The aim of this research is to estimate a gasoline demand function for the U.S. This isachieved by applying the structural time series modelling technique over the period 1949-2005 but with a general specification that allows for both asymmetric price responses (fortechnical progress to impact endogenously) and an underlying gasoline demand trend (fortechnical progress and other factors to impact exogenously, but in a non-linear way). It isexpected therefore that price increases above previous historical levels result in a greaterresponse from consumers than a price increase below the previous maximum; which in turnwill result in a greater response (in absolute terms) than a price reduction. It is found that forU.S. per capita gasoline consumption the estimated income elasticity is 0.45, the estimatedprice-max elasticity is -0.20, the estimated price-recovery elasticity is -0.10, and the estimatedprice-cut elasticity is 0 (given it is insignificant). In addition, the estimated underlying energydemand trend for US gasoline demand is non-linear with periods when it is increasing andperiods when it is decreasing.
展开▼