT feed the city centre, since they are bordered by low-density uses and setbacks that have been mandated as part or more recent roadbuilding practices. The congestion-reduction rationale also justifies the transit ‘biz’ dichotomy of “choice” and “captive” riders: The former are peak-hour-only users who are assumed to own cars that they use (have to use) for non-peak, non-commute trips; while the latter are those who don’t have cars (at one for their exclusive personal use) and whose transit patronage can be taken for granted. I just added to my blog, http://www.hearthhealth.wordpress.com yesterday about the second difference between the two groups: ‘Choice’ patrons use the service almost exclusively during peak periods, while the latter, many of them seniors, caregivers with kids, and workers with part-time or odd-hours jobs, use a system with much different characteristics, the difference being justified by need to provide better service to ‘compete’ with personal car use. As a result, the off-peak user has to walk further to reach a service that is less frequent and slower, to a closer destination; all for the same fare. No wonder off-peak ridership is so low and why so many non-commuters are looking for alternatives to it.
Scott
I assume by “congestion” you’re referring to the phenomenon where the road becomes so full of cars that their average speed is reduced. But I’m not sure that’s the best way to look at it. What people care about is how fast they INDIVIDUALLY can get from where they are to where they want to go, when they want to go there. When you increase transportation capacity—whether by adding road space or increasing frequency of buses—then you enable more people to get where they want, when they want, faster than if the same number of people used the existing infrastructure. There may be the same level of “congestion” in the sense that the road is still full of cars and the average speed is not increased, but there are more people going places, so it’s not like you haven’t gained something. The problem I have with the induced demand argument is that it assumes the additional trips people are making aren’t valuable.
Kuba
I find Andrew’s response quite telling, price (inducements) and fairness in accessibility (social equity) are elements of a shift away from excessive VMT. Building transit/alt. networks, in light of stronger accountability and controlling investments (where do we spend transpo $’s), and factoring in external factors like health and climate/carbon impacts brings us around to a better long term approach. And, it’s not just rigorous studies that tell the story (such as the UC Davis one), there are others that could be brought to bear – well being, social values, public space, quality of life, etc.
Quimby
Does that University of Toronto study really indicate that expanding transit capacity does not affect overall VMT? I can see how more people using transit may not necessarily improve congestion, but if it doesn’t result in any decrease in VMT that seems to be a wholly different problem of much greater concern. In that case, transit isn’t reducing our dependence on oil. And if it doesn’t improve congestion (avoiding the need for highway expansions) or reduce oil use, it’s losing its primary environmental justifications.
Bernadette Keenan
I am curious about how they calculate the change in congestion? A change in the number of vehicles? And are the buses and trains empty? Because if they are full and the number of cars stays the same with little or no net increase then the congestion problem if not relieved, has at least been contained. There may not be a reduction in cars, but if there is no increase over time, the transit is a success in aleviating the problem.
Todd Litmanq
Clark, I am happy that you used information from my report on generated traffic, but for this column you might also want to review another report, “Smart Congestion Reductions II: Reevaluating The Role Of Public Transit For Improving Urban Transportation,” (http://www.vtpi.org/cong_reliefII.pdf ). A summary version of this paper was published as, “Evaluating Rail Transit Benefits: A Comment,” in Transport Policy, Vol. 14, No. 1, January 2007, pp. 94-97.This research indicates that high quality (comfortable, convenient, affordable, and integrated into a community), with grade separated routes (bus lanes or rail lines) can significantly reduce urban traffic congestion on parallel roadways. Several studies described in the study show that cities with such transit systems do have substantially lower per capita congestion delay. Congestion does not disappear, but it is never as bad as it would be if the transit system did not exist. High quality transit complements other congestion reduction strategies such as road pricing. A Puget Sound region study found that motorists were four times more responsive to congestion fees if they live in an area with high quality transit service than in other areas.The take-away message is that transit can make significant contributions to congestion reductions, but it needs to be high quality so it actually attracts people out of cars. This is only one of many reasons to implement high quality transit.
Clark Williams-Derry
Todd -As always, you make important points. That said, I wonder if simply building high-quality transit, all by itself, would really reduce congestion. It seems to me like a great complement for road pricing, to increase car travel elasticities. Quimby – Even if transit didn’t reduce congestion per se, it might still have excellent environmental justifications. Cities with good transit can often support higher population densities—so transit might let you fit more people into the same space. And that, in turn, can affect per capita energy consumption, both for residences and for transportation. In short, you can have a heavily congested city—think New York City—that also has dramatically lower energy consumption per capita than a sprawled out city. And good transit is one of the things that allows a place like New York to work—without buses & trains, NYC, or the central areas of Seattle, Portland, Vancouver, DC, Boston, etc., are essentially impossible.
Eric Doherty
“I wonder if simply building high-quality transit, all by itself, would really reduce congestion”
This is a ‘how many angels can dance on the head of a pin’ kinda question. Who cares when transit improvement are normally done along with complementary measures such as reducing parking supply? Lets look at success in the real world with the full set of complementary measures working with improved transit.
PS – The Koch Bro’s agents up here in Vancouver BC love this post.
Clark Williams-Derry
For the record, were I a Vancouver resident I’d DEFINITELY vote in favor of more Translink funding. Not necessarily because it cuts congestion, but for reasons of economic equity, energy efficiency, long-run land-use impacts, and the many other benefits of transit.
I think your mention of the U-Pass/B-line experience is a useful counterpoint to my argument. And there’s also this interesting take on a Los Angeles transit strike:
http://are.berkeley.edu/~mlanderson/pdf/Anderson_transit.pdf
“Using a simple choice model, we predict that transit riders are likely to be individuals who commute along routes with the most severe roadway delays. These individuals’ choices thus have very high marginal impacts on congestion. We test this prediction with data from a sudden strike in 2003 by Los Angeles transit workers. Estimating a regression discontinuity design, we find that average highway delay increases 47% when transit service ceases.”
Todd Litman also has a useful description of one way that grade-separated transit can affect congestion, here:
http://www.vtpi.org/tranben.pdf
“Traffic congestion tends to maintain equilibrium: it increases until delays discourage additional peak-period vehicle trips. High quality, grade-separated transit can reduce traffic congestion costs by reducing the point of equilibrium, offering travelers an alternative to driving, and by supporting compact development which reduces travel distances.”
This makes good sense to me, though I’m not sure how it plays out in conditions of substantial latent demand.
A final note — In your comments from 2011 (above) you say:
“I have no ambition to eliminate traffic congestion. I enjoy watching cars stuck in traffic as I whiz by in a transit or bike lane.”
I agree with you on the first sentence…not because I enjoy congestion or watching people in congestion, but because I think that congestion is a misguided metric of transportation success. Reducing travel times & transportation seems like a reasonable ambition, but congestion is a poor proxy for those goals.
But your second sentence might give some readers pause…if you’re going to laud a plan for reducing congestion, it’s probably not a good idea to say that you like watching cars stuck in traffic.
Weezy
“For the record, were I a Vancouver resident I’d DEFINITELY vote in favor of more Translink funding. Not necessarily because it cuts congestion, but for reasons of economic equity, energy efficiency, long-run land-use impacts, and the many other benefits of transit.”
Clark: So try justifying Sound Transit’s rail costs. Unlike any peer, it confiscates decades of heavy regressive taxes as security for a mountain of long-term debt. Seattle has become MORE inequitable since rail operations began. It has not decreased “long-run land use impacts” and it can not do so. It primarily serves people using SeaTac Airport (many of whom are not locals). Despite its grandiose taxing and spending it has not increased the percentage of Settle residents who use public transit to commute (that figure remains at 18%) — that figure comes from the US Census ACS survey:
http://en.wikipedia.org/wiki/List_of_U.S._cities_with_high_transit_ridership
Do you have any idea of the regressive tax cost to the public of the aberrant financing plan that unaccountable board is setting in to place? Serious question . . . “greenies” seem to take pride in acting oblivious about the negative impacts on households and communities regressive taxes cause when the subject of commuter rail comes up.
Todd Litman
Clark,Yes, transit oriented cities often have more intense congestion (automobile traffic speeds are significantly reduced), but residents of those cities drive less and the distances they must travel are shorter, so they experience less total per capita congestion delay. I suggest that you read the recommended paper (http://www.vtpi.org/cong_reliefII.pdf ), and the ‘Congestion Costs’ chapter of my report “Transportation Cost and Benefit Analysis” (www.vtpi.org/tca ), which discuss the different ways that congestion can be measured.Yes, high quality transit that attracts travelers who would otherwise drive and stimulates more compact, mixed land use development can significantly reduce per capita energy consumption and pollution emissions, but that is just one of the total benefits. It also provides substantial consumer financial savings (less money spent on vehicles and fuel), more dollars circulating in the local economy, accident reductions, improved public fitness and health, road and parking facility cost savings, and improved mobility for non-drivers, to name a few. For more information see “Rail Transit in America: Comprehensive Evaluation of Benefits”(www.vtpi.org/railben.pdf ).
Quimby
Those are good points, Clark, but don’t they go more to land use decisions? The UCD study you cite, for one, seems to say transit alone may not do it—it has to be combined with changes in development patterns. I can see how congestion is a separate issue, but I would have thought that transit even in isolation would bring down VMT. For congestion, people may change their routes or travel times to fill the highway back up after transit is brought in. But I assumed that those people riding the bus/train would represent reduced VMT. As the comments indicate, maybe it’s the case that the transit is helping to offset population growth and an otherwise increasing VMT, although I would have thought studies would factor that in before stating that VMT remains flat. I guess my take away is that it seems like to reduce oil use transit may need to be combined with other incentives (higher gas prices; appropriate housing options).
Daniel Henderson
Without having time to read the research, I look around on the light rail in East Portland each day and see hundreds of people right there with me who aren’t driving. The population keeps growing, and we knew that was going to happen. The total of 9 interstate lanes going into and through the Portland area are still adequate, if not ideal, because of all of the transit options. And then there are the folks who don’t own a car at all because of other options. Congestion, however it’s measured, is just one way to look at it. Pride is another. I’m reminded of the strides Portland has made to be livable every time I step on the light rail.
John Karras
I like this article and think it’s actually very important. So many cities (and their public and private sector leaders) talk about the need to reduce traffic congestion. And so much money and time is wasted in the name of clearing away bad traffic. The reality is that expanding roads creates more traffic and while building new transit lines won’t get rid of traffic jams, it does provide options…something that roadway expansion does not. And I like the mention of the many other benefits of transit at the end of the article.
Eric Doherty
Oh come off it Clark. When transit is greatly improved, people experience less congestion delays – just not the creme de la creme who refuse to ride transit. But people who ride transit are generally not white and wealthy, so they don’t count in most measures of ‘congestion’.
And there are clear examples of transit improvement reducing congestion, even for the white creme de la creme in their imported SUVs. (Yes major transit improvements almost always come bundled with other measures such as increased parking costs. This is the real world.)
Lets start with U-pass & B-line bus improvements at the University of BC in Vancouver Canada. Traffic congestion almost disappeared within weeks over a significant segment of the city. Google it.