HIGHWAY FINANCE IN THE UNITED STATES: AN EMPIRICAL MODEL

This thesis seeks to construct an empirical model of highway finance in the United States, and in particular, to examine the relationship between highway-user revenues and highway spending. It provides a general overview of the current highway system, including the federal-aid highway program, and the flow of highway funds between different levels of government. It also examines issues relating to highway-user revenues. A review of the literature failed to provide any 'standard' model of highway spending and no previous studies of spending across all levels of government. Using data from the 50 states and the District of Columbia over the three-year period 1998-2000, regressions were run on the dollars spent on highways in each state from all levels of government. The independent variables included highway-user revenues (as defined by the Federal Highway Administration) in each state from all levels of government, lane-miles, daily vehicle-miles of traffic, land area, percent of land area classified as urban, population, gross state product, annual average wage, percent of traffic consisting of trucks, and average winter temperature. OLS estimates using the classical linear regression model were found to be unreliable, and attempts at using a growth rate model provided poor overall fit. Opportunities for future research are identified, as this is an important issue that should be of interest in public policy decision-making.