The Impact of Energy Prices on Sprawl

Abstract

This dissertation uses ordinary least squares regression analysis to measure the impact of gasoline price changes on sprawl defined as land density of population, employment and light visible from space over multiple decades. Using census data compiled and normalized to 2000 boundaries in the Geolytics Neighborhood Change Database (NCDB) for the decennial census years 1970 through 2000 and energy price data published by the Energy Information Agency (EIA), it tests the effect of gasoline prices on eight population density-derived measures of sprawl for 72 Primary Metropolitan Statistical Areas and 48 states. Using data contained in the Zipcode Business Patterns (ZPB) database maintained by the Census Bureau, it tests the impact of gasoline prices published by ACCRA on a variety of employment dispersion measures of sprawl for Metropolitan Statistical Areas (MSAs). And using a dataset created by Burchfield et al of light visible from space, it tests the effect of gasoline prices published by ACCRA on the Burchfield et al sprawl index for 275 MSAs. The regressions control for a variety of socioeconomic, demographic, legal and physical variables and use instrumental variables and other techniques to test for causal relationships.

The dissertation finds robust evidence that gasoline price changes did influence sprawl at the 95% confidence level over the periods studied with higher prices leading to less marginal sprawl and lower prices leading to more marginal sprawl. In turn, it discusses ways that policymakers might intervene using this pathway to influence sprawl, notably through the imposition of gasoline taxes.