Household Gasoline Demand in the United States

Continuing rapid growth in U.S. gasoline consumption threatens to exacerbate environmental and congestion problems. We use flexible semiparametric and nonparametric methods to guide analysis of household gasoline consumption in 1988 and 1991. The number of licensed drivers has a strong effect on consumption, and including this variable cuts the estimated income elasticity in half. Slower projected future growth in licensed drivers points to slower growth in gasoline consumption. A parsimonious representation of age, income, lifecycle and location effects is developed and tested. We show how flexible methods also helped reveal fundamental problems with the available price data. * Sloan School of Management, E52-455, Massachusetts Institute of Technology, Cambridge, MA 02139. This research was supported by the MIT Center for Energy and Environmental Policy Research, the U. S. Department of Energy and the National Science Foundation. The authors wish to acknowledge numerous helpful comments from A. Denny Ellerman, Howard Gruenspecht, Edward D. Porter and the audiences of the CEEPR Spring meetings, the Neyman Seminar at the Department of Statistics, University of California at Berkeley and the Institute for Fiscal Studies in London, England.