Gasoline taxes can be employed to correct externalities associated with automobile use, to reduce dependency on foreign oil, and to raise government revenue. Our understanding of the optimal gasoline tax and the efficacy of existing taxes is largely based on empirical analysis of consumer responses to gasoline price changes. In this paper, we directly examine how gasoline taxes affect consumer behavior as distinct from tax-exclusive gasoline prices. Our analysis shows that a 5-cent tax increase reduces gasoline consumption by 1.3 percent in the short-run, much larger than that from a 5-cent increase in the tax-exclusive gasoline price. This difference suggests that traditional analysis could significantly underestimate policy impacts of tax changes. We further investigate the differential effect from gasoline taxes and tax-exclusive gasoline prices on both the intensive and extensive margins of gasoline consumption. We discuss implications of our findings for the estimation of the implicit discount rate for vehicle purchases and for the fiscal benefits of raising taxes.