A lesson in "phantom" trips.
Some of the most trusted planning tools used to manage vehicular traffic have shown themselves to be pretty harmful to city life in certain ways. A metric known as Level of Service, which aims to minimize automobile delay at an intersection, can act as a huge obstacle to public transportation projects. A design book calling for 12-foot lanes, an engineering staple across the country, can speed up car flows and endanger public safety as a result.
It might be time to add one more established tool to the questionable list: the Trip Generation Manual from the Institute for Transportation Engineers, a common guide that tells traffic planners how many car trips will be generated by a new commercial or residential development project.
That's the argument made by environmental scholar Adam Millard-Ball of UC-Santa Cruz, who challenges the merits of the Trip Generation Manual in an upcoming research paper (nicely summarized in ACCESS magazine). Millard-Ball reports that the ITE manual may overestimate the number of trips generated from a new development by as much as 55 percent—"phantom trips," he calls them. The result is that cities may build way more roads than necessary, perpetuating sprawl and leaving less street space for non-drivers in the process.
"[The ITE manual] is used pretty much everywhere for small or medium-sized development," Millard-Ball tells CityLab. "In some places it's a starting point rather than a final word, though I think that's still pretty rare."
A quick primer for non-engineers out there. The Trip Generation Manual tells planners how much weekday traffic they can expect to emerge from various types of local development. So a 24-hour convenience store, for instance, will put about 738 new daily trips on the road for each 1,000 square feet of retail space. The project types get very specific: from "automated car wash" to "baby superstore" to "coffee/donut shop with drive-through window and no indoor seating."
Ideally, the manual helps traffic planners prepare a road network for all the new trips—adding a lane here, perhaps, or a turn signal there. But while the tool is supposed to be reserved for suburban, car-centric areas with negligible transit shares, it's often misappropriated as a blanket guide. Some cities do adjust trip estimates for higher-density areas or use other measures they consider more precise, says Millard-Ball, but the ITE manual is nevertheless a standard tool for smaller-scale development.
"Maybe it's a five-to-ten unit apartment building, and the cost of having some consultants come in to run a more sophisticated model is completely disproportionate," he says. "That's when you just look up the trip generation rates in the ITE manual."
For his analysis, Millard-Ball compared the number of development-related trips estimated by the ITE manual to the number of trips people report taking in actual household travel surveys. As the table below shows, the ITE manual predicts people will take many more trips than they truly do, pretty much across the board. Take an average school. Whereas the ITE manual predicts it will generate about 41 million trips a year, the 2009 household travel survey suggests the real trip number is closer to 13.7 million—overestimating traffic by 198 percent.
Altogether the manual overshot new trips related to residential development by 56 percent and to commercial development by 54 percent, according to Millard-Ball's calculations, with total trips overestimated by 55 percent.
That trend holds true in pretty much every metro area across the country. The figure below shows residential trips per housing unit made each weekday in 2009 in more than 50 U.S. metros. In nearly every city, the number of trips reported by household surveys (the thick bars on the left) came in significantly lower than the number estimated by the ITE manual (the thin bars on the right). For the entire country, the average household reported about five trips per weekday, against the ITE prediction of roughly eight.
Here's why those "phantom trips" matter: the more car trips that local planners believe will emerge from a new development, the more road space they dedicate to cars. This unnecessary space creates higher taxpayer maintenance costs and reduces room on the road network for pedestrians or alternative transport modes. Phantom trips can also encourage sprawl. If the manual estimates more new trips than the local road network can accommodate, for instance, developers might prefer to move the project to a more remote area.
"Certainly the inclination of cities is to err on the side of estimating higher trips," says Millard-Ball. "It's conservative in this context to expect more car traffic, even if another way of interpreting conservatism is that you shouldn't build roadway infrastructure or widen lanes unless you're absolutely sure they'll be needed."
The fundamental reason the manual is such a poor predictor of trips, he says, is its flawed core premise that new development automatically generates new trips. Take a new courthouse, for instance. Building the court doesn't inherently increase the number of crimes or legal procedures that occur in an area. It's far more likely that people who'd already be traveling to some court will simply go to this new one.
In other words, the courthouse relocates old trips far more than it generates new trips. Distinguishing truly new trips from merely relocated ones is the first step toward more accurate estimates and fewer unnecessary roads.
But to frustrated drivers across the country, Millard-Ball's insights raise an obvious question: If metro areas are building more roads than necessary based on "phantom" trips, then how come there's still so much non-phantom traffic on city streets? Part of the answer, he says, is that new road space alone encourages new traffic—a concept known as "induced demand." So even though new development itself may not generate as much traffic as the ITE manual suggests, the new roads created to accommodate that development will entice some people to drive.
"It's not that we'd have induced traffic all the way to the level of the original prediction, but there would likely be some modest increase in traffic because we're catering to phantom trips," he says. "To some extent, it's a case of if we build it, they will partially come."