| This dissertation looks at both the supply and the demand side of the market for gasoline. We rely on the Residential Energy Transportation Energy Consumption Survey (RTECS) for gasoline demand data and the Monthly Report of Prime Supplier Sales of Petroleum Products for Local Consumption for gasoline supply data. Both are collected by the Energy Information Administration.; Schmalensee and Stoker (1999) utilized nonparametric modeling to specify a regression structure of the 1988 and 1991 RTECS data. Chapter 1 applies this methodology to the 1994 data to determine the structure of gasoline demand with respect to age. We then estimate elasticities of demand with respect to income and the number of drivers in the household for each of 36 possible geographic locations. We find that income elasticities are generally much smaller than number-of-driver elasticities.; Chapter 2 tests a simple model of gasoline demand using all three RTECS. The major question is whether variation in gasoline demand is due to miles traveled or vehicle fuel efficiency. Our results strongly suggest that miles driven is the major source of variation. Miles driven, in turn, is most affected by geographic location, the age of the adults in the household, and the number of drivers in the household.; Chapter 3 considers the relationship between market power of gasoline producers and environmental regulation of gasoline. We show that a region with less strict environmental standards can realize lower gasoline prices by harmonizing its requirements with those of a region with more strict environmental regulation. The reduction in market power due to entry of new firms to the combined market can more than offset the higher marginal cost. The result is expressed as an upper limit of the ratio of the cost of cleaner gasoline to the cost of dirtier gasoline. The EIA prime supplier data is used to inform parameter choices for numeric simulations of the limit of the ratio of marginal costs. The simulations are used to determine how the result is affected by price elasticity of demand, number of suppliers, and the probability of refinery outage. |
