Gasoline Prices, Fuel Economy, and the Energy Paradox

It is often asserted that consumers undervalue future gasoline costs relative to purchase prices when they choose between automobiles, or equivalently that they have high “implied discount rates” for these future energy costs. We test this by examining how time series variation in gasoline prices a¤ects the relative prices of vehicles with di¤erent fuel economy ratings, using a detailed dataset based on 86 million transactions at auto dealerships and wholesale auctions between 1999 and 2008. Our preferred empirical estimate is that vehicle prices move as if consumers are indi¤erent between one dollar in purchase price and 72 cents in discounted future gas costs, although we document how plausible alternative assumptions can a¤ect this result. We introduce a new approach to behavioral welfare analysis in a discrete choice setting which generates two stark results. First, even if consumers undervalue gasoline prices by a relatively small amount, this distorts vehicle markets more than the failure to impose a Pigouvian carbon price. Second, even if consumers undervalue gas prices by a relatively large amount, CAFE standards increase average fuel economy by much more than can be justi…ed by undervaluation. JEL Codes: D03, L62, Q41. — — — — — — — — — — — — — — — — — — — — — — — — — — We thank Alberto Abadie, Josh Angrist, Orley Ashenfelter, David Autor, Lucas Davis, Henry Farber, Kelly Gallagher, Penny Goldberg, Michael Greenstone, Jerry Hausman, Mark Jacobsen, James Kahn, Ryan Kellogg, Lutz Kilian, David Laibson, David Lee, Henry Lee, Robin Lee, Lars Lefgren, Erin Mansur, Erich Muehlegger, Sendhil Mullainathan, Richard Newell, Ariel Pakes, Nancy Rose, Jesse Rothstein, Stephen Ryan, Jim Sallee, Jorge Silva-Risso, Rob Stavins, Chris Timmins, Roger von Haefen, Glen Weyl, Matt White, and seminar participants at Analysis Group, Carnegie Mellon, Colorado School of Mines, Harvard, Kent State University, Mathematica Policy Research, MIT, NBER Summer Institute, Northwestern, Ponti…cia Universidad Catolica de Chile, Princeton, Stanford University, Stanford Institute for Theoretical Economics, Texas A&M, University of California Energy Institute, University of Delaware, University of Wisconsin, University of Toronto, and Washington University for their feedback. Funding for this project was provided by Princeton’s Industrial Relations Section, Harvard’s Energy Technology Innovation Policy group, the Harvard University Center for the Environment, and the Sloan Foundation. We thank JD Power for providing vehicle transaction price data. Thanks to Lonnie Miller at RL Polk for helpful conversations. We thank Manheim Consulting for providing auction data and Gary George, Nirmeet Kacheria, Pete Sauber, and Tom Webb at Manheim for sharing their expertise. We thank Carmen Collyns and Jordon Ricks for research assistance on this and related projects. We especially appreciate meticulous research assistance by Sounman Hong. Code for replicating the estimation is available from Hunt Allcott’s website. Allcott: New York University and NBER. [email protected]. Wozny: Mathematica Policy Research. [email protected].