New Understanding of Traffic Congestion

Todd Litman's picture

Congratulations to this year's high school, college and university graduates! The current crop includes our son, who was recruited by a major corporation. The location of his new job will affect his travel patterns and therefore the transportation costs he bears and imposes for the next few years: until now he could get around fine by walking, cycling and public transport, but his new worksite is outside the city center, difficult to access except by automobile. As a result he will spend a significant portion of his new income to purchase and operate a car, and contribute to traffic congestion, parking costs and pollution. This is an example of how land use decisions, such as where corporations locate their offices, affects regional transport patterns and costs. It illustrates research showing that where people work and shop has as much impact on their travel habits as where they live.

Other recent research offers additional insights. A report titled Land Use and Traffic Congestion, published by the Arizona Department of Transportation, is changing the way we think about congestion and solutions. It found that residents of higher-density neighborhoods in Phoenix, Arizona drive substantially less than otherwise similar residents located in lower-density, automobile-dependent suburban neighborhoods. For example, the average work trip was a little longer than seven miles for higher-density neighborhoods compared with almost 11 miles in more suburban neighborhoods, and the average shopping trip was less than three miles compared with over four miles in suburban areas. These differences result in urban dwellers driving about a third fewer daily miles than their suburban counterparts. 

That is unsurprising. There is plenty of evidence that land use factors such as density, mix and road connectivity affect the amount people travel. However, the study made an important additional discovery. It found that roadways in more compact, mixed, multi-modal communities tend to be less congested. This results from the lower vehicle trip generation, particularly for local errands, more walking and public transit travel, and because the more connected street networks offer more route options so traffic is less concentrated on a few urban arterials. This contradicts our earlier assumptions.

In the past experts often assumed that, although more compact development increases overall accessibility (it increases the number of goods and activities within an area and therefore reduces the travel costs required to reach them), it increases local roadway congestion intensity, a point used to criticize smart growth.

It turns out that this criticism is untrue. It fails to account for real world conditions. Not only does smart growth significantly reduce automobile trips, by offering better accessibility options it allows people to respond to congestion by shifting mode and route. For example, when congestion is a problem you walk or bike to local stores rather than driving to a more distant shopping center, some commuters shift to alternative modes, and motorists can shift to less congested routes for some trips. These solutions are not possible on newer suburban communities where destinations are dispersed; walking, cycling and public transport inferior; and hierarchical road networks channel all traffic onto major arterials.

This has important implications for transport and land use planning. It indicates that smart growth development policies have smaller costs and greater benefits than usually recognized, including local and regional traffic congestion reductions, but it also indicates that these benefits are contingent; they require an integrated set of policies including increased density, mix, connectivity and transport options. As a result, the best response to smart growth criticism is more smart growth, for example, more density and mix, additional pedestrian and public transit improvements, more connected transport networks, more parking management, and additional incentives to shift travel mode.

Critics often assume that smart growth consists only of increased development density. If that were true then some of their criticisms could have merit, but it is inaccurate, as discussed in a previous column, An Inaccurate Attack On Smart Growth. Smart growth involves a combination of increased development density and mix, more connected paths and roads, and improved transport options. Together, these land use reforms can provide a host of direct and indirect benefits.


For more information

Md Aftabuzzaman, Graham Currie and Majid Sarvi (2011), "Exploring The Underlying Dimensions Of Elements Affecting Traffic Congestion Relief Impact Of Transit," Cities, Vol. 28, Is. 1 (, February, Pages 36-44.

Wendell Cox (2003), How Higher Density Makes Traffic Worse, Public Purpose (

J. Richard Kuzmyak (2012), Land Use and Traffic Congestion, Report 618, Arizona Department of Transportation (; at

J. Richard Kuzmyak, Jerry Walters, Hsi-hwa Hu, Jason Espie, and Dohyung Kim (2012), Travel Behavior and Built Environment: Exploring the Importance of Urban Design at the Non-Residential End of the Trip, Lincoln Institute of Land Policy (; at 

Todd Litman (2011), Smart Congestion Relief: Comprehensive Analysis Of Traffic Congestion Costs and Congestion Reduction Benefits, Victoria Transport Policy Institute (; at

Todd Litman is the executive director of the Victoria Transport Policy Institute.



Lower speeds = less congestion

Great post. It's not surprising that your referenced study found more compact areas to have less congestion than less compact suburban areas. Yes, one reason is fewer and shorter trips. There's another reason, and one very much underappreciated: lower speeds in more compact areas. Road lane volume capacities differ at different speeds. Maximum lane capacity occurs around 25-30 mph, and this is well over twice the capacity than at 50 mph. The operating dynamic is simple: motorists travelling at 25 mph are more "bunched" together than at 50 mph. There's more benefits to slower speeds that affect traffic congestion. Here's one example: left turning motorists on a 25 mph street have a much easier/quicker time completing that left turn than a driver facing identical traffic volume in the opposing lane but travelling at 50 mph.
Keith Tianen

Todd Litman's picture

Optimal Traffic Speeds

Thanks for your comments, Keith.

That is a good point: according to traffic engineering studies traffic volumes tend to be maximized at moderate speeds.

This is a reminder that a problem such as traffic congestion can be evaluated from several different perspectives, including traffic engineering (maximizing vehicle flow and route options), transport planning (maximizing modal options), economics (efficient pricing to ration road space and encourage use of efficient modes) and land use planning (optimize land use mix to reduce travel distances). Each of these perspectives suggests different solutions. It is important that everybody involved in the planning process understand all of these perspectives and solutions so we can identify the truly optimal set.

Todd Alexander Litman
Victoria Transport Policy Institute
"Efficiency - Equity - Clarity"

Local vs regional traffic congestion

Dear Dr. Litman,

This a very interesting post indeed. It seems to contradict the so-called “paradox of intensification” as aptly described by Melia et al. (2011). However, I am a little skeptical about the interpretation of this study results. Less traffic congestion in these areas may be the result of fewer people choosing their cars for their activities within the compact area “boundaries”. However, there may also be another large share of people living in these areas who just do not have the option to choose for the alternative means of transport to cover their mobility needs. This means that even if “congestion is a problem you -cannot- walk or bike to local stores rather than driving to a more distant shopping center (just because local stores do not offer the same consuming opportunities), or some commuters -cannot- shift to alternative modes (just because their workplaces are far way for their home located in the compact community)”. As a result a large share of the compact area residents may act as captive car users contributing to regional but not to local traffic congestion. Consequently, compact communities may have less traffic congestion within their “boundaries”, but they may also contribute to higher traffic congestion (..and energy consumption, pollution etc) out of them. The question here is which would be the optimum combination of local and regional planning policies that could offer a more balanced approach as regards the reduction of automobile dependence.

Melia, S., Parkhurst, G., & Barton, H. (2011). The paradox of intensification. Transport Policy, 18(1), 46-52. Elsevier. doi:10.1016/j.tranpol.2010.05.007

Todd Litman's picture

Local vs Regional Trip Generation and Congestion Impacts

Thank you Dimitris Milakis for your comments.

Yes, Melia, Parkhurst and Barton's study reflects the common assumption that more compact development increases traffic congestion intensity (I am pleased to see that they rely on my report, "Use Impacts on Transport, How Land Use Factors Affect Travel Behavior" [] for their analysis). However, their analysis is based on general assumptions about how land use factors affect per capita vehicle travel. In contrast, Kuzmyak's study used actual traffic density data on a specific corridor, which takes into account important additional factors related to how travelers actually respond to roadway congestion. It shows that a combination of improved travel options (better walking, cycling and public transit services), land use mix which reduces travel distances, and a more connected road system that improves route options, significantly reduces both per capita vehicle travel (which should reduce regional traffic congestion) and reduces local traffic congestion.

These findings are consistent with your conclusion; if smart growth consisted only of increased development density then it probably would cause more intense traffic congestion, but true smart growth also includes increased development mix and centricity, improved travel options (better walking, cycling, carsharing, ridesharing, public transit and delivery services), more connected path and road systems, and more efficient parking management. Kuzmyak's research shows that together these reduce local traffic congestion.

In other words, to the degree that more compact development by itself increases congestion, the solution is more complete smart growth policies that include other VMT reduction strategies.

This supports your point, that a key factor is the quality of transport options available, so travelers can and will reduce their driving on major congested corridors. This can consist of a combination of being able to walk and bike for local errands (in automobile dependent areas people often drive when traveling short distances due to very harsh walking conditions; such short trips can make a major contribution to congestion due to the friction of exiting and entering driveways), and using public transit for longer-distant trips. That is why high quality public transit (fast, frequent, integrated, attractive) can provide significant traffic congestion reductions, as discussed in my report, "Smart Congestion Relief: Comprehensive Analysis Of Traffic Congestion Costs and Congestion Reduction Benefits" ( ).

Our new understanding of traffic congestion indicates that the best solutions are integrated policies that improve travel options, provide incentives to use space-efficient modes, more connected roadway networks, plus land use policies that improve accessibility. Conversely, automobile-dependent sprawl development increases traffic congestion. I think this is an overall positive finding because it expands the range of congestion reduction strategies to include smart growth land use policies.

Mr. Todd Alexander Litman (Note, I am not a "Dr.")
Victoria Transport Policy Institute
"Efficiency - Equity - Clarity"

Compact communities and metropolitan level smart growth policies

Dear Mr. Litman,

Thank you very much for your reply. I totally agree with your arguments regarding the necessity of more complete smart growth policies that include a variety of VMT reduction strategies. Please let me share some of my thoughts on this issue, using data from the Kuzmyak's/ADOT study. According to this research results for the 4 urban corridors/zones analyzed, the internal capture rate for all trips falls between 20% (for the suburban area-SA; Bell Road) to 47.7% (for the compact area-CA; Scottsdale). This percentage falls to 13.4% (SA) - 21.2% (CA) for HBW trips. In addition, just a small percentage of internal-external trips are made by public transport (0.8% SA - 10% CA). This probably means that a large share of residents (even in CAs) is captive to car. Specifically 53.3% (CA) - 80% (SA) of all trips and 78.8% (CA) - 86.6% (SA) of HBW trips bound for regional destinations with extremely high car use. Consequently, the CAs may contribute significantly to regional traffic congestion, although less than the SAs. The question is whether a sufficient network of bus lines or even light rail lines (as in the case of this city) is adequate to absorb these regional car trips (and consequently reduce regional traffic congestion) or a metropolitan-level smart growth planning approach is also necessary for this purpose.

Dimitris Milakis

Todd Litman's picture

Evaluating Congestion Impacts

Dear Dimitris,

Thank you for your feedback.

I think we agree on the conclusion that improving travel options (particularly grade-separated public transit or HOV lanes) on major travel corridors can be an effective congestion reduction strategy. This is supported by both theoretical and empirical evidence, as discussed in my "Smart Congestion Relief" report. Improving roadway connectivity (more grid, less channeling of traffic onto major arterials and highways) can also significantly reduce congestion. Congestion pricing could achieve similar results, but it has high implementation costs and faces strong political resistance. Other strategies, such as efficient parking pricing, more land use mix, improved public transit convenience and comfort, commute trip reduction programs, and improving walking and cycling conditions can also help, but are probably only marginally effective by themselves.

I don't think that overall mode share statistics are suitable for this type of analysis since, for congestion analysis, we are only concerned with peak-period trips on major corridors, which is typically only 10-20% of total trips. Thus, a large 20% reduction in peak-period vehicle trips, which can significantly reduce traffic congestion, is only a 2 to 4 point shift in overall mode share. The key question is, how responsive are travelers to congestion. Residents might make lots of vehicle trips, but if they have good mode, route and destination options, they will make changes to avoid congestion. For example, you might say, "I normally shop at Walmart but since the road is congested I'll buy dinner at the local grocery store," and "Now that Central Avenue has a bus lane I can get to work faster by bus than driving, so I use it several days each month."

Unfortunately, few travel models effectively incorporate congestion feedback - they do not recognize all of the ways that travelers respond when roadway congestion increases and are not very sensitive to factors such as walking and cycling conditions, public transit convenience and comfort factors, land use mix, or long-term impacts on development patterns. As a result, they tend to exaggerate the severity of future congestion problems, underestimate the congestion reduction benefits that would result from qualitative improvements to alternative modes (other than increased speed or reduced fares), and tend to use low price elasticity values which understate the long-term effects of pricing on vehicle travel.

This is often an issue in debates about the value of public transit improvements. Critics make statements such as, "We are spending billions of dollars to improve public transit to increase transit mode share from 2% to 4%. That's inefficient!" But such analysis ignores the much larger mode shifts on major urban corridors, often 10-40%, which significantly reduces traffic congestion, and the leverage effects that high quality public transit has on per capita vehicle travel. High quality transit can be a catalyst for transit-oriented development, where residents tend to own fewer motor vehicles, drive less and rely more on walking, cycling and public transit. Their vehicle travel reductions are usually much greater than just the trips shifted to transit. Where this occurs, each additional transit passenger-kilometer represents 2-9 reduced vehicle-kilometers. For discussion of this see "Evaluating Public Transit Benefits and Costs" ( ) and the "Transit Land Use Multiplier" ( ).

Todd Alexander Litman
Victoria Transport Policy Institute
"Efficiency - Equity - Clarity"

New Traffic Tactic?

I am intrigued by this discussion, the topic of which is still very new to me (I am not an urban planner - there, I said it!). The urbanization of the globe is, however, of great interest to me, and I am particularly enthralled by some of the projects I see around me in Europe (I live in Berlin). We've all heard about the bikes-only neighborhood in Freiburg, but I recently read about another initiative launched in Austria. I think some of your readers might find it interesting, particularly given the tradition of high density neighborhoods in Europe's oldest cities:

Prepare for the AICP* Exam

Join the thousands of students who have utilized the Planetizen AICP* Exam Preparation Class to prepare for the American Planning Association's AICP* exam.
Starting at $245

Essential Readings in Urban Planning

Planning on taking the AICP* Exam? Register for Planetizen's AICP * Exam Preparation Course to save $25.
Melissa and Doug Town Blocks Wooden Play Set

Block play meets role play

New! Town block set from Melissa and Doug Toys

Get the "Green Bible" of planning books

Understand the complexities of planning at the local level while preparing for the AICP* exam.