Travel Demand and Location Efficiency
As revealing as were earlier studies of travel demand, they were limited by the lack of data on transportation choices made at the household level. Later studies have therefore gone to great lengths to more closely scrutinize the same relationships with fine-grained, neighborhood-level data.
This has necessarily involved the laborious compilation of new information. John Holtzclaw, in a 1994 paper, "Using Residential Patterns and Transit to Decrease Auto Dependence and Costs," developed a methodology for predicting household automobile travel from density and transit access in 28 California communities. His work later became part of an analysis conducted collaboratively by the National Resources Defense Council, the Center For Neighborhood Technology and the Surface Transportation Policy Project, calculating the transportation value, or "location efficiency," of a given place.
The Center For Neighborhood Technology, in cooperation with the Natural Resources Defense Council and the Surface Transportation Policy Project, developed a model to predict vehicle miles traveled in the Chicago, San Francisco and Los Angeles metropolitan areas in 1997. While earlier work, such as that carried out by Pushkarev and Zupan, looked at metropolitan regions on a city-wide scale, the LEM and subsequent modeling was able to predict vehicle miles traveled for small geographies, in this case traffic analysis zones in San Francisco and Los Angeles, and quarter sections in Chicago.
Such a focus on small scales allowed as many variables as possible to be accounted for, thus removing suspicions that factors other than density (such as income level, geography, or culture) influenced travel choices. "Direct comparison of neighborhoods is necessary," Holtzclaw writes, "to determine if neighborhood characteristics like density, transit service and pedestrian and bicycle friendliness characteristics that can be influenced by public policy truly influence auto ownership and driving."
The model designed by Holtzclaw and colleagues predicts household vehicle ownership and use based on household income and size, vehicle ownership, residential density, block size (used as a surrogate for pedestrian accessibility), vehicle miles traveled, transit routes, and frequency of transit service. These factors are brought together in a statistical model to describe the transportation efficiency attributable to a location: the degree to which any trip can be made quickly and efficiently. High levels of efficiency indicate conditions favorable to transit, and to high levels of pedestrian activity. Not surprisingly, in such circumstances, people consistently own fewer cars, drive less, and therefore produce fewer emissions.
The LEM study marked an advance in three respects: Geographic Information Systems unavailable prior to the 1980's allowed land use patterns and their effects to be made plainly visible in cartographic form; the massive collection of household data in three cities allowed for trip origins (rather than total trips) per household to be tightly correlated to residential density; and the relative cost to households having to make more trips.
By incorporating into statistical analysis the travel habits of different income groups, as well as neighborhood-level data from geographically and historically distinct cities (Chicago, San Francisco, and Los Angeles), the 2000 location efficiency study found that the strong inverse correlation of residential density with auto ownership held true across three distinct urban environments. (Figure 3.1) "Urban design and transportation infrastructure," concludes the location efficiency study, "have a highly significant influence on auto ownership and distance driven for neighborhoods," thus refining the twenty-five year old insight of Pushkarev and Zupan, and moving beyond it with the introduction of the concept of location efficiency into discourse on travel demand.