This dissertation consists of three chapters that analyze issues relating to automobile fuel economy. Chapter 1 discusses automobile fuel economy regulations in the United States. The new U.S. Corporate Average Fuel Economy (CAFE) standards not only tighten the target fuel economy to be achieved by automakers, but also make significant changes to the design/structure of CAFE standards by introducing three policy instruments (footprint-based targets, intra-firm transferring of fuel efficiency credits between passenger cars and light trucks, and inter-firm trading of fuel efficiency credits). While there are a number of previous studies on the impact of tightening CAFE standards, economists have paid little attention to the design of CAFE standards. I use policy simulation to evaluate these policy instruments relating to the design of CAFE standards. First, I model and estimate the demand- and supply-sides of the U.S. vehicle market using various data sets. Then, based on the estimation results, I simulate the vehicle market and the demand for driving under four counterfactual CAFE policies with different designs, and examine the impacts of the three policy instruments. Simulation results suggest: (1) footprint-based targets have little impact on market shares, producer profits, consumer surplus, and gasoline use; (2) inter-firm credit trading lowers overall compliance costs by about $110-140 million, and thus increases social welfare; and (3) allowing intra-firm credit transferring (but not inter-firm credit trading) reduces aggregate gasoline consumption by 0.1-0.25%.

Chapter 2 proposes a new approach to analyzing how automobile fuel economy is valued in the market, using a hedonic regression framework. A distinctive feature of my approach is the use of each vehicle's miles traveled: a consumer's marginal willingness to pay (MWTP) for fuel economy is inferred with her vehicle's miles traveled. With the inferred MWTP, we apply the steps of the standard hedonic method backward and estimate each vehicle's marginal and total price of fuel economy, and consumers' discount rate for future fuel cost savings. We find that the standard hedonic method may not provide a stable and reasonable estimate of the value of fuel economy, likely due to the omitted variable bias from vehicle attributes such as safety features, interior equipment and reliability. This method makes it possible to separate the portion of vehicle price that is attributable to fuel economy, and significantly alleviates the omitted variable bias. Applying the procedure to model year 2001 vehicles in the U.S. market, we estimate that consumers discount future fuel cost savings at the annual rate of 26-43%, that for the middle case of the discount rate of 34%, the price of a 0.1 gallon per 100 miles improvement in fuel efficiency is on average $75 (in 2000 U.S. dollars), and that for the same case, the average total price of fuel economy is $1,950. We also find that larger, less fuel efficient vehicles tend to have higher marginal and total prices of fuel economy.

Chapter 3 examines whether Japanese fuel economy regulations established in the 1990s induced technological progress in Japanese automakers' technology for providing fuel economy. By observing how fuel economy of automobiles has improved after controlling for changes in vehicle characteristics such as weight and power, I find that fuel economy improvement accelerated after regulations were introduced, implying induced innovation in fuel efficiency technology.