Congestion and Size
The Texas Transportation Institute has just released the latest version of its much-criticized Urban Mobility Report. Although the conclusions and recommendations made by the TTI tend to reflect its funding sources (APTA, American Road and Transportation Builders Association), the underlying data seems sound, and suggests conclusions orthogonal to those made by the report. In addition, looking at the correlations more closely suggests obvious hazards coming from any simplistic analysis of linear regression. It even showcases how we could use data dishonestly and lie with statistics. So let’s take the data that’s relevant right now and see what we can conclude ourselves.
First, the size of an urban area is a very strong correlate of its level of congestion. The linear correlation between size and per capita congestion cost is 0.71. The correlation increases to 0.8 if we take the log of population and the log of congestion, or if we consider congestion in the absence of public transportation; in both cases, it comes from the fact that New York is far below the population-congestion regression line.
Now, more freeways do not really lead to congestion reduction. There’s some correlation between freeway miles per capita and congestion per capita, going in the expected direction, but it’s weak, -0.2, and while it’s statistically significant, the p-value is an uninspiring one-tailed 0.025. Looking at a scattergram doesn’t make any nonlinear relationship obvious.
Moreover, size is a correlate of both congestion (0.71 as above) and freeways (-0.23). This is fully expected: literature on cities’ economies of scale (here is a story of one controversial example) suggests that congestion and the economic activity causing it grow faster than linearly in city size while the amount of required energy and infrastructure grows slower than linearly. I open the floor to anyone with more powerful tools than OpenOffice Calc to do multiple regression; again, the sanitized data is here.
Even without controlling for population, freeways are not a very strong correlate. The regression coefficient is -233: increasing the number of freeway miles per thousand people by 1 (the range is 0.13-1.4, with few large metros above 1 or below 0.35) reduces the congestion cost per capita by $233 per year, also uninspiring.
The regression number alone can be used as a dishonest trick when arguing on the Internet. If we overinterpret weak correlations, we can declare that the only way to decrease congestion is to build an unrealistic number of freeways, and thus declare the problem unsolvable. Of course, for most cities we can find other cities of comparable size with much less congestion and without enormous amounts of asphalt – this is why the correlation is so weak. But a good hack should not bother himself with such caveats to talking points.
So if making an urban area larger makes it more congested, independently of and much more strongly than all else, should we give up on cities? Well, no. Assuming no change in traffic policy, congestion results from more economic activity. It then becomes straightforward to institute congestion pricing. It’s no different from how big cities can use their resources to hire more cops to deal with the crime that could result from extra interactions between people. On top of this, in very large cities, mass transit becomes a serious option: this not only reduces the amount of congestion per capita, but also removes many people from the highways to the point that congestion becomes irrelevant to their daily lives, except perhaps through higher transportation prices, which they can fully afford given the extra wealth.
Another thing to consider is that most American cities have added more freeways than people since 1982, the first year for which TTI data is available, while also becoming much more congested. If a simple relationship between freeway miles per capita and congestion held, it would be robust to these changes over time. Of course, traffic has grown even faster, leading the main report to showcase on PDF-page 21 how congestion increased the fastest in regions where road demand outgrew supply the most. But this raises the question of whether the main issue is one of demand, rather than one of supply. This is not just an issue of size: the log-log regression coefficients with cost and time is 0.42, i.e. doubling an urban area’s population will raise its per-driver congestion cost and travel delay by a factor of 2^0.42; since 1982, the average urban area on the list has seen its population grow by a factor of 1.46 and its travel delay per driver grow by a factor of 2.85 = 1.46^2.77. Cost has grown even faster, because of higher value of time.
That said, quantity of freeways does not equal quality (from the drivers’ perspective, of course, rather than the city’s). On paper, Greater New York has added freeway lanes about 9% faster than people over the last thirty years. In practice, none has addressed the major chokepoints within and into the city itself, where traffic is worst. Of course, commutes involving Manhattan are overwhelmingly likely to be done on public transportation, but diagonal commutes within the city are more likely to be done by car than on transit.
On a parenthetical note, the units of comparison here are TTI-defined urban areas. TTI’s belief about urban area population growth trends is sometimes at odds with that of the Census Bureau, but the raw population numbers are close enough. More important is the question of what to do about urban areas that are really exurbs of larger areas, such as Poughkeepsie-Newburgh and the Inland Empire. My first instinct was to lump them in with their core metro areas, but their congestion level per capita is not high. Their commutes are long, but not very congested for their size. Finally, although most correlations here are with congestion cost, the correlation numbers with travel delay and excess fuel consumptions are very similar; the one exception I’ve checked, for which I have no explanation, is log-log congestion-fuel correlation (0.84, with regression coefficient 0.73).
Pedestrian Observations 
All the shenannigans in Texas lead back to Rick Perry. Transportation, education, environment, it’s all demonstrably corrupt.
An recent article in the Texas Observer opined, “Texas Commission on Environmental Quality deliberately low-balled radiation levels in drinking water…Instead of reporting to EPA the radiation measured in community drinking water samples, plus the margin of error, TCEQ simply subtracted the margin of error.”
Texas agencies that are headed by Rick Perry appointees cannot be trusted, that is the issue. Data from the State of Texas cannot be trusted.
Chris!
Is the TTI really funded by the state of Texas? It’s an independent institute at Texas A&M, and the report only claims sponsorship by industry groups, not by the government. Is there a Perry appointee brooding there?
I don’t really know the answer to that, but if you dig around the State of Texas anywhere Perry;s influence is found you will find science deniers, hoky reports, and other unreliable information. it will take years and years to undue the corruption that Rick Perry has injected in our state. We are truly a pay to play state and many of Perry’s appointees simply reject any facts that they find inconvenient, or they hide those facts like they did with the EPA.
Chris
Well… does it mean I should reject out of hand anything that comes out of Rice or UT? I’m pretty sure they have respectable scientists doing good work even on issues that the deniers rage about.
Here is my rule of thumb, if data benefits Perry then it should be scrutinized thoroughly. If it comes from one of his appointees then double scrutinize it. And universisites tend to be better insulated from riff raff but certainly not immune.
Speaking of education, the last three heads of the Texas Board of Education were people who honestly believe the world is less than 10,000 years old, that creationism should be taught in public science class, that Thomas Jefferson’s influence should be marginalized, just to name a few. And then the science deniers appointed by Perry who occupy the office of Texas Commission on Environmental Quality laugh at global warming.
Mr. Christopher, this king of political bricking is just as low-level as people accusing any scientist studying climate and publishing something on the direction of “the Earth is warming and humans maybe responsible for that” of being part of a conspiracy to impose New World Order government etc.
It’s also a disrespect for 2 of the worldwide prestigious universities with dozens of awarded scholars in many different areas.
Should we say anything coming from UCLA or UC-Berkley that benefited Schwazzeneger policies should have been scrutinized deeply, but after elections now it’s the Democrat governor who should have underground connections checked out?
The most worrisome is what your reasoning implies: “if I’m a scholar and I have some research that supports a political group whose broader platform I oppose, I’d rather withheld than give them more ‘munition'”.
Rather sad.
Mr. Christopher needs to make a distinction. Information coming out of a report from the Texas Commission on Environmental Quality needs to be scrutinized of course: Perry probably has his hands all over that. Saying UT and Rice are in on it is a giant leap. These are prestigious institutions, and as Andre Lot points out, the logical extension of that line of reasoning leads to an unraveling of any credibility for any state university, which is unbelievable to say the least.
And Alon, I am confused. Was not your original point that the TTI tends to be skewed towards whichever organization is funding it (i.e. APTA funded study=pro transit and opposite for the ARTBA)?
The APTA and ARTBA funding sources don’t create much of a mode war skew, but they do skew the study toward proposing that cities build their way out of congestion. The literature on transit’s long-term ability to reduce congestion is more mixed than the literature on the short-term effect, and as for freeways, they barely make any dent in congestion independently of any control.
Does removing choke points even help? It seems to me there’s a negative feedback loop between driving and congestions. Reduce congestion at the choke point, people drive more, then that simply results in a compensating degree of congestion everywhere else. It seems the #1 knob to turn on congestion is cost of driving, as you suggest. That encourages use of public transit which drives improvement in that transit which further attracts others to use it: positive feedback. So giving the choice between addressing transportation needs with a negative feedback loop (investing in auto infrastructure) or a positive feedback loop (increasing the cost of vehicle use) the choice is obvious.
I don’t think removing choke point really helps – it would just move the source of congestion elsewhere. But I mention this because it’s a potential counterpoint to the argument that American freeway supply has increased: even under the traditional build-your-way-out-of-congestion theory, adding freeway lane-miles at random locations isn’t effective.
A bit tangential to the main topic of this post, but I wonder if there is any connection between the increase in “congestion” since the early 1980s and the decline in traffic fatalities since that time (or really, in VMT terms, since the rise during the 1960s). There’s seems to be some supporting literature out there about this:
http://usj.sagepub.com/content/34/4/679.abstract
Two things:
1. People who make journals closed-access are the scum of humanity.
2. Not being able to read the study I don’t know whether it looked at congestion and fatalities in different regions and countries. But it would have to do that to convince me, because the trend of reduction in traffic fatalities per VMT has been constant in percentage terms going back to when data was first kept, in the Model T days.
I guess there is reasonable evidence vehicles got safer by any measure considered.
Highway fatality rate is rather low compared to city traffic fatalities, and have been decreasing constantly. The revocation of stupid 55mph law and subsequent maximum (and actual) speed increases didn’t bring “massive deaths” as claimed by its detractors.
The clearest conclusion I got from your analysis, Alon, is that adding more freeways does not reduce congestion in a substantive manner. A coefficient of 0.23 is less than a standard deviation from base congestion, I think?
A secondary problem is that addressing chokepoints only shifts them. For example, if you had a tree diagram from Point X where Road A split into Roads B and C, when then split into Roads D, E, F, and G, and so on, the natural chokepoint would be where roads B and C merge into Road A (since Roads B and C are handling all the traffic coming off of roads D et al.) Fixing this chokepoint–adding a Road H between Point X and the junction of Roads A, B, and C, would actually serve to shove the chokepoints out to where Roads B and C meet Roads D et al. The only natural way of solving this would seem to be to have Roads D et al. all independently converge on Point X–but then Point X would become so overwhelmed with traffic it would become a chokepoint. And so on. So congestion is in part a supply/demand function, where demand outstrips supply; supply/demand functions are themselves corollaries of economic activity, and thus congestion is significant of a great deal of economic activity being generated in one place, enough to influence the supplies and demands of a whole lebensraum around it. Ergo economic solutions are the best solutions to deal with congestion (and ideas like congestion pricing and the privatization of parking are first and foremost economic solutions).
But wait, there’s more! One of the major functions of a freeway is bypass–that is, allowing traffic wishing to pass the city on their journey from elsewhere to elsewhere to efficiently bypass it. To that end, urban bypasses are still needed, and need to be built in such a way as to bypass the regions of maximal congestion. As congestion is a corollary of economic demand (i.e. the desire to locate in the city), such bypasses need to be built to an area of minimal economic demand (some distance away from the city). A good example is the NJ Transit’s bypass of Philadelphia. But the kicker is that certain functions will eventually locate alongside the bypass, due to the opportunities it presents (warehousing, for example); on occasion, these activities will feed on themselves until there is a major economic center right alongside the bypass originally built to, er, bypass the major economic centers. And voilà! An edge city.
Really, any attempt to analyze transportation and mobility on any sort of linear scale, and without recourse to economics, is doomed to fail. The problem, in many ways, is intractable.
Wait, what do you mean “less than a standard deviation”? Correlation coefficients are not like that; they measure how much the variance can be explained by the regression line (namely, the square of the correlation). The -0.2 correlation between freeway lane-miles per capita and congestion per capita is statistically significant, with a p-value equal to that of being about 2 standard deviations below the norm, but neither the statistical significance nor the regression coefficient is very high.
The situation you describe for freeways creates more chokepoints, yes. So maybe I should clarify that when I talk about relieving chokepoints, what I mean is increasing the number of lanes at the chokepoint, wherever it is. So in your example, if initially roads A-G all have two lanes per direction, the chokepoint has two lanes, and if road H then has two lanes per direction as well, the chokepoint is increased to four lanes. Of course it’s just going to lead to more traffic and to jams at other junctions, but the point is that from the point of view of the standard fluid dynamics-like theory of traffic held by the road engineers, certain freeway upgrades will reduce congestion more than others.
I can’t really argue with you on bypasses, except to say that in many ways it’s best to have the freeway network only offer bypasses, and make people drive into the city on local streets. London, for all its many arterial street faults, has zero freeways in the urban core. The urban expressway, Robert Moses’ contribution to road building, creates more problems than it solves.
I’m not too strong with math…I had no idea the correlation was statistically significant.
Agreed with your broad point: a freeway network should be bypass-oriented rather than commuter-oriented. As you note, this is the case in many European cities. Combining this with a network of major avenues creates a highly livable yet highly traversible city for all modes.
Bleh. It’s 100% due to unclear writing on my part, when I said that the correlation was statistically significant but not by a large amount and with a weak regression coefficient. These are two different things: regression measures how much congestion reduction you get out of building more freeways, and correlation measures how close the distribution is to the regression line.
The dishonest argument I give in the post is dishonest only because the correlation is weak; if the correlation is perfect and the regression line is still weak, then the argument is correct, because congestion reduction comes exactly with freeways (or maybe population loss…). A shallow regression line with perfect correlation then means something. In other words: because the size-congestion correlation is so strong, I’m comfortable saying that there isn’t all that much a large city can do to reduce congestion except shrink, or maybe build a transit system that bypasses congestion for most users; however, because the freeways-congestion correlation is weak even before we control for size, it is dishonest to say that a city can only reduce congestion by tripling its freeway network size or more.
Alon, if you haven’t seen the CNU Salon with TTI’s Tim Lomax and CEOs for Cities Joe Cotright it’s really worth having a look. The arguments on both sides are as usual interesting. I argue as you do above that there are certain people (such as myself) that never see the congestion. I’m always on the train or riding my bike. So it’s really more of an urban roads congestion index rather than a general congestion index. In the Q&A session you might hear some familiar voices including mine, Jarrett from Human Transit and Yonah from TP.
http://www.youtube.com/user/CEOsforCities1
In the Q&A, Tim goes through all the funding sources, which is not the State of Texas directly.
The problem with such studies is that they ignore, not for scholarly incompetence but lack of data, other factors that come play their part on long-term traffic trends.
One of those factors that shouldn’t be ignored is the probably increase on the average number of daily trips a person takes. This is tied with the decline of share of the once super-majority of household: a family of 2 adults, one staying home, other working on a fixed scheduled with kids.
Throw in old trends like kids attending schools closer to their houses (no magnet schools, no school choices – and rolling more back not even busing -) and not taking 8 extracurricular, scattered-around-town activities; and you get more reduction on average trips per household or per person.
Of much importance is also the fact the average size of a business unit facility, be it an office or a factory, decreased a lot. Factories were difficult to relocate, offices and stores, much less. Then, as work mobility increased (a very positive trend on itself, IMO) throughout one’s career, it started to make less and less sense to “live close to your work” if you are expected not to stay 15 years working in the same place to begin with.
These kinds of factors I cited, among many others, influence one of the inputs of such regressions of congestion on a/b/c.
The problem is that those aren’t peak trips. Congestion does not come from driving to the supermarket and back or from shuttling the kids to after-school activities in the neighboring suburban subdivision. It comes from peak-hour work trips (and the raw data indicates nearly all peak-hour trips are indeed for commuting) – and ones involving intense job sites, for edgeless cities tend to have very low job density precisely because of the gridlock the edge cities have.
The tidy little tables the Census Bureau publishes, count people who are over 16. So the very long commutes some school children take aren’t in the reports.
Shuttling the precious little scholars to school happens in the peak of the morning rush hour. All traffic halts when the flashing red lights come on. On a busy street that means there’s clumps of traffic being created. Taking them to after school activities means they have to be taken back. During the shoulder or right in the middle of evening rush.