Modeling gasoline demand in the United States: A flexible semiparametric approach

- Model the demand for gasoline in the U.S. using a semiparametric approach.
- Using several instruments, the state level gasoline price is found to be exogenous.
- The price/income elasticity of gasoline demand varies across states and over time.
- States with more trucks, higher population density, etc. have more elastic demand.
- Gasoline demand becomes more elastic over time due to several time-varying factors.

The focus of this paper is on the modeling and estimation of quarterly state-level gasoline demand in the United States. The existing literature may not appropriately evaluate the price elasticity and income elasticity of gasoline demand. Most studies fail to address the possible heterogeneity in gasoline demand elasticities that may arise from a variety of sources. The endogeneity issue of gasoline price has remained redundant throughout the literature. I address these challenges using a flexible demand model and a recently developed estimation technique. The econometric approach allows for functional coefficients to accommodate the heterogeneity in demand elasticities. Several instrumental variables are used to investigate the endogeneity of gasoline price. The estimation results provide strong evidence of heterogeneous gasoline demand elasticities across states and over time. Some state-level attributes along with income and macroeconomic shocks are the potential sources of heterogeneity.