Seattle traffic congestion seen from Rizal Park

Is Traffic Congestion a Con?

Since I’m traveling this week, I’m just going to pass along this excellent section from a recent Surface Transportation Innovations newsletter from Robert Poole at Reason: (highlights mine)

Is Traffic Congestion a Con?

Last month, INRIX issued a new report on urban traffic congestion worldwide, with new rankings for U.S. metro areas. Based on a revised method of assessing commuter traffic (counting trips to more than just the central business district of each metro area—finally) it identified Boston, Chicago, and Philadelphia as having the highest amount of delay hours per driver, worse than Washington, DC, Los Angeles, and San Francisco. Due to their much greater size, the New York, Los Angeles, and Chicago metro areas are ranked as first, second, and third in the total cost of congestion, which is quite plausible.

These numbers are based on 2019 INRIX traffic data, whereas the well-known data in the most recent 2019 Urban Mobility Report (UMR) from the Texas A&M Transportation Institute are from 2017. Despite INRIX having worked with TTI on that report, its delay and cost numbers are very different, with INRIX now reporting an average delay per driver nationwide of 99 hours per year compared with 54 hours per year in the 2019 UMR. Even more startling, the UMR estimated total U.S. congestion cost as $179 billion a year, while the new INRIX report’s number is a mere $88 billion. Due to TTI’s more sophisticated methodology, I’m inclined to accept their congestion numbers.

Has Freeway Expansion Failed to Curb Traffic Congestion?

Meanwhile, about the same time as the INRIX report reached my inbox, Transportation for America (T4A) released a new report called “The Congestion Con.” Its thesis is that our main national strategy for dealing with congestion is expanding freeway capacity—and that this has utterly failed. The report uses data from the 2019 Urban Mobility Report together with Federal Highway Administration (FHWA) highway statistics and census bureau data in an attempt to demonstrate its case—and utterly fails. An excellent rebuttal, well informed with data and reasoning, is Randal O’Toole’s response, “The Induced-Demand Con.”

The T4A authors use some odd measurements to make their case. First, they compare the percentage growth in freeway lane-miles with the percentage growth in population between 1993 and 2017, finding that the lane-mile increase was generally higher than the population increase. O’Toole points out a major problem with this. FHWA highway statistics reveal that large amounts of exurban freeway mileage that already existed in 1993 were outside the 1993 definition of urbanized areas, but are included in their 2017 boundaries. He plausibly estimates that “well over a third of the 30,511 [lane-miles] that T4A implies were built in that time period already existed in 1993.”

A second problem is that the percentage increase in congestion is a poor basis for comparing metro areas. O’Toole notes that low-congestion metro areas (like Bakersfield, CA) still have low 2017 congestion but may have had a high percentage increase since 1993. By contrast, a highly congested area like Los Angeles typically has relatively small percentage increases in its already enormous traffic congestion. He suggests a better metric would be to compare freeway vehicle miles of travel (VMT) per lane-mile over time. That metric allows us to discern differences among metro areas that did and didn’t add a lot of freeway lane-miles.  For example, neither Portland nor Seattle added many lane-miles between 1993 and 2018, but their freeways got increasingly clogged: VMT/ln-mi. up 84 percent in Portland and 44 percent in San Francisco. By contrast, the increases in metro areas that added a lot of capacity are much lower: VMT/ln-mi. up only 21 percent in Houston and 14 percent in Phoenix.

T4A’s assertion that adding capacity has failed to reduce congestion is also falsified by data that used to be included in TTI’s Urban Mobility Reports but has been omitted since their 2012 report. With TTI’s permission, I reproduced a graph from that report in my book, Rethinking America’s Highways (page 258). It shows that 17 metro areas that had capacity growth within 10 percent of traffic growth actually had a declining trend in congestion increases between 1998 and 2010, compared with strong and ongoing congestion increases from 1982 through 2010 for 84 metro areas that added far less capacity.

Another major problem is that T4A continues to put forth the “induced-demand” thesis: that it is futile to add freeway capacity because new lanes will simply fill up, and congestion will soon get back to what it was before. If that were literally true, as O’Toole and others have pointed out, every freeway would have high levels of congestion, yet, for example, Los Angeles freeways are overloaded with 23,000 VMT per lane-mile per day, while Pittsburgh freeways (for example) breeze along with just 9,000. In my book, I devoted several pages to a critique of the most-cited source on induced demand, the 2011 paper in American Economic Review titled “The Fundamental Law of Road Congestion: Evidence from U.S. Cities.” If you don’t have my book, you can read a condensed version of this sidebar here.

The bottom line of T4A’s report is that America should stop expanding freeways and radically reform land-use policy to discourage or prohibit outward expansion of metro areas (which they deride as “sprawl”). National policy, they urge, should reorient the federal surface transportation program away from reducing delay to encouraging better “access.” By that, they mean the densification of urban areas, on the premise that people could, therefore, walk, bike, or use transit to get to “jobs” and other destinations. The fallacy in all of that “access” research is that it treats jobs as generic—as long as you can get to some kind of a job, problem solved. But as researchers like Alain Bertaud of NYU (Order Without Design, MIT Press) and others have shown, this is a recipe for reducing the productivity of urban areas. Urban agglomeration effects only come about when people and companies are able to find each other and engage in positive-sum transactions—individuals finding the best jobs and companies finding the best people. This is what happens in urban areas that enable fast commutes over long distances—exactly the opposite of what “smart-growthers” like T4A recommend. A growing body of research finds that metro areas with fast, region-wide transportation infrastructure have significantly higher economic productivity.

P.S: I will give T4A credit for two good points in their 2020 policy agenda (and included in this new report). I agree that federal and state political considerations have led to a serious underfunding of highway maintenance. Also, I agree that road pricing is underused, and could make a real difference as part of reducing traffic congestion. But a lot of that requires new construction, such as adding priced managed lanes to congested freeways, which T4America does not support.

This piece first appeared on Houston Strategies.


Tory Gattis is a Founding Senior Fellow with the Center for Opportunity Urbanism and co-authored the original study with noted urbanist Joel Kotkin and others, creating a city philosophy around upward social mobility for all citizens as an alternative to the popular smart growth, new urbanism, and creative class movements. He is also an editor of the Houston Strategies blog.

Photo: Ron Clausen via Wikimedia under CC 4.0 License.