Policy commentary

Should Automobile Fuel Economy Standards be Increased?

Sep 17, 2007 | Ian W.H. Parry

The U.S. Corporate Average Fuel Economy (CAFE) program requires auto manufacturers to meet minimum standards for the average fuel economy of their car and light-duty (minivan, pickup and sport utility vehicle) fleets. The standard for cars, last increased 22 years ago, is 27.5 miles per gallon, while that for light-duty trucks is rising to 23.5 miles per gallon by 2010. To many people, it seems a no-brainer that fuel economy standards should be tightened to reduce carbon dioxide (CO2) emissions and oil dependence. After all, passenger vehicles account for 20 percent and 45 percent of nationwide CO2 emissions and oil use, respectively.

Legislation to this effect will likely be introduced in the second session of the present Congress, where CAFE has been debated many times before. But before we can conclude whether or not tightening CAFE is a good idea, an economic assessment of the benefits and costs is appropriate. To think about this, it is helpful to separate out the effect of tighter standards on gasoline use, vehicle miles of travel, and the costs of automobile manufacture.

Higher fuel economy standards would reduce the demand for gasoline, thereby producing "externality" benefits (societal benefits that are not taken into account by individuals) in the form of avoided CO2 emissions and reduced nationwide dependence on oil. Most estimates of economic damages from future global warming--agricultural impacts, rising sea levels and increased storm intensity, health effects from spreading tropical disease, and so on--are around $5-20 per ton of current CO2 emissions, or 5-20 cents per gallon of gasoline (burning a gallon of gasoline produces nearly 0.01 tons of CO2). Obviously, damages are higher if more weight is given to the wellbeing of future generations or extreme climate risks.

The broader external costs of oil dependence include the risk of macroeconomic disruption costs from oil price shocks that might not be fully taken into account by the private sector, such as some costs associated with the temporary idling of labor and capital. And while the United States as a whole has an influence on the world oil market, individual oil importers do not consider the impact of their own infinitesimal consumption on increasing the world oil price, which imposes an external cost by increasing the amount of money transferred from other oil importers in the United States to foreign oil suppliers. Paul Leiby estimates that external costs from macroeconomic disruption risks and US market power amount to, very roughly, 30 cents per gallon of gasoline. Dependence on oil also constrains U.S. foreign policy and possibly undermines national security. Politicians may be reluctant to challenge oil-producing countries on human rights and other issues, and oil revenues may help certain hostile governments, terrorists, and other unsavory groups. Putting an additional dollar figure on these broader foreign policy and national security costs is extremely difficult however. (A later commentary in this series will discuss the military costs of protecting Persian Gulf oil supplies.)

Motorists already pay, at least in part, for the external costs of fuel consumption through federal and state gasoline taxes, which add about 40 cents per gallon to the price at the pump. According to basic tax theory, reducing gasoline use produces net benefits to society only to the extent that CO2 and oil dependence externalities exceed fuel taxes. Our discussion suggests, albeit very tentatively, that external costs that have been quantified (say, about 10 cents per gallon for CO2 and 30 cents for oil dependence based on the above figures) might be roughly offset by prevailing fuel taxes. However, accounting for national security and other costs would seem to imply net benefits overall from reducing gasoline use, though the magnitude of the gain is very difficult to gauge.

Critics of CAFE standards sometimes point to the perverse effect of higher fuel economy on lowering fuel costs per mile and increasing the incentive to drive, which can increase highway congestion, accidents, and pollution. However, according to a recent study by Kenneth Small and Kurt Van Dender, less than 10 percent of the fuel savings from better fuel economy are offset by increased driving; most likely, the costs of this "rebound effect" are probably fairly modest.

To the extent that fuel economy regulations are binding, they induce auto manufacturers to incorporate more fuel-saving technologies into new vehicles, leading to higher vehicle production costs and prices. However, a number of studies, such as one in 2002 by the National Research Council, suggest that fuel-saving benefits over the vehicle life would outweigh the up-front installation costs for many emerging technologies. Some analysts argue that these apparent "win-win" technologies may not be adopted without tighter CAFE regulations, however, because consumers may under-appreciate the benefits of better fuel economy if they are pre-occupied with other vehicle attributes like power, comfort, and safety. On the other hand, others argue that forcing technology adoption may be costly if consumers would instead prefer new technologies be used to improve other vehicle characteristics, such as increased horsepower, rather than fuel economy. Another possibility is that manufacturers may meet higher fuel economy requirements by reducing vehicle weight and size; this can raise injury risks for occupants of these vehicles, though it makes the roads a little safer for other drivers.

In short, the case for tightening CAFE can be argued either way, because it is difficult to judge precisely how manufacturers will respond, and how consumers will value changes in vehicle technology. But, most importantly, the climate and national security benefits from reduced gasoline use are so difficult to measure. Another policy option is to raise fuel taxes, which unlike CAFE, would reduce congestion and other highway externalities, through reducing vehicle miles traveled. While the case for higher fuel taxes is, in my view, more clear-cut, this option is still highly controversial.

When I first began studying CAFE, the case for tightening the standards looked rather dubious to me. However, my perspective has changed somewhat as the difficulties in doing a nice, clean cost-benefit analysis have become more apparent. Moreover, colleagues of mine who have thought hard about the issue--like Carolyn Fischer, Lawrence Goulder, Winston Harrington, Richard Newell, William Pizer, Paul Portney, Philip Sharp, and Kenneth Small--are sympathetic to higher standards, if they are not introduced too rapidly and reforms permit more trading of fuel economy credits (trading will be discussed in a later CAFE commentary). My own view is that if the argument comes down to doing nothing or tightening CAFE, then the latter is what you do. As new technologies are developed over time, a progressive tightening of CAFE seems to make sense, given that the downside costs are probably not that great, while 20 years from now we may be very glad that serious measures were taken during the intervening years to reduce the dependency of the transport system on conventional fossil fuels.

Views expressed are those of the author. RFF does not take institutional positions on legislative or policy questions.

To receive the Weekly Policy Commentary by email, or to submit comments and feedback, contact comments@rff.org

Further Readings: 

For an extensive overview of the CAFE program and emerging fuel saving technologies, see:

  • National Research Council, 2002. Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards. National Academy of Sciences, Washington, DC, National Academy Press.

For recent discussions of the costs and benefits of tightening fuel economy regulations, though with mixed findings, see:

For evidence on the rebound effect see:

For a recent discussion of the economic costs of global warming see:

For a recent discussion of the economic costs of oil dependence see: