Road Expansion, Urban Growth, and Induced Travel
Final Report - Summary
Robert Cervero
Dept. of City and Regional Planning
Institute of Urban and Regional Development
University of California
228 Wurster Hall
Berkeley CA 94720
University of California Transportation Center Year 12 (1999-2000)
DISCLAIMER
The contents of this report reflect the views of the authors, who are responsible
for the facts and the accuracy of the information presented herein. This document
is disseminated under the sponsorship of the Department of Transportation, University
Transportation Centers Program, in the interest of information exchange. The
U.S. Government assumes no liability for the contents or use thereof
Acknowledgements
The research on which this report is based was carried out for the University of California Transportation Center (UCTC). Funding for the Center comes in equal parts from the US Dept. of Transportation and the California Department of Transportation (Caltrans.) The University of California also provides support though reduction of overhead charges and the donated time of many faculty, students, and administrative staff. The researchers of the UCTC are grateful for all of this support.
Statement of the Problem
Few issues in the urban transportation field have sparked as much controversy and threatened proposed road projects as claims of “induced demand”. For decades, highway critics have charged that building new roads or expanding existing ones to relieve traffic congestion is a futile exercise. Improved roads simply spur additional travel or divert trips from parallel routes, quickly returning a facility to its original congested condition.
The preponderance of empirical evidence to date suggests that induced effects are substantial. One widely cited study, based on 18 years of data from 14 California metropolitan areas, found every 10 percent increase in lane miles was associated with a 9 percent increase in vehicle miles traveled (VMT) four years after road expansion, controlling for other factors. Similar findings have been recorded by others. Past research, however, can be faulted for failing to introduce a normative behavioral framework for tracing impacts, one that accounts for intermediate steps between road improvements and traffic growth and that allows for two-way causality.
Using data for a panel of California freeways, this study sought to fill past methodological gaps by postulating and empirically testing a path model of induced travel. A short-run model, which focuses on relationships within a one-year time frame, holds that changes in road supply affect travel speeds, which nearly instantaneously affect traffic levels. In contrast to most recent analyses of induced demand that measure VMT growth as a direct function of lane-mile additions, this analysis introduces an important intermediate step – namely, that road improvements confer benefits, in the form of higher travel speeds, and that it is changes in operating conditions that influence demand, not the physical attributes (e.g., lane miles) of a project. A longer-run model traces how road investments induce major building activities over a multi-year time horizon, and how resulting land-use shifts in turn lead to increased travel. A feedback loop is also modeled, capturing how traffic growth influences road investment decisions. The model postulates that the combination of current operating speeds on a roadway and previous-year changes in urban development influence current-period demand levels. Both factors are thought to increase VMT -- the former in the near term, the latter over the longer run.
2. Toward a Normative Theory: A Path Model
A path model proposed for tracing the effects of road improvements on travel demand as well as urban development is presented in Figure 1. The figure also accounts for “induced investment” effects. Notably, changes in a project’s share of countywide VMT over time can be expected to influence future shares of countywide road improvements targeted at the corridor, as will trends in travel speeds. Roads not only stimulate but also respond to demand.
Figure 1. Hypothesized Path Model
The near-term model – i.e., the influences of freeway expansions on VMT are channeled
through an intermediate step, operating speeds – was confirmed. From the econometric results, every 10 percent increase in a facility’s share of countywide freeway lane mileage was associated with a 4.2 percent increase in mean operating speed on that facility. Rising travel, however, eroded some of the speed benefits conferred by a road. Based on elasticity estimates, VMT increases were not totally offsetting – that is, the speed-enhancing benefits of freeway expansions exceeded the speed-eroding impacts of rising VMT. For the near term analysis, an “induced travel” elasticity of 0.238 was found, considerably smaller than elasticities estimated in previous county-level studies drawn from California experiences that used lane-miles as a direct predictor. In that some of the recorded travel increases represented route diversions, the “induced demand” elasticity of newly produced VMT is likely even smaller.
The longer-run analysis substantiated the existence of an “induced growth” effect.
The share of countywide building square footage and valuations along a corridor increased with the share of countywide freeway lane-mileage added three years earlier. Building activities were also found to be highly responsive to average operating speeds two years before.
By type of land uses, home-building activities were most responsive to the addition of roadway capacity. Lane-mile additions two to four years previously, and in the case of apartments and multi-family units, operating speeds two years earlier, significantly explained residential construction, with elasticity estimates well above one. Non-residential activities were most responsive to changes in operating speeds two to four years previously, with lane-mile additions three to four years earlier exerting more modest effects on office, industrial, and public-use construction.
The longer-run analysis yielded an elasticity estimate that, consistent with theory and past research, was considerably higher than short-run effects -- 0.637. The longer-term model also reveals that a smaller but nonetheless appreciable increase in VMT is attributable to heightened development activity along impacted corridors. Still, the influences of behavioral shifts (e.g., latent trips, modal changes, route diversions) was found to be nearly four times as strong as those of structural changes (e.g., land-use shifts).
Lastly, a significant induced-investment effect was also uncovered. Every 10 percent increase in the share of countywide VMT on a corridor two years previously is associated with a 4.9 percent increase in the current share of countywide lane-mile capacity, ceteris paribus. Besides VMT levels, previous-year operating speeds were also statistically associated with freeway expansion. The fact that variables measuring both VMT and operating speeds appeared as direct and statistically significant predictors of freeway expansion could reflect
the influences of multiple criteria in investment decisions – that is, a combination of both traffic growth and performance levels could have played into political decisions to expand freeway capacity.
Conclusion
Fairly firm positions have been taken on the induced demand debate despite the methodological shortcomings of past research. Simple mode structures have often been used to reach the conclusion that road investments provide only ephemeral congestion relief, with most added road capacity absorbed by increases in traffic. The path model presented in this research study attempted to sort through the ways in which road improvements affect travel demand, and vice-versa. As with past research, evidence of induced demand, induced growth, and induced investment was uncovered. While the magnitude of induced growth effects found in this study is generally consistent with that of previous research, the magnitude of induced demand effects is generally less. To the degree the path model better captures causal relationships than previous studies, many past elasticity estimates are likely inflated.
The problems people associate with roads – congestion, air pollution, and the like – are not the fault of road investments per se. These problems stem mainly from the unborne externalities from the use of roads, new and old alike. They also stem from the absence of thoughtful and integrated land-use planning and growth management around new interchanges and along new corridors. While the induced demand phenomenon is important and not to be trivialized, far more energies need to go toward figuring out how to best invest and manage scarce transportation and land resources – e.g., should we be building more bus rapid transit systems, applying “value-pricing” on current carpool lanes, and more closely integrating transportation and land use, and if so, when, where, and under what conditions? Whether new roads are on balance beneficial to society cannot be informed by studies of induced demand, but rather only through a full accounting and weighing of social costs and benefits.
Cervero, R. 2001. Induced Travel Demand: An Urban and Metropolitan Perspective. Washington, D.C.: Paper presented at the Conference on Working Together to Address Induced Demand, Eno Transportation Foundation and the U.S. Environmental Protection Agency.