New Highways & Urban Growth Patterns: Analyzing the Development Impact of the Orange County Toll Roads
Marlon G. Boarnet
Department of Urban and Regional Planning
UC-Irvine
Irvine, CA 92697-7075
Ph. 949-824-7695
Fax: 949-824-8566
Email: mgboarne@uci.edu
Funded by a UCTC Year 12 Research Grant
Overview:
The nature of the link between urban development and highway infrastructure is still poorly understood yet recent prominent discussions of the link between highways, urban decentralization and induced automobile travel have created a need to better understand the specific nature of any influence that new highways have on urban development. This research will uses econometric models of house sales prices and census tract population and employment growth to examine whether and how toll roads have changed land values and, by extension, development patterns in Orange County, California. The essence of the research is to carefully examine how house prices and census tract population and employment were influenced by the opening of the county's extensive toll road network. Such a test has never been done using advanced empirical techniques, with the extensive data that are currently available, and in the context of a road building project as extensive as the recent construction of the three major toll roads in Orange County. The results of this research will provide the first statistically and theoretically sound "before and after" test of the effect of highways on urban growth patterns.
Key Findings:
The results, taken collectively, provide strong evidence for the hypothesis that highway construction influences the development pattern of urban areas. Specifically, the research revealed the following:
Multiple regression of home sales prices showed that a negative house price gradient appeared after the initial links of the toll road system were built. That gradient was not present before the roads were constructed. As urban economic theory would predict, home prices declined with distance from the toll roads. The absence of this effect before the toll roads began construction suggests a causal influence – that the roads influenced land values.
An econometric model of population and employment location suggests that census tracts containing the toll roads were high population growth tracts, controlling for other factors, before the roads were built. Those census tracts remained high population growth tracts after the roads were built.
The same econometric model of population and employment growth suggests that the census tracts containing the toll roads were low employment growth tracts, controlling for other factors, before the roads were built. After the roads were built, the tracts containing the toll roads experience employment growth that, controlling for other factors, was not different from the county average.
Overall, the toll roads were sited in high population growth areas. Yet once the roads were built, there is evidence that the roads exerted an independent effect on both house prices (and by inference land values) and employment growth along the corridors near the roads.
References:
Boarnet, Marlon G. and Saksith Chalermpong, “New Highways, Induced Travel, and Urban Growth Patterns: A “Before and After” Test,” Final report submitted to
The Environmental Protection Agency and the University of California Transportation Center, October, 2002.
Note: The U.S. EPA and UCTC Year 12 funded different portions of the research on the growth impacts of the toll roads. The UCTC Year 12 funding was originally proposed to extend the U.S. EPA research by applying spatially weighted regression. During the course of the UCTC research, several more fundamental specification issues were uncovered, and the UCTC research was applied to examining how econometric specification issues influenced the results of the basic population-employment regression model. Because the house price model, the basic population-employment regression model, and the extended specification tests all are linked to the same research question, the detailed results are reported in one unified final report.
Note: The above paper won the Fannie Mae Foundation Award for best paper on housing or community development presented at the 2000 Association of Collegiate Schools of Planning meetings.
Acknowledgements: Saksith Chalermpong provided excellent research assistance throughout the course of this project Mike Greenwald also provided valuable research assistance at various stages. Robert Noland and Lewison Lem provided helpful comments at various stages. John Karevoll provided assistance in obtaining the house price data, which were provided by Dataquick, Inc. The California Department of Transportation and Geographic Data Technology, Inc. provided the GIS maps used in this research.