An important issue for future improvement and extensions of highways will be the ability of projects to sustain challenges to Environmental Impact Statements based upon forecasts of regional growth. A legal precedent for such challenges was established in 1997 when a U.S. District Court judge ruled that the EIS for a proposed Illinois toll road was deficient because the growth projections were the same in the build and no-build scenarios. This paper incorporates popular regional growth forecasting models into a quasi-experimental research design that directly relates new highway investments in three California counties to changes in population and employment location, while controlling for no-build historical counterfactuals. The authors model simultaneous employment and population growth from 1980 to 2000 in Merced, Orange, and Santa Clara counties, three California counties that received substantive highway improvements during the mid-1990s. The strategy permits a comparison of the before-and-after tests for effects of investments on economic growth and land use in three regions that contrast how increased highway access affects development patterns: (1) for an urban center in Santa Clara County, (2) for an exurban region in Orange County, and (3) for a small town in Merced County. We find that traditional forecast approaches, which lack explicit control selection, can lead to erroneous conclusions about an impact. Our integrated form of the lagged adjustment model confirms results from a conventional form of the model that includes all cross-sectional units as observations; in both forms of the model we estimate a statistically significant increase in employment development in the exurban region in Orange County where new toll roads were constructed. In the case of Santa Clara County, neither our quasi-experimental integrated approach nor the conventional lagged adjustment approach estimates a significant effect on population or employment growth that can be attributed to the new highways constructed in the urban center. For the small town environment in Merced County, the conventional simultaneous growth regressions produce a materially different estimate than the approach we develop and examine in this paper. Isolating effects to local spatial units where the intervention occurred and their no-build counterfactual produces estimates of a statistically significant decrease in employment growth in the small town near the newly constructed highway bypass.
R. G. Funderburg, Hilary Nixon, and M. G. Boarnet. "Linking highway improvements to changes in land use with quasi-experimental research design: A better forecasting tool for transportation decision-making" Mineta Transportation Institute (2009).