The Harvard Business Review has a piece this month on research by Lawrence Frank, Bombardier Chair in Sustainable Transportation at the University of British Columbia, on the effects of cul-de-sacs in neighborhoods in King County, Washington. He found that residents in areas with the most interconnected streets travel 26% fewer miles by automobile than those in areas with many cul-de-sacs. Meanwhile, recent studies by Frank and others show that the higher a neighborhood’s overall walkability, the greater the amount of walking and biking— which means a drop in per capita air pollution, fuel use, and body mass index.

The theory behind cul-de-sacs was that they lessened traffic, since they change the primary function of local streets — rather than offering a way to get anywhere, now they simply provide access to private residences. The problem is that this design inherently encourages car use, even for the shortest trips. It also limits the growth of communities and transportation options. Consider the above maps of one-kilometer walks in two different Seattle suburbs — the first, in Woodinville, is all cul-de-sacs that result in a disconnected jumble of streets with no walking or bike paths, while the second, in Ballard, offers an interconnected network of streets that provide easy access to shopping, parks, and other destinations. The argument that cul-de-sacs increase safety because they limit traffic is also misguided — the more empty and desolate a suburban (and often affluent) street is, the more likely crime is to occur. Also, it’s much harder for emergency vehicles to reach these homes if they’re sequestered in the belly of a web of disconnected dead-ends.

As more and more direct evidence piles up that these dead-end developments are doing no one any good, the cul-de-sac tides are beginning to change: Last year, the Virginia legislature passed a law limiting cul-de-sacs in future developments. And if other states see the benefits for VA –  more efficient streets that are cheaper to maintain, as well as other savings from not having to widen arterial roads that otherwise were overburdened by cul-de-sacs — perhaps they’ll follow suit.

Image: Urban Design 4 Health

CLICK TO ENLARGE

While states with the highest populations unsurprisingly tend to use the most gas, the real fuel efficiency picture comes when you examine each state’s fuel consumption per capita. With that factor added, a very different scenario emerges: High-use states like New York actually have low per-capita usage, while states like Alabama, Iowa, Mississippi, Missouri, New Hampshire, and North Dakota — all states with smaller populations and large distances required for drivers — have higher than average consumption, with the highest per capita usage falling on Wyoming. One surprisingly low per capita consumer is Utah — perhaps because of the Mormon population’s tendency to inhabit smaller areas that don’t require long drives?

Over the past month, economist Ed Glaeser has explored the benefits of high-speed rail inan occasional series over at the New York Times website. To put it mildly, his reception in the blogosphere has been wretched. Ryan Avent at Streetsblog has been a particularly devastating critic, picking apart Glaeser’s analysis strand by and strand and characterizing the overall effort as “daft and indefensible.”

But what’s been missing thus far is a numbers-based rebuttal of Glaeser’s “back-of-envelope calculations.” He figures three categories of benefits from high speed rail: travel (for example, fewer car accidents and reduced highway congestion), environmental (lower carbon emissions than car or plane travel, etc.), and improved land use (the rail project encouraging denser, more walkable cities, etc.). Through this combination of factors, Gleaser examines a hypothetical HSR link between Dallas and Houston and calculates annual benefits of $158 million. Not bad perhaps, but they pale in comparison to annual costs of $648 million. The gap between costs and benefits–an annual loss to society of $500 million–would seem to be so huge as to kill the prospect of US high speed rail in its cradle.

That may even to have been Glaeser’s intent in writing the series. The problem is that–through a sorry mix of omission, oversimplification, distortion, and deficiency–his calculations bear no relation to the effects he is claiming to consider. So it’s important to show that “the numbers” do not at all undermine the viability of HSR in the US, even outside the northeast and California. In fact, they tend to support it.

By populating his model with a better set of assumptions, we hope to show how badly the economist missed the mark even on his handpicked example of an HSR link between Houston and Dallas. In reality, a well-designed high speed intercity rail project between the two largest cities in Lone Star State would likely produce a net economic benefit–not at all the white elephant Glaeser suggests. In this more comprehensive model that takes into account trivialities like regional population growth and a reality-based route, the annual benefits total $840 million compared with construction and maintenance costs of $810 million. Which is to say, our numbers show that HSR pays for itself rather handily.

And this would be early in the lifecycle of the system, with those benefits likely to grow in future decades.

The Basics: A Better Set of Assumptions

Rather than looking at Glaeser’s hypothetical 240-mile rail line directly and exclusively between Dallas and Houston, I’ll base my argument on a line actually under consideration called the Texas T-Bone that would run roughly 300 miles between the cities, with intermediate stops at Waco, Temple, and College Stations. For simplicity’s sake, in this piece I’ll ignore the roughly 140-mile proposed extension of the line south to Austin and San Antonio but factor in connecting slow-speed trains from those locales.

Despite the fact that an HSR system would take more than a decade to build, Glaeser calculations are all for 2008. Why? We have no idea. Unlike some other US states, Texas is projected to grow steadily in coming years. Assuming the project gets underway relatively soon, the Texas T-Bone HSR line ought to be hitting full stride around 2030. So our model focuses on that year. Texas is projected to have 33 million people (up from 24 million today) with the metropolitan areas of Houston and Dallas each growing by more than 4 million inhabitants to populations of 9.9 million and 10.6 million, respectively.

Ridership: Using Real-World Examples

Glaeser argues that a Houston-Dallas line would be roughly one-half as popular, relative to population, as the current slow Amtrak service is in the Northeastern Corridor. His reasoning is that both Dallas and Houston are less transit-friendly areas, and therefore less conducive to train travel. So, assuming a 50 percent lower per capita ridership rate, he comes up with 1.5 million annual customers for the line – this is similar to the number of people who currently fly directly between the two cities.

There are a number of major flaws with this approach though. First, while transit-friendly conditions are desirable – and it bears mention that both Dallas and Houston are expanding their transit systems significantly – there is little evidence those networks are vital in attracting customers to high-speed rail.

Second, Amtrak services between Washington and Boston have never fit the international definition of high-speed (186 mph). They reach a high of 165 mph for a short segment in Rhode Island and Massachusetts, but Acela trains average only 72 mph overall, so their ridership is hardly an apt point of comparison, as Glaeser suggests.

Third, Glaeser neglects to mention that a number of commuter rail agencies in the Northeast serve long-distance travel, such as MARC between Baltimore and Washington and MBTA between Boston and Providence. These agencies cannibalize Amtrak market share, carrying more than 60 million passengers a year on track shared by Amtrak Northeast Regional and Acela trains; Glaeser’s calculations do not account for these rail riders at all.

So instead of deeply flawed attempts to project ridership based on the Northeast, we should be focusing on high-speed rail’s noted ability to take substantial market share away from the airlines and even from automobile commuters. Evidence from overseas to this effect is plentiful, though Glaeser doesn’t even mention it. In France, for instance, the 200 mph TGV Est line between Paris (metro population 11 million) and Strasbourg (600,000) carried 11 million passengers in its first year of operation. Rail now commands 70% of total travel market share, including automobiles, versus 30% before the line opened. Today, roughly 10 million people a year travel between Dallas and Houston either by plane or by car.

Looking to the future provides even more evidence of the route’s potential ridership. We can reasonably speculate that trains would take 80% of the market from air services offered between Houston, Dallas, College Station, Waco, and Temple, each of which would be located directly along the line. By 2030, about 4.3 million people will take flights between those cities, based on current growth rates. In addition, people connecting to the line on routes between Austin and Houston (900,000); Austin and Dallas (1.2 million); and Dallas and San Antonio (800,000) would have an improved time incentive to take the train (up to 30% of market share for the first two and 10% for the latter). In total, around 4.1 million people every year would choose the high-speed route instead of air travel.

Of the 12 million people who will likely drive between Dallas and Houston by 2030, looking to international example leads us to conclude that 25% could be expected to abandon their cars for the train ride as long as services are provided at a competitive rate. Fewer people would likely choose the train on trips between Austin and Dallas (15%); Austin and Houston (10%); and Dallas and San Antonio (5%), simply because rail would be less convenient than on direct Dallas-Houston journeys. Based on reasonable assumptions, 1.3 million car drivers could be expected to switch to the train on trips to and from College Station, Waco, and Temple. In total, a full 7.3 million trips might be removed from the highways.

This adds up to a projected more than 11.4 million annual riders riding the train (27,000 a day), far higher than Glaeser’s oversimplified assumptions. This figure is well supported by a comparison with Spain, whose AVE line between Madrid (7.1 million) and Barcelona (3.2 million) will serve 8 million passengers a year by 2011, taking 50% of total market share.

With this level of ridership, we calculate annual travel benefits alone of $578 million (see this PDF for detailed breakdown of how we got this figure).

Environmental Effects:  21st Century Rail Travel Should Be Zero Carbon

The construction of a high-speed rail line would require a large environmental sacrifice – construction crews would need to shape the land, poor concrete, lay the tracks, and build the stations. This work would release millions of tons of carbon dioxide into the atmosphere. But building a new highway such as Texas’ planned I-69 would require similar work and would almost certainly be just as ecologically damaging. On a somewhat smaller scale, the same can be said for new terminals or runways at airports.

In a rapidly growing state like Texas, though, a serious need for a transportation capacity upgrade is bound to arise over the next decades – especially between the state’s two biggest cities. The construction of this infrastructure would require carbon emissions on a large scale–but since we don’t yet have competing plans for highway or airport capacity expansions if the high-speed system is not built, the most meaningful question for us is the rail system’s environmental effects in operations rather than construction.

Glaeser himself demonstrated the efficiency advantages of rail, showing that it releases less than a fifth of the emissions per passenger-mile of those of automobiles and less than a fourth of those of airplanes. Air travel emissions are particularly damaging to the environment because the nitrogen oxides and water vapor they release magnify the global warming effect.

Approximately 150 daily flights would be eliminated if the ridership model suggested above plays out. In comparison, Dallas’ Love Field hosts only 130 daily departures today. As such, the opening of the high-speed line would represent significant relief for airports and it would delay the need to expand terminals and runways, projects that are typically multi-billion-dollar operations.

The reduction in carbon emissions from people choosing not to drive cars or fly airplanes would be quite significant – especially if the rail system is powered by renewable energy. These savings are particularly evident on the very short flights on this corridor, such as from College Station to Houston or from Waco to Dallas, which could be replaced entirely with rail service.

Glaeser argues the power plants that produce the electricity used by high-speed trains would produce significant carbon emissions, reducing the environmental gain from switching away from air or car travel.

Yet he fails to account for the green potential of an electric rail line: it can operate without releasing any carbon at all. California, which is developing a 220 mph line between San Francisco and Los Angeles, has pledged to run its trains with electricity obtained only from carbon-neutral sources, such as wind turbines and solar panels. Texas could make the same commitment and dramatically expand the environmental benefits of the high-speed system. Texas is uniquely positioned to build such facilities, too – its western and northern sections are sunny, windy, and sparsely populated.

Needless to say, airplanes and automobiles cannot match high-speed’s promise of carbon-free long-distance travel. Just as significantly, this allows the system’s capacity to be ramped up – more and more trains, serving a growing population – with no additional carbon output.

With greater ridership, all the benefits that Glaeser calculates–based on better safety, congestion reduction, and lower carbon emissions–are all amplified. All told, we figure the annual environmental benefits would be $150 million in 2030 — compared to the $26 million Glaeser’s calculates for 2008.

Land Use: Give Density A Chance

With the exception of two flights a day to Atlanta, the small airports in Waco, Temple, and College Station each only provide service to Dallas and Houston today. If high-speed rail replaced plane circulation, these airports could be downsized and redeveloped for the public’s benefit. This is especially true for Easterwood Field in College Station, which is located just a few blocks from Texas A&M University. The use of this land for new housing or office space could reap significant new tax revenue for the city.

Glaeser himself assumes that high-speed rail would change the habits of some Americans: he argues that perhaps 100,000 households could be convinced to switch from suburban areas to inner city neighborhoods in both Dallas and Houston. Doing so provides a net benefit to society through carbon reduction.

Those estimates, however, are likely low since both metropolitan areas will be expanding by more than four million people by 2030. What if the state were proactive in encouraging dense, infill development around station cores? That straightforward strategy–which has a long history of success in cities around world–would both encourage ridership and amplify the projects benefits. It would also open up significant new economic opportunities for inner-city growth in those two cities as well as in Waco, Temple, and College Station. But Glaeser chooses to ignore that a high-speed system would attract offices and retail around stations. With both residential and commercial development being drawn to a downtown magnet rather than agglomerating in loosely around the cities’ outskirts, transit use and density would increase accordingly, both to society’s net benefit.

By plugging in a few of these conservative assumptions we get annual land use benefits of $72 million generate by a high speed rail project.

Conclusion: Build, Baby, Build

This reevaluation of Glaeser’s argument seems to upend his primary conclusion that the construction costs of the high-speed line would vastly outweigh the corridor’s benefits. While he figures that a 240-mile train system would result in a net annual loss of around $500 million, this analysis – using his own economic benefits model – shows a net benefit of $30 million a year (see our PDF for the math). High-speed rail between Dallas and Houston, then, seems like an eminently sensible thing to do.

See the full numerical breakout on this PDF.

Note: Car Travel Data from the 1995 American Travel Survey; updated by the author to compensate for the growth in the population of metropolitan areas. Population estimates based on existing population growth of 7% a year in San Antonio metro area; 8% in Dallas and Houston; and 9% in Austin. Car travel estimates for smaller cities (Waco, Temple, and College Station) calculated based on assumption of 50% car/air market distribution, which seems to be low-balling the data since the Houston-Dallas market is 3:2 car:air.

***

Yonah Freemark blogs about transportation and land use issues at The Transport Politic and is a regular contributor to The Infrastructurist.

Remember a few years ago when millions of our fellow Americans started gorging on bacon and cheeseburgers in order to lose weight? The Atkins diet fad was an odd moment in our culture and probably one best politely forgotten. But one reason the scheme took off like it did is that human beings are innately fascinated by counter-intuitive effects. Most examples you hear about on teevee–”Rock-hard abs without getting off your couch!”–are malarkey, of course. But in certain charmed cases, it is possible to get thin by eating lard, so to speak.
.
One example is reducing traffic congestion by eliminating roads. Though our transportation planners still operate from the orthodoxy that the best way to untangle traffic is to build more roads, doing so actually proves counterproductive in some cases. There is even a mathematical theorem to explain why: “The Braess Paradox” (which sounds rather like a Robert Ludlum title) established that the addition of extra capacity to a road network often results in increased congestion and longer travel times. The reason has to do with the complex effects of individual drivers all trying to optimize their routes. The Braess paradox is not just an arcane bit of theory either – it plays frequently in real world situation.

Likewise, there is the phenomenon of induced demand – or the “if you build it, they will come” effect. In short, fancy new roads encourage people to drive more miles, as well as seeding new sprawl-style development that shifts new users onto them.

Of course, improving congestion is not the main reason why a city would want to knock down a poorly planned highway–the reasons for that are plentiful, and might include improving citizen health, restoring the local environment, and energizing the regional economy. More efficient traffic flow is just a wonderful side benefit.

Sound dubious? Here are several examples of how three cities (and their drivers) have fared better after highways that should never have been built in the first place were taken down.

CASE 1: SEOUL, SOUTH KOREA – CHEONGGYCHEON HIGHWAY

In 2002, Mayor Lee Myung Bak pledged to renew South Korea’s capital Seoul by eliminating a 1970s-era highway that literally represented a paving over of the Cheonggyecheon River. His radical plan replacing it not with another road, but with a restored stream along the old riverbed. The immediate result of the intervention was a beautiful new 1000-acre park in the center of the city, lower pollution, cooler temperatures city-wide. What wasn’t expected, however, was the city’s reduced traffic volumes. After all, the road carried 160,000 cars a day before it was closed. But the highway’s closing was enough to convince thousands of people to drive less, or change their habits, as the city offered better public transportation options.

Before:

(Photo)

After:

(Photo)

CASE 2: PORTLAND, OREGON – HARBOR DRIVE

The idea that it’s possible to remove a major road without creating traffic jams is not exactly a recent one: Portland proved its merits more than 30 years ago. Until the early 1970’s–a period when the city’s now-thriving downtown area was losing the battle with urban blight–there was a four-lane freeway known as Harbor Drive occupying the western shore of the Willamette River, creating a barrier between the downtown area and the waterfront. Even citizens and a few politicians began arguing in favor of taking down the road in the late ’60s though, Oregon’s Highway Department wanted to widen the thoroughfare.

Ultimately, the most important advocate for its demolition was then-governor Tom McCall. After a long and contentious political battle, McCall prevailed and the highway was closed in 1974. On the first day it was shut off to traffic, one of the highway engineers who predicted gridlock catastrophe reportedly called one of McCall’s lieutenants to congratulate him: there hadn’t been “a ripple” of disturbance in the city’s traffic flow.

By 1978, a greenway occupied the land where the Harbor Drive once stood. Twice expanded since then, the Tom McCall Waterfront Park is an integral part of Portland’s success in recreating itself as a 21st century city.

Before:

(Photo)

After:

(Photo)

CASE 3: SAN FRANCISCO – EMBARCADERO FREEWAY

Arguably the US city where freeway removal has most improved urban life is San Francisco. The Embarcadero Freeway once stood elevated on the city’s waterfront. Two levels of concrete divided downtown from the bay. Though there had been a public push to demolish it since it was constructed, only after it was damaged in the 1989 Loma Prieta earthquake did the efforts crystallize, and it was never rebuilt. In its place today: a waterfront boulevard with bike trails, parks, and public exhibitions.

Before:

After:

(Photos)

CASE 4: SAN FRANCISCO – CENTRAL FREEWAY

Also damaged in the Loma Prieta quake was the Central Freeway, which ran as a spur into the city. The thoroughfare was closed in 1992 and a few years ago rebuilt as a surface road named Octavia Boulevard. Though the boulevard is well-used, it’s no more congested than the far larger highway that it replaced, showing that traffic responds the environment in which it is placed.

Before:

(Photo)

After:

Of course, this evidence doesn’t suggest that all road expansions are unnecessary, or that all highways should be removed. All of the highway demolitions cited above are in densely packed urban areas where other highways and reliable and convenient public transportation options are available. But the lesson is clear: If a major road is making a city a less livable and vital place that it would otherwise be, in many cases everyone benefits when politicians have the vision and guts to tear it down.