Evaluating US Fuel Economy Standards in a Model with Producer and Household Heterogeneity

Abstract

This paper employs an empirically estimated model to study the equilibrium effects of an increase in the US corporate average fuel economy (CAFE) standards. I identify and model heterogeneity across firms and find that the profit impacts of CAFE fall almost entirely on domestic producers. The welfare analyses consider the simultaneous household decision of vehicle and miles traveled, allowing direct comparison with a gasoline tax. Finally, I consider dynamic impacts in the used car market. I find these comprise nearly half the gross welfare cost of CAFE and fall disproportionately on low-income households. Contrary to previous results, the overall welfare costs are regressive. (JEL H24, L51, L62)

Vehicle Scrappage and Gasoline Policy

“…This allows us to demonstrate how our estimated scrap elasticities control the Gruenspecht effect in a general setting. More richly detailed simulations could identify favored and discouraged vehicles at the level of model names and configurations, producing analysis of specific rules in Europe, Japan, or the recently adopted "footprints" in the U.S. 22 …”

Substitution between Cars within the Household

“…Early studies include, for example, Hausman andNewey (1995), West (2004) and West and Williams (2005). More recent papers by Bento et al (2009) andJacobsen (2013) incorporate the supply side of the car market into the analysis, confirming that fuel prices and fuel efficiency standards have substantial distributional effects.…”

The effect of vehicle fuel economy standards on technology adoption

“…The European standards are based on weight, such that lighter vehicles are subject to lower emissions rate standards. 8 The fuel consumption rate decreased noticeably in 2010 for several manufacturers. Starting in 2010, the fuel consumption rate calculated from Wards data for flex-fuel vehicles corresponds to the fuel consumption rate using 85 percent ethanol rather than gasoline.…”

“…While this approach controls for segment-level shocks that affect technology adoption, if segment shocks happen to be correlated with stringency we would be concerned that subsegment shocks may also be correlated with stringency. We can assess whether segment shocks are correlated with stringency by omitting the triple interactions in equations (7) and (8). Appendix Tables 3 and 4 report the same specifications as in Tables 4 and 5, without the triple interactions.…”