National Academies Press: OpenBook

Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards (2002)

Chapter: 5 Potential Modifications of and Alternatives to CAFE

« Previous: 4 Impact of a More Fuel-Efficient Fleet
Suggested Citation:"5 Potential Modifications of and Alternatives to CAFE." Transportation Research Board and National Research Council. 2002. Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards. Washington, DC: The National Academies Press. doi: 10.17226/10172.
×
Page 83
Suggested Citation:"5 Potential Modifications of and Alternatives to CAFE." Transportation Research Board and National Research Council. 2002. Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards. Washington, DC: The National Academies Press. doi: 10.17226/10172.
×
Page 84
Suggested Citation:"5 Potential Modifications of and Alternatives to CAFE." Transportation Research Board and National Research Council. 2002. Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards. Washington, DC: The National Academies Press. doi: 10.17226/10172.
×
Page 85
Suggested Citation:"5 Potential Modifications of and Alternatives to CAFE." Transportation Research Board and National Research Council. 2002. Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards. Washington, DC: The National Academies Press. doi: 10.17226/10172.
×
Page 86
Suggested Citation:"5 Potential Modifications of and Alternatives to CAFE." Transportation Research Board and National Research Council. 2002. Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards. Washington, DC: The National Academies Press. doi: 10.17226/10172.
×
Page 87
Suggested Citation:"5 Potential Modifications of and Alternatives to CAFE." Transportation Research Board and National Research Council. 2002. Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards. Washington, DC: The National Academies Press. doi: 10.17226/10172.
×
Page 88
Suggested Citation:"5 Potential Modifications of and Alternatives to CAFE." Transportation Research Board and National Research Council. 2002. Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards. Washington, DC: The National Academies Press. doi: 10.17226/10172.
×
Page 89
Suggested Citation:"5 Potential Modifications of and Alternatives to CAFE." Transportation Research Board and National Research Council. 2002. Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards. Washington, DC: The National Academies Press. doi: 10.17226/10172.
×
Page 90
Suggested Citation:"5 Potential Modifications of and Alternatives to CAFE." Transportation Research Board and National Research Council. 2002. Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards. Washington, DC: The National Academies Press. doi: 10.17226/10172.
×
Page 91
Suggested Citation:"5 Potential Modifications of and Alternatives to CAFE." Transportation Research Board and National Research Council. 2002. Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards. Washington, DC: The National Academies Press. doi: 10.17226/10172.
×
Page 92
Suggested Citation:"5 Potential Modifications of and Alternatives to CAFE." Transportation Research Board and National Research Council. 2002. Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards. Washington, DC: The National Academies Press. doi: 10.17226/10172.
×
Page 93
Suggested Citation:"5 Potential Modifications of and Alternatives to CAFE." Transportation Research Board and National Research Council. 2002. Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards. Washington, DC: The National Academies Press. doi: 10.17226/10172.
×
Page 94
Suggested Citation:"5 Potential Modifications of and Alternatives to CAFE." Transportation Research Board and National Research Council. 2002. Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards. Washington, DC: The National Academies Press. doi: 10.17226/10172.
×
Page 95
Suggested Citation:"5 Potential Modifications of and Alternatives to CAFE." Transportation Research Board and National Research Council. 2002. Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards. Washington, DC: The National Academies Press. doi: 10.17226/10172.
×
Page 96
Suggested Citation:"5 Potential Modifications of and Alternatives to CAFE." Transportation Research Board and National Research Council. 2002. Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards. Washington, DC: The National Academies Press. doi: 10.17226/10172.
×
Page 97
Suggested Citation:"5 Potential Modifications of and Alternatives to CAFE." Transportation Research Board and National Research Council. 2002. Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards. Washington, DC: The National Academies Press. doi: 10.17226/10172.
×
Page 98
Suggested Citation:"5 Potential Modifications of and Alternatives to CAFE." Transportation Research Board and National Research Council. 2002. Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards. Washington, DC: The National Academies Press. doi: 10.17226/10172.
×
Page 99
Suggested Citation:"5 Potential Modifications of and Alternatives to CAFE." Transportation Research Board and National Research Council. 2002. Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards. Washington, DC: The National Academies Press. doi: 10.17226/10172.
×
Page 100
Suggested Citation:"5 Potential Modifications of and Alternatives to CAFE." Transportation Research Board and National Research Council. 2002. Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards. Washington, DC: The National Academies Press. doi: 10.17226/10172.
×
Page 101
Suggested Citation:"5 Potential Modifications of and Alternatives to CAFE." Transportation Research Board and National Research Council. 2002. Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards. Washington, DC: The National Academies Press. doi: 10.17226/10172.
×
Page 102
Suggested Citation:"5 Potential Modifications of and Alternatives to CAFE." Transportation Research Board and National Research Council. 2002. Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards. Washington, DC: The National Academies Press. doi: 10.17226/10172.
×
Page 103
Suggested Citation:"5 Potential Modifications of and Alternatives to CAFE." Transportation Research Board and National Research Council. 2002. Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards. Washington, DC: The National Academies Press. doi: 10.17226/10172.
×
Page 104
Suggested Citation:"5 Potential Modifications of and Alternatives to CAFE." Transportation Research Board and National Research Council. 2002. Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards. Washington, DC: The National Academies Press. doi: 10.17226/10172.
×
Page 105
Suggested Citation:"5 Potential Modifications of and Alternatives to CAFE." Transportation Research Board and National Research Council. 2002. Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards. Washington, DC: The National Academies Press. doi: 10.17226/10172.
×
Page 106
Suggested Citation:"5 Potential Modifications of and Alternatives to CAFE." Transportation Research Board and National Research Council. 2002. Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards. Washington, DC: The National Academies Press. doi: 10.17226/10172.
×
Page 107
Suggested Citation:"5 Potential Modifications of and Alternatives to CAFE." Transportation Research Board and National Research Council. 2002. Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards. Washington, DC: The National Academies Press. doi: 10.17226/10172.
×
Page 108
Suggested Citation:"5 Potential Modifications of and Alternatives to CAFE." Transportation Research Board and National Research Council. 2002. Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards. Washington, DC: The National Academies Press. doi: 10.17226/10172.
×
Page 109
Suggested Citation:"5 Potential Modifications of and Alternatives to CAFE." Transportation Research Board and National Research Council. 2002. Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards. Washington, DC: The National Academies Press. doi: 10.17226/10172.
×
Page 110

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

83 5 Potential Modifications of and Alternatives to CAFE WHY GOVERNMENTAL INTERVENTION? Why should the government intervene in the fuel econ- omy decisions of consumers and manufacturers? This sec- tion discusses the underlying rationales. Environmental and International Oil Market Impacts Fuel economy decisions can be distorted if the market price of gasoline—the price that motivates decisions—fails to take account of the environmental impacts of gasoline use, the impacts of oil consumption on world oil prices, or the impacts of oil consumption on vulnerability to oil market disruptions. And, absent intervention, the resulting distor- tions would result in a fleet of new vehicles with fuel economy lower than what is optimal for the United States as a whole. Appropriately designed and scaled interventions can successfully mitigate these distortions and thereby en- hance overall welfare. This chapter examines the appropri- ate scale of interventions and explores alternative policy in- struments that could reasonably be expected to enhance overall welfare. The primary environmental issue is the emissions of car- bon dioxide (CO2), a normal combustion product of hydro- carbons, and the resultant impacts of atmospheric accumula- tions of CO2 on global climate change. 1 The amount of CO2 released from driving is directly proportional to the amount of gasoline consumed. There is also an environmental cost associated with releases of hydrocarbons and toxic chemi- cals from the gasoline supply chain. The second issue is the impact of increased oil consump- tion on the world oil market and on oil market vulnerability. The price of oil imported into the United States exceeds the competitive level because of the Organization of Petroleum Exporting Countries’ (OPEC’s) market power; greater U.S. oil consumption could further increase the import price. In addition, international oil market disruptions could lead to economic losses in the United States. The greater the con- sumption of petroleum products, the more vulnerable the United States is to such disruptions. These factors together imply that there are costs of in- creasing gasoline use in addition to those seen by the indi- vidual consumer. These additional costs are referred to col- lectively as externalities (external costs). Since the rationale for fuel-use-reducing market interven- tions is the existence of external costs, the magnitude of these external costs determines the appropriate scale, or strength, of the interventions. Economic efficiency requires that con- sumers face the full social cost (including the external cost) associated with gasoline use, or be induced to act as if they faced those full costs. Therefore, quantification of these ex- ternal costs is important for policy analysis. A later section discusses this quantification. Unresolved Issues of Governmental Intervention Some analysts argue that, even in the absence of any en- vironmental and international oil market impacts, the United States should intervene in automobile markets to require higher fuel economy than would be chosen by manufactur- ers and consumers absent market intervention. These ana- lysts argue that there are reasons to believe that the market choices for fuel economy are not efficient, even absent these externalities.2 The net value of major increases in fuel economy is, at most, some hundreds of dollars to new car buyers even if 1Vehicles also emit criteria pollutants, but these pollutants are regulated on a grams-per-mile basis, with allowable emissions not dependent on auto- motive fuel economy. Therefore, varying the fuel economy of new vehicles is unlikely to cause significant variations in emissions of criteria pollutants. 2An intermediate position, held by at least one committee member, is that there should be no intervention absent environmental and international oil market impacts, but that when one combines environmental externalities and oil market problems with the imperfections in the markets for fuel economy, the case for action becomes strong.

84 EFFECTIVENESS AND IMPACT OF CORPORATE AVERAGE FUEL ECONOMY (CAFE) STANDARDS fuel savings over the entire life of the vehicle are considered. This is because the net value is the difference between the discounted present value of fuel savings and the cost to the consumers of added technologies to achieve these savings. But if buyers consider the fuel costs only over the first few years they intend to own the vehicle, the perceived net value from these costly changes could be very low—even nega- tive—and consumers might prefer to not pay for such changes. In order to implement significant fuel economy increases, manufacturers must completely redesign all the vehicles they make. Manufacturers would decide whether to make very expensive and risky investment decisions, in the expectation of a small, uncertain advantage. According to this view, there is a lot of inertia in the market choices determining fuel economy, and one cannot be sure that manufacturers and consumers would ever arrive at the optimal fuel economy level. In addition, although there is good information available to consumers on the fuel economy of new vehicles, the in- formation is not perfect. And, since consumers do not know what the future price of fuel will be, they may underestimate or overestimate future fuel costs. Consumers typically do not actually compute the discounted present value of fuel savings before buying a car. Most importantly, there is no pure “price” of higher fuel economy facing car buyers; in- stead, they must infer how much greater fuel economy will cost by comparing different vehicles. Thus, these consumers may buy vehicles with fuel economy that is higher or lower than what they would have chosen had perfect information been available. Finally, from this point of view, although automobile companies compete intensely with each other, the automo- bile market is not perfectly competitive—in fact, it is more adequately described as oligopolistic. In oligopolistic mar- kets, companies may choose levels of fuel economy that are higher or lower than they would have chosen if the markets were perfectly competitive. For all of these reasons, this viewpoint maintains that there is no guarantee that markets will achieve economically optimal levels of fuel economy; rather, the levels could be either too high or too low. Committee members differ in their beliefs about the quan- titative importance of these issues and whether they justify government intervention to regulate fuel economy. Those supporting the viewpoint described here suggest that because automobile manufacturers might systematically produce vehicles with lower-than-optimal fuel economy, the government should intervene in the markets and should re- quire manufacturers to increase fuel economy to some “cor- rect” level. This viewpoint requires calculation, external to the automobile manufacturers, of the correct level, at least for each type of vehicle. That correct level of fuel economy might be taken to be the cost-efficient level—that is, the level at which the estimated cost of additional fuel economy im- provements would be just equal to the estimated discounted present value of additional fuel cost savings over the entire life of the vehicle, using some estimate of future gasoline prices and some specified discount rate for future fuel cost savings. These fuel economy levels might correspond to the 14-year case described in Chapter 4, if all assumptions un- derlying that calculation turned out to be accurate. Those rejecting the viewpoint argue that it is in the inter- ests of automobile manufacturers to estimate the preferences of their customers and others they wish to attract as their customers. Therefore, manufacturers provide levels of fuel economy that, in their estimation, best reflect the trade-offs potential customers would make themselves. But manufacturers realize, too, that vehicle buyers differ greatly from one another, including in the trade-offs they are willing to make between vehicle purchase price and fuel economy. Accordingly, the various manufacturers offer dif- ferent makes and models for sale, with a range of fuel econo- mies. Potential customers are free to choose vehicles that correspond to their particular preferences. Some will wish to purchase vehicles with fuel economy corresponding to the 14-year case. Others value vehicle purchase price and may prefer vehicles that use more gasoline but are less expensive. For them, the best choice might be a vehicle with fuel economy corresponding to the 3-year case from Chapter 4. From this perspective, if the government requires fuel economy to correspond to the 14-year case, then those people who prefer to purchase vehicles corresponding to the 3-year case would be harmed. They would have to pay more money to purchase a vehicle. Although they would subsequently spend less on gasoline, the gasoline savings would not be sufficient to compensate them for the increased vehicle pur- chase price. Conversely, if the government were to require fuel economy to correspond to the 3-year case, then those who would prefer to purchase vehicles corresponding to the 14-year case would be harmed. Although they would save money on new vehicle purchases, the savings would not be sufficient to compensate them for the additional price they would pay for gasoline over the life of the vehicle. This per- spective notes that absent government intervention, each type of consumer can be satisfied because the competing manufacturers will offer a range of options from which con- sumers can select. If fuel economy is regulated, the range of consumer choice may be sharply diminished, and some people will be harmed. This debate has not been resolved within the committee, nor within the community of policy analysts. However, as is clear from Tables 4-2 and 4-3 in Chapter 4, the difference between the cost-efficient fuel economies of vehicles in the 14-year case and the 3-year case is large. Thus, if most con- sumers had preferences corresponding to the 3-year case, yet fuel economy standards were set to correspond to the 14- year case, most consumers would be made economically worse off by such governmental regulation. Conversely, if most consumers had preferences corresponding to the 14- year case, yet automobile manufacturers offered only ve-

POTENTIAL MODIFICATIONS OF AND ALTERNATIVES TO CAFE 85 hicles corresponding to the 3-year case, most consumers would be made economically worse off by the government’s failure to intervene to increase fuel economy. Other Issues Not Considered The committee has explicitly not relied on rationales other than environmental and international oil market impacts. It is sometimes asserted that increases in fuel economy would reduce tailpipe emissions of criteria pollutants such as NOx or volatile organic compounds. But these criteria pollutants are regulated independent of fuel economy, so that each ve- hicle, when new, must have emissions below a federally mandated (or a tighter state-mandated) number of grams per mile. Since allowable grams per mile can be expected to remain the binding constraint and since allowable grams per mile does not depend on fuel economy, criteria pollutants do not provide a rationale for intervening to increase fuel economy. However, some research suggests that once the control systems of vehicles deteriorate, there is a relation- ship between fuel economy and emissions. And some par- ticularly high-fuel-economy vehicles now operate well be- low the grams-per-mile standards, although this is less likely to occur with higher fuel economy standards and/or with tightened standards for criteria pollutant emissions. The trend toward increasing the required warranty times for pol- lution control systems is likely to render this phenomenon moot for the purpose of evaluating fuel economy standards. Second, some critics believe manufacturers overestimate costs, that it would cost little to improve fuel economy, and that there could even be manufacturing cost savings associ- ated with such improvements. The cost estimates presented by the committee in Chapter 3 are, as might be expected, lower than some from industry. Overall, however, the com- mittee concludes that improving fuel economy significantly will raise the price of vehicles significantly. Third, some critics believe that consumers who care more about performance characteristics, such as acceleration, than about fuel economy are irrational. But this represents a dif- ference in tastes—the consumers and the critics value differ- ent things—not a difference in rationality. Finally, other industry observers have estimated external costs of driving in addition to the costs identified above, such as the costs of road congestion and policing. However, such costs will be unaffected by fuel economy and therefore do not provide a reason for market intervention to improve fuel economy. Quantifying Environmental and International Oil Market Costs One product of the combustion of hydrocarbon fuels such as gasoline or diesel in internal combustion engines is CO2. Scientific discussions of greenhouse gases typically refer to CO2 emissions in terms of the weight of the carbon (C) con- tained in the CO2, and it is this terminology that the commit- tee uses in this report. The combustion of each gallon of gasoline releases 8.9 kg of CO2, or 2.42 kg of C in the form of CO2. 3 The environmental and economic consequences of these releases are not included in the price of gasoline and are part of the environmental externalities of gasoline use. To quantify the environmental externalities associated with such CO2 releases, in principle one could directly mea- sure the various consequences of additional CO2 releases and place monetary values on each consequence. Although esti- mates have been made of the costs to agriculture, forestry, and other economic activities, estimating the marginal costs of environmental degradation, species extinction, increased intensity of tropical storms, and other impacts beyond com- mercial activities has proven highly controversial. A wide range of estimates appears in the literature, from negative values to values well over $100 per metric ton (tonne) of C. Public debate suggests that many people would estimate val- ues even outside of the published range of estimates, par- ticularly because there are many possible, although highly unlikely, events that could be very harmful or very benefi- cial. The committee has used a figure of $50/tonne C as an estimate of the environmental externality of additional car- bon emissions, although this figure is significantly higher than typical estimates in the published literature.4 This esti- mate translates into a cost of $0.12/gal (gasoline), the value used in the examples in this chapter. A range of cost esti- mates from $3/tonne to $100/tonne would give a range of estimated external costs from $0.007/gal to $0.24/gal of gasoline. A second environmental cost of gasoline use is related to the hydrocarbon and toxic chemical releases from the gaso- line supply chain, including oil exploration and recovery, oil refining, and distribution (tanker, pipeline, or tanker truck distribution, and gasoline retail sales). The more gasoline used, the greater will be the amount of hydrocarbons and toxics released. However, the supply chain is tightly regu- lated, and releases per gallon of gasoline now are very lim- ited. Marginal costs of this environmental impact are small. 3Combustion of gasoline releases about 19.36 kg of C per million Btu (MMBtu) of gasoline. A barrel of gasoline has an energy content of about 5.25 MMBtu; there are 42 gallons per barrel. 4At one of the committee’s public meetings, a representative of an envi- ronmental advocacy organization indicated that there was much uncertainty but offered a figure of $50/tonne. That figure is viewed as high, but not implausibly high, by committee members who have been involved in the global climate change debates. 5This issue was examined in Delucchi et al. (1994). That study showed various estimates for the value of hydrocarbon reductions from fuel economy improvements, depending on assumptions about the value of re- ducing hydrocarbon emissions, upstream control effectiveness, possible benefits of refueling, and evaporative losses. Based on this work, $0.02/gal seemed a reasonable estimate, given current emissions control trends and using $1,000/ton (1990 dollars) marginal damage of hydrocarbon emissions.

86 EFFECTIVENESS AND IMPACT OF CORPORATE AVERAGE FUEL ECONOMY (CAFE) STANDARDS The committee used an estimate of a $0.02/gal for the total of these external costs in its calculations.5 The Organization of Petroleum Exporting Countries (OPEC) operates as a cartel that restricts the supply of oil to escalate its price above the free-market level. The greater the consumption of oil, the higher will be its price. Since the higher price would apply to all oil imports, not just to the increased consumption, the financial cost to the United States of increased oil use exceeds the market payment for the increased amount. The additional financial cost of im- porting more oil, often referred to as the monopsony compo- nent of the oil import premium, was much studied after the energy crises of 1973–1974 and 1979–1980. The volatility of oil prices also creates problems. Past oil shocks may have caused significant macroeconomic losses to the U.S. economy. As U.S. oil consumption rises, so does the vulnerability to such disruptions.6 The additional vulner- ability cost, often referred to as the security component of the oil import premium, was also much studied in the wake of past energy crises. Observations of the world oil market since the 1980s sug- gest that the monopsony component of the oil import pre- mium is small, primarily because the impact of U.S. oil con- sumption on world prices has proven to be smaller than once thought. These observations also suggest that the security premium also has become smaller, because the U.S. econ- omy is now less vulnerable to oil price volatility than in the 1970s and the early 1980s: the United States has become significantly more energy efficient, its expenditure on oil relative to the gross domestic product (GDP) has declined, there now are more mechanisms for cushioning oil shocks, and the nation’s ability to manage the overall economy has improved greatly. Therefore, the marginal cost to the United States of oil consumption is now considerably smaller than previously estimated and is likely to remain so. However, this issue may become more important if the world price of oil rises. In addition, some analysts would argue that the concentration of oil use for transportation might be relevant, since there are few substitutes for oil in this sector. For its examples, the committee used an estimated exter- nal marginal cost of oil consumption of $5.00/bbl of oil for the combined monopsony component and security compo- nent of the oil import premium, although the cost could be smaller or larger than this figure.7 This estimate translates into a cost of $0.12/gal of gasoline. An oil import premium range from $1/bbl to $10/bbl would give a range of esti- mated external costs from $0.02/gal to $0.24/gal of gasoline. It should be emphasized that the monopsony component of the oil import premium is the marginal cost of increasing oil use. It includes neither the entire benefit to the United States of “solving” the problem of noncompetitive pricing by the OPEC nations nor the entire benefit of increasing in- ternational stability in world oil markets (or, equivalently, the cost of not solving these problems). These problems can- not be solved completely by changing the amount of oil con- sumed in the United States. Combining the $0.12/gal marginal cost estimate for CO2 externalities, the $0.12/gal for international oil external mar- ginal costs, and the $0.02/gal figure for externalities in the gasoline supply chain, the committee uses a total external marginal cost of additional gasoline use of $0.26/gal in all of the examples, although estimates as high as $0.50/gal or as low as $0.05/gal are not implausible and estimates well out- side of that range cannot be rejected out of hand. ALTERNATIVE POLICIES—SUMMARY DESCRIPTION Presentations to the committee, a review of published lit- erature, and committee deliberations identified many pos- sible modifications to the current CAFE system as well as other approaches to improving fuel economy. The various approaches are generally not mutually exclusive but nor- mally can be used alone or in combination with others. These changes can generally be grouped into four broad classes: • Retain the basic CAFE structure. This approach would keep CAFE basically intact but would modify some elements that are particularly troublesome. • Restructure fuel economy regulations. This approach would restructure CAFE with alternative regulatory or incentive policies directed at the fuel economy of new vehicles. • Adopt energy demand-reduction policies. This broader approach is designed to reduce either gasoline con- sumption or consumption of all fossil fuels. • Pursue cooperative government/industry technology strategies. This approach would attempt to advance au- tomotive technologies to greatly improve fuel economy. Retain the Basic CAFE Structure This class of policies would keep CAFE basically intact but would modify some troublesome elements, particularly those involving domestic versus import production and the defini- tions used to classify vehicles as trucks or passenger cars. 6Some people believe that military expenditures will also increase as a function of gasoline price (above and beyond mere fuel costs). However, the committee has seen no evidence to support that belief. 7Work by Leiby et al. (1997) provides estimates of $3.00/bbl. The En- ergy Modeling Forum (1982) study World Oil, using nine different math- ematical models, estimated that in 2000, oil price would be increased by between 0.8 percent and 2.9 percent for every million barrels per day of oil import reduction. Applying that same percentage to current prices would give a monopsony component between $2.20 and $8.20/bbl. The vulner- ability component was much smaller in that study.

POTENTIAL MODIFICATIONS OF AND ALTERNATIVES TO CAFE 87 The Two-Fleet Rule Differentiating Between Domestic and Imported Cars Currently, each manufacturer must meet the CAFE stan- dard separately for its domestically produced fleet of new passenger cars and for its imported fleet. Averaging8 is al- lowed within each of these two fleets but is not allowed across the two fleets. A domestically produced fleet that sig- nificantly exceeded the CAFE standard could not be used to compensate for an imported fleet that failed to meet the stan- dard and vice versa. This requirement that the domestically produced fleet and the imported fleet each separately meet the CAFE standard is referred to as the two-fleet rule. No such requirement exists for light trucks. This distinction could be removed and the fuel economy standard could ap- ply to the entire new car fleet of each manufacturer. Classification of Vehicles As Trucks vs. Passenger Cars The distinction between passenger cars and light-duty trucks could be redefined to correspond more closely to the original distinction between the two classes—passenger ve- hicles and work/cargo vehicles. Alternatively, the incentives for manufacturers to classify vehicles as trucks could be re- duced or eliminated. The provisions for flexible-fuel vehicles could be eliminated or redefined to ensure they will apply only to vehicles that will often use alternative fuels. Restructure Fuel Economy Regulations These policies more fundamentally restructure CAFE with alternative regulatory or incentive policies directed at fuel economy of new vehicles. Tradable Fuel Economy Credits There would be an increase in the economic efficiency and the flexibility of the CAFE system if a market-based system of tradable fuel economy credits were created, under which automobile manufacturers could sell fuel economy credits to other manufacturers and could buy credits from other manufacturers or from the government. This system would be similar in many respects to the successful system now used for trading sulfur emission credits among electric- ity power plants. Feebates Feebates is an incentive mechanism that uses explicit gov- ernment-defined fees and rebates. Vehicles with a fuel economy lower than some fuel economy target pay a tax, while vehicles with a fuel economy higher than the target receive a rebate. Such systems could be designed to be rev- enue neutral: the tax revenues and the subsidies would just balance one another if the forecasted sales-weighted average fuel economy (or average per-mile fuel consumption) turned out as predicted. Attribute-Based Fuel Economy Targets The government could change the way that fuel economy targets9 for individual vehicles are assigned. The current CAFE system sets one target for all passenger cars (27.5 mpg) and one target for all light-duty trucks (20.7 mpg). Each manu- facturer must meet a sales-weighted average (more precisely, a harmonic mean—see footnote 8) of these targets. However, targets could vary among passenger cars and among trucks, based on some attribute of these vehicles—such as weight, size, or load-carrying capacity. In that case a particular manu- facturer’s average target for passenger cars or for trucks would depend upon the fractions of vehicles it sold with particular levels of these attributes. For example, if weight were the cri- terion, a manufacturer that sells mostly light vehicles would have to achieve higher average fuel economy than would a manufacturer that sells mostly heavy vehicles.10 Uniform Percentage Increases There have been proposals that would require each manu- facturer to improve its own CAFE average by some uniform percentage, rather than applying the targets uniformly to all manufacturers. This is often referred to as the uniform per- centage increase (UPI) standard. Adopt Energy Demand-Reduction Policies There are several alternatives aimed more broadly at re- ducing motor fuel consumption or all fossil fuel consump- tion. They could be part of a more comprehensive energy policy. If these more broadly based alternatives were imple- mented, they could be used in place of or along with the instruments aimed directly at new vehicle fuel economy. 8Under CAFE, the “average” fuel economy is the sales-weighted har- monic mean of fuel economies of the individual vehicles sold by the manu- facturer. Mathematically, a standard on the sales-weighted harmonic mean of fuel economies of individual vehicles is exactly equivalent to a standard on the sales-weighted average of per-mile fuel consumption of individ- ual vehicles. In this chapter, the word “average” or “averaging” is used to denote this harmonic mean of fuel economies or average of per-mile consumption. 9Throughout this chapter, the word “target” is applied to the goal for fuel economy, or per-mile fuel consumption, of individual vehicles or groups of vehicles. The word “standard” is used to denote a regulatory rule that must be met. Under the current CAFE system, regulations do not require that each car or truck meet any particular target, although CAFE requires the corporation to meet the standards for the aggregate of all passenger cars and the aggregate of all trucks. 10Targets could also be normalized, for example, by expressing them in terms of weight-specific fuel consumption—for example, gallons used per ton of vehicle weight per 100 miles.

88 EFFECTIVENESS AND IMPACT OF CORPORATE AVERAGE FUEL ECONOMY (CAFE) STANDARDS Gasoline Tax The current federal excise tax on gasoline could be in- creased. A tax increase would provide direct incentives for consumers to buy and for manufacturers to produce higher- fuel-economy new vehicles and would also provide incen- tives to reduce the use of all new and existing vehicles. Carbon Taxes/Carbon Cap-and-Trade Systems In order to address problems of global greenhouse gas re- lease, the United States could impose a carbon tax or could adopt a carbon cap-and-trade system. Under these systems, the total annual emissions of carbon dioxide would be limited or capped, rights to emit carbon would be allocated or auc- tioned off, and these rights would then be tradable among firms. In either system, the price of energy would increase, on a fuel-by-fuel basis, roughly in proportion to the amount of CO2 released from combustion of that fuel. Such plans would provide a broad-based incentive to use less of all fossil fuels, especially those that are particularly carbon-intensive. Pursue Cooperative Government/Industry Technology Strategies The final class of strategies would attempt to create dra- matic changes in automotive technologies, changes that could greatly alter the economy of fuel consumption or the types of fuels used. Partnership for a New Generation of Vehicles (PNGV) A particular, ongoing example of such strategies is the Partnership for a New Generation of Vehicles (PNGV) pro- gram, a private-public research partnership that conducts precompetitive research on new vehicle technologies. One of its goals is to create marketable passenger cars with fuel economy up to 80 mpg. Technology Incentives The government could provide tax or other incentives to manufacturers or consumers for vehicles that embody new high-efficiency technology. Such incentives would encour- age manufacturers to pursue advanced technology research and to bring those new technologies to market and would encourage consumers to purchase vehicles that use them. MORE COMPLETE DESCRIPTIONS OF THE ALTERNATIVES Retain the Basic CAFE Structure Classification of Vehicles When CAFE regulations were originally formulated, dif- ferent standards were set for passenger vehicles and for work/cargo vehicles. Work/cargo vehicles (light-duty trucks that weigh less than 8,500 lb gross vehicle weight) were al- lowed higher fuel consumption because they needed extra power, different gearing, and less aerodynamic body con- figurations to carry out their utilitarian, load-carrying func- tions. At that point, light-truck sales accounted for about 20 percent of the new vehicle market. However, as one observer noted, “the 1970s working definition distinction between a car for personal use and a truck for work use/cargo transport, has broken down, initially with minivans, and more recently with sport utility vehicles and other ‘cross-over’ vehicles that may be designed for peak use but which are actually used almost exclusively for personal transport.”11 The car/truck distinction bears critically on fuel economy considerations. Trucks are allowed to meet a lower CAFE standard, 20.7 mpg versus 27.5 mpg for cars, and their mar- ket share has increased enormously. Vehicles classed as light-duty trucks now account for about half the total new vehicle market. The car/truck distinction has been stretched well beyond its original purpose. For example, the PT Cruiser, a small SUV that can carry only four passengers and cannot tow a trailer, is considered a truck, while a large se- dan that can carry six passengers while towing a trailer is considered a car. Two kinds of change might alleviate these problems: • Redefine the criteria determining whether a vehicle is classified as a car or a truck or • Sharply reduce the economic incentives for manufac- turers to classify their vehicles as trucks. Fuel economy regulators might tighten the definition of a truck. The Environmental Protection Agency (EPA) has already done so: for example, it classifies the PT Cruiser as a car for purposes of the emissions standards. EPA and the National Highway Traffic Safety Administration (NHTSA) have considerable regulatory discretion to implement such changes after a rulemaking process. The economic incentives for manufacturers to classify their vehicles as trucks come from both the CAFE standards and the gas guzzler tax. Because CAFE standards require much greater fuel economy for cars than for trucks and be- cause they impose a binding constraint on manufacturers, CAFE standards create a strong incentive to make design changes in vehicles that allow them to be classified as trucks, whenever such changes are possible at modest cost. (For example, the PT Cruiser was designed with removable rear seats, which allows it to be classified as a truck.) Reductions in the differential between CAFE standards for trucks and standards for cars would therefore reduce these economic incentives. At one extreme, eliminating the differ- 11J. Alston, EPA, Letter to the committee dated April 16, 2001.

POTENTIAL MODIFICATIONS OF AND ALTERNATIVES TO CAFE 89 ential between truck and car CAFE standards would elimi- nate all CAFE-derived incentives for manufacturers to have vehicles classified as trucks.12 This is the approach EPA will use for emission standards: beginning in 2009, it will treat light-duty trucks the same as it treats cars under the Tier 2 emission regulations.13 In addition, the gas guzzler tax, discussed in Chapter 2, im- poses a large tax on passenger cars but not on trucks. For ex- ample, a passenger car with a fuel economy of 20 mpg would face a gas guzzler tax of $1,000, while a truck with similar fuel economy would face no tax. At 12 mpg, the tax on a car would be $7,000. This tax provides a large financial incentive to de- sign any vehicle expected to have low fuel economy in such a way as to assure that it will be classified as a truck, so that reducing or eliminating the tax would likewise reduce the in- centive to ensure that vehicles are classified as trucks. The distinction between cars and trucks extends to the processes for determining fuel economy standards. CAFE standards for passenger cars are set legislatively, with a long time horizon. CAFE standards for trucks are set by rule- making within NHTSA, with shorter time horizons. Integra- tion of the processes, so that trucks and passenger cars are subject to equivalent processes and equivalent time horizons for regulatory decision making, might lead to more consis- tency of treatment among vehicle types. A related problem involves flexible-fuel vehicles. In cal- culating new car fuel economy for CAFE compliance pur- poses, a flexible-fuel vehicle is currently deemed to have a fuel economy 1.74 times as high as its actual fuel economy, with a 1.2 mpg maximum total increase per manufacturer from this flexible-fuel vehicle adjustment. This adjustment is based on a legislative assumption that 50 percent of the fuel such vehicles use would, on average, be E85, including only 15 percent petroleum.14 However, few of these vehicles ever use any fuel other than gasoline. Estimates from the Energy Information Ad- ministration (EIA) suggest15 that for 1999, there were 725,000 vehicles capable of using E85, but only 3.1 percent of them were using any E85 at all. Total E85 consumption in 1999 was 2 million gallons, or only 92 gallons for each of the 3.1 percent of the vehicles using some E85 (EIA, 2001). Therefore, it is likely that even these vehicles were using E85 for less than 25 percent of their fuel requirements. In total, less than 1 percent of the fuel used in these vehicles seems to be E85, and more than 99 percent seems to be gaso- line. Thus, the current incentives to produce such vehicles lead to increased costs and lower fleet fuel economy without corresponding benefits. Distinction Between Domestic and Imported Passenger Cars Current CAFE regulations make a distinction between domestic and imported passenger cars. The distinction is based on the proportion of the car that is manufactured in the United States; an import is defined as a car with less than 75 percent domestically produced content. Imports and domes- tically produced vehicles constitute two separate car fleets under CAFE regulations. Under these regulations the do- mestic fleet and the import fleet of passenger cars must sepa- rately meet the same 27.5 mpg standard. There is no such rule for trucks. The two-fleet rule was added to the original CAFE legis- lation to protect domestic employment. There was a concern that U.S. manufacturers might decide to import their small cars rather than continue to make them in the United States, with a resultant loss of jobs. At one time the two-fleet rule made it impossible for the Big Three to increase imports to help meet their domestic fleet CAFE obligations. Over time, however, foreign manufacturers have moved production to the United States. Now the two-fleet rule can just as well provide incentives for manufacturers to reduce domestic production and increase import production to help meet CAFE obligations. In addition, under the North Ameri- can Free Trade Agreement (NAFTA), manufacturers can count vehicles produced in Canada as part of their domestic fleet and will soon be able to count those produced in Mex- ico. Thus it appears to the committee that the two-fleet rule no longer serves to protect U.S. employment. Presentations to the committee during its open meetings noted that the rule produces perverse results and increases costs. For example, representatives of American Honda Motors testified that Honda ships Accords from Japan to as- sure that Honda Accords are classified as an import under CAFE. Reporting on the testimony on March 19, 2001, in Automotive News, Stoffer (2001) wrote as follows: “Last year, by selling about 87,000 Japan-built Accords, to go with the 317,000 built in Ohio, Honda was able to keep the average domestic content of the whole model line below the cutoff point of 75 percent,” Honda spokesman Art Gar- ner said. That means the Accord remains classified an im- port model line and helps Honda keep the average fuel economy of its imported fleet comfortably above the 27.5 mpg standard. Without the more efficient Accord, the im- ported fleet would consist entirely of the more performance- oriented Acuras and the Honda Prelude. A second example was discussed in the same Automotive News story: “In the early 1990s, Ford Motor Co. intention- ally reduced domestic content of the Crown Victoria and Grand Marquis because of CAFE.” 12This statement is not a recommendation that the standards for trucks be made identical to those for passenger cars, rather a simple observation about the incentives that are created by the differential nature of these standards. 13The bin structure of Tier 2 emissions standards provides more flexibil- ity for manufacturers to meet the standards, but the requirements are identi- cal for passenger cars and light-duty trucks. 14E85 consists of 85 percent ethanol and 15 percent gasoline. 15The data appear online at <http://www.eia.doe.gov/cneaf/alternate/ page/datatables/atf1-13_00.html>.

90 EFFECTIVENESS AND IMPACT OF CORPORATE AVERAGE FUEL ECONOMY (CAFE) STANDARDS The United Auto Workers (UAW) union stated during the committee’s open meetings that it supports continuation of the two-fleet rule and indicated it believes the rule contin- ues to protect American jobs. In response to UAW commu- nications, the committee sought to identify research or analy- sis that tends to validate that position but could not find any. It is possible that the rule provides some protection for exist- ing jobs, but it appears likely that removing it would have little industrywide impact. Since the two-fleet rule increases costs to consumers, the committee believes it is no longer justifiable and should be eliminated. Restructure Fuel Economy Regulations Tradable Fuel Economy Credits The existing CAFE system already allows a manufacturer to accumulate CAFE credits if its fleet mix exceeds the stan- dard. These credits may be carried forward and used to off- set future CAFE deficits by the same manufacturer. The idea of tradable fuel economy credits (tradable credits, for short) carries this flexibility one step further: Manufacturers could also be allowed to sell and buy credits among themselves or to buy credits from the government. Under this system, fuel economy targets would be set, either uniform targets as in the current CAFE system, or at- tribute-based targets, as discussed below. Each manufacturer would be required either to meet these targets or to acquire sufficient credits to make up the deficit. The credits could be purchased from the government or from other automobile manufacturers. A manufacturer whose new vehicle fleet had greater fuel economy than the overall target would acquire credits. These credits could be saved for anticipated later deficits or could be sold to other manufacturers. Credits would be equal to the difference between the projected gaso- line use over the life of the vehicle, using a legislatively deemed total lifetime vehicle miles, and the projected life- time gasoline use of vehicles just meeting the target. As an example, assume that a uniform target for cars of 30 mpg is legislated and the vehicle lifetime is deemed to be 150,000 miles. This implies a lifetime target fuel consump- tion of 5,000 gal/vehicle (150,000 miles; 30 mpg). A manufacturer that sold 1 million cars with an average ex- pected lifetime fuel consumption of 4,500 gal each (33.33 mpg) would acquire 500 credits per car for each of 1 million cars, or 500 million credits. A manufacturer that sold 1 mil- lion cars with an average expected lifetime fuel consump- tion of 5,500 gal each (27.27 mpg) would need to purchase 500 million credits. The government would assure that prices for tradable credits would not exceed some ceiling price by offering to sell credits to any manufacturer at some predetermined offer price. The offer price could be set equal to the estimated external costs per gallon of gasoline use.16 If external costs (e.g., greenhouse gas emissions and international oil market) are estimated to be $0.26/gal, the government would offer to sell credits at a price of $0.26 per 1-gallon credit. The availability of credits from the government is impor- tant because it represents a safety valve preventing exces- sive costs to manufacturers (and consumers) in the event that unforeseen market changes or errors in setting targets make attaining the target more costly than originally projected.17 The market-clearing price of tradable credits would never exceed the government offer price, because a buyer of cred- its could always turn to the government if the price of credits were above the government offer price. Suppose the marginal cost of reducing gasoline use enough to meet the target was greater than the sum of the gasoline price plus the market price of credits: Manufactur- ers could buy credits without being forced to install overly expensive technology or to make changes to vehicle at- tributes that could damage sales. Conversely, if the marginal cost of reducing gasoline use to meet the fuel economy tar- get was less than the sum of gasoline price plus market price of credits, the manufacturer would choose to make changes necessary to meet or to exceed the fuel economy target. Since the decisions would be made by and the resulting financial costs borne by the manufacturer, the manufacturer would have a motivation to correctly estimate the costs of fuel economy increases. Under this system, the manufacturer could respond to automotive market conditions but would still have an enhanced incentive to increase fuel economy. In comparison with the current CAFE system, a tradable credits system would increase the range of options available to manufacturers. Currently, manufacturers have two op- tions: They can meet the standards or they can pay the gov- ernment a civil penalty.18 Under a system of tradable fuel economy credits, manufacturers would have more options: They could meet the targets, they could pay the government for credits, or they could purchase credits from other manu- facturers. They would be free to choose. Similarly, in comparison with the current CAFE system, a tradable credits system would increase the range of attrac- tive options available to manufacturers whose fuel economy exceeds the target. Under CAFE, such manufacturers have no incentives to further increase fuel economy. But under a tradable fuel economy credits system, they would have the option of further increasing fuel economy and receiving ad- ditional credits that they could sell to other manufacturers. 16Applying this rule and using a reasonably accurate estimate of the life- time miles of vehicles is economically efficient as long as the external cost per gallon of gasoline use during the future vehicle life has the same value for current decision making as the estimate of external costs used for the regulation. 17In addition, the safety valve limits the exercise of market power in the market for tradable fuel economy credits. Such market power could other- wise become an important problem if only a very small number of manufac- turers were selling tradable credits. 18The penalty is currently $5.50 for every 0.1 mpg by which the manu- facturer misses the standard.

POTENTIAL MODIFICATIONS OF AND ALTERNATIVES TO CAFE 91 The tradable credits system would have another advan- tage, especially if the sales price of tradable credits were made public.19 Debates about environmental standards usu- ally involve disputes about implementation cost: Those fa- voring regulation contend that the standards will be cheap to implement, while manufacturers contend that the standards will be too expensive. The sales price of credits will reflect the marginal costs of fuel economy improvements, since manufacturers can be expected to increase fuel economy to the point at which the marginal cost of fuel savings equals the sum of gasoline price plus the market price of tradable credits. It should be noted that a similar tradable credits scheme has been used for some time in the electrical power industry to reduce sulfur emissions.20 There is general agreement that tradable credits have been highly successful: They have re- duced the economic cost of compliance, and they have reached the achieved environmental goals. Feebates Like tradable fuel economy credits, a feebate system21 is an incentive mechanism that can be used with almost any method of specifying fuel economy targets. Under a feebate system, target fuel economies would be set, either uniform targets or attribute-based targets. Fees would be imposed on new vehicles with mpg’s lower than the target; the lower the mpg, the greater the fee. Rebates would be provided to manu- facturers of new vehicles with mpg’s above the target. These fees and rebates would be aggregated across all vehicles sold by a single manufacturer, which would receive, or make, a single payment. Feebates could be designed to be revenue neutral, with fee revenues and rebates balancing each other. However, actual revenue neutrality would depend on the accuracy of sales forecasts. If it proves inexpensive to increase fuel economy and manufacturers greatly exceed the targets, there would be a net payment from the government to the auto- motive industry. Conversely, if it is very costly to increase fuel economy and manufacturers fall well short of the tar- gets, there would be a net payment from the industry to the government. Like tradable credits, feebates would provide direct in- centives for all manufacturers to increase the fuel economy of their vehicles.22 However, unlike the tradable credits sys- tem, particular fees and rebates would always be determined by legislation and would not be influenced by market conditions. Attribute-Based Targets Figure 5-1 illustrates the principles behind the current CAFE standards. Each dot represents a specific passenger car model—for example, the four-cylinder Accord and the six-cylinder Accord are separate dots. Only those car models that sold at least 1,000 vehicles per year in the United States are shown. The vertical axis shows fuel consumption mea- sured not in miles per gallon but in the amount of gasoline each car needs to travel 100 miles—for example, 25 mpg implies 4.0 gallons to drive 100 miles. The dark horizontal line shows the current CAFE standards: It is placed at 27.5 mpg, which is 3.64 gallons per 100 miles on the vertical axis. The horizontal axis shows the weight of the car. Cars on the right-hand end weigh more and consume more fuel: They are above the dark CAFE line, which means they are con- suming more fuel than the average allowed by the standards. To get back into compliance, a manufacturer that sells heavy cars must also sell some light cars—the left-hand end of the graph. Those cars consume less fuel than the standard. If the manufacturer sells enough light cars, it can produce a fleet average that complies with the CAFE standard of 27.5 mpg. When the CAFE standards were first implemented, the average car was considerably heavier than today’s cars. To comply with the new standard, manufacturers had a consid- erable incentive to downweight their very largest cars, to produce more small cars, and to encourage consumers to buy those small cars. Thus one effect of CAFE was to reduce the average weight of cars, and this had an undesirable side effect on safety. There were also equity effects across manufacturers. Some manufacturers had a product mix that emphasized small and medium-weight cars—these manufacturers found it cheap and easy to meet the CAFE standards. Other manu- facturers were producing a mix that was more toward the right-hand end of the curve—those manufacturers had to spend a considerable amount of money to develop and sell lighter cars so they could create enough CAFE credits to bring them into compliance with the standards. These problems arise because the CAFE standards hold all cars to the same fuel economy target regardless of their weight, size, or load-carrying capacity. This suggests that19Regulations could be established that require the prices of credits to be made public. Absent such regulations, manufacturers might include confi- dentiality provisions in their agreements to buy or sell credits. 20The tradable fuel economy credits system would differ from tradable credits for sulfur emissions in that the proposed system includes a safety valve, whereas the sulfur emissions system does not. One reason for the difference is that exercise of market power for sulfur emissions credits is much less likely than would be the case for tradable fuel economy credits. 21For a more comprehensive study on feebates, see Davis et al. (1995). 22There is a duality between feebates and tradable credits; if the targets are identical, the feebate system is linear with respect to deviations in fuel consumption rates from the targets, and the average fuel consumption rates are above the average target. In this case, feebates and tradable credits have identical incentives. They differ from each other if the average fuel con- sumption rates are below the average targets.

92 EFFECTIVENESS AND IMPACT OF CORPORATE AVERAGE FUEL ECONOMY (CAFE) STANDARDS consideration should be given to developing a new system of fuel economy targets that responds to differences in vehicle attributes. For example, the standards might be based on some vehicle attribute such as weight, size, or load. If such attribute-based targets were adopted, a manufacturer would still be allowed to average across all its new vehicle sales. But each manufacturer would have a different target, one that depended upon the average size of the criterion attribute, given the mix of vehicles it sold.23 A tradable fuel economy credits system, as described above, could be implemented in combination with the attribute-based targets. The choice of method for setting ve- hicle economy targets could be separate from choice of in- centives to meet targets. In the current fleet, size, weight, and load-carrying capac- ity are highly correlated: large cars tend to be heavier, to have room for more people, and to have more trunk capacity than small cars. Choice of a particular attribute as the basis for CAFE measurement will result in incentives for engi- neers to design vehicles with new combinations of the at- tributes and to respond to incentives by further varying that particular attribute. An attribute-based system might use vehicle weight as the criterion. The dashed, upward sloping line in Figure 5-1 shows the average relationship between vehicle weight and fuel consumption. A weight-based CAFE system would use that upward sloping line as its target rather than the current horizontal line.24 While a weight-based CAFE system has a number of at- tractive features, it also has one major disadvantage: It re- moves incentives to reduce vehicle weight. Judging by re- cent weight and profit trends, it seems likely the result would be an increase in the proportion of very large vehicles, which could cause safety problems as the variance in weight among vehicles increased. It could also cause an increase in fleet- wide fuel consumption. (These issues are discussed at more length in Attachment 5A.) Figure 5-2 illustrates an alternative that combines most of the desirable features of the current CAFE standards and the weight-based standard. The target for vehicles lighter than a particular weight (here, 3,500 lb) would be proportional to their weight (e.g., the dashed line in Figure 5-1). But to safe- guard against weight increases in heavier vehicles, the target line turns horizontal. Cars heavier than this weight would be required to meet a target that is independent of their weight. (The details of positioning the lines are discussed in Attach- ment 5A). These targets provide a strong incentive for manufactur- ers to decrease the weight of heavier cars—and even a small incentive to increase the weight of the lightest cars. The safety data suggest that the combined effect would be to en- hance traffic safety. Accordingly, the committee has named it the Enhanced-CAFE standard (E-CAFE). The Enhanced- CAFE standard may be calibrated separately for cars and for trucks, or it is possible to create a single standard that applies to both types of vehicles, thereby removing the kinds of manipulation possible under the current dual classification system. The committee views the Enhanced-CAFE system as a serious alternative to the current CAFE system. It holds real promise for alleviating many of the problems with the cur- rent regulations. Attachment 5A presents a full description and analysis. Uniform Percentage Increases Another possible change would be to require each manu- facturer to improve its own CAFE average by some target, say 10 percent; this is often referred to as the uniform per- centage increase (UPI) standard. Thus, a manufacturer that was now right at the 27.5 mpg CAFE standard would have to improve its performance to 30.25 mpg. A manufacturer that 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 1,500 2,500 3,500 4,500 5,500 Curb Weight G al lo ns U se d pe r 10 0 M ile s CAFE standard 27.5 mpg = 3.64 gals/100 miles (20 mpg) (17 mpg) (33 mpg) (50 mpg) FIGURE 5-1 The operation of the current CAFE standards: pas- sengers cars, gasoline engines only, 1999. 23The manufacturer could average actual gallons per mile and compare that average with the average of target gallons per mile. Alternatively, the manufacturer could average deviations, plus and minus, between actual gallons per mile and target gallons per mile. Whether averaging is done first and deviation calculated second or deviations are calculated first and aver- aging is done second is mathematically irrelevant. 24These possible weight-based targets do not begin to exhaust the possi- bilities. Many alternative weight-based targets could be designed, or the targets could be based on load-carrying capacity, interior volume, exterior volume, other utility-related attributes, or a combination of these variables (e.g., weight and cargo capacity). The committee did not try to identify and analyze all such possibilities—that would have been well beyond the scope of this study.

POTENTIAL MODIFICATIONS OF AND ALTERNATIVES TO CAFE 93 was exceeding the current standard at, say, 33 mpg would have to improve its performance to 36.3 mpg. The UPI system would impose higher burdens on those manufacturers who had already done the most to help reduce energy consumption. The peer-reviewed literature on envi- ronmental economics has consistently opposed this form of regulation: It is generally the most costly way to meet an environmental standard; it locks manufacturers into their relative positions, thus inhibiting competition; it rewards those who have been slow to comply with regulations; it punishes those who have done the most to help the environ- ment; and it seems to convey a moral lesson that it is better to lag than to lead. In addition to fairness issues, the change would not elimi- nate the problems of the current CAFE system but would create new ones. Implementation of such rules provides strong incentives for manufacturers to not exceed regulatory standards for fear that improvements will lead to tighter regu- lations. Thus, such rules tend to create beliefs counterpro- ductive for longer-term goals. Adopt Energy Demand-Reduction Policies Several alternatives would be aimed more broadly at re- ducing total gasoline consumption or at reducing all fossil fuel consumption, not simply at reducing the per-mile gaso- line consumption of new vehicles. Alternatives include gaso- line taxes, carbon taxes, and a carbon cap-and-trade system. Either gasoline taxes or carbon trading/taxes might be part of a comprehensive national energy policy. If these more broadly based policies were implemented, policies aimed directly at fuel economy of new cars might be used along with the broadly based policies, or they could be used in place of one another. The committee did not devote much time to discussing carbon trading, carbon taxes, and fuel taxes. This does not imply that it considers these options to be ineffective or in- appropriate. In fact, such policies could have a much larger short-term and mid-term effect on fuel consumption and greenhouse gas emissions than any of the other policies dis- cussed in this report. The committee did not address these policies comprehensively because they were not part of its charge; instead, it presents here a basic, though incomplete, discussion of these options. Gasoline Taxes One alternative, addressed directly at gasoline use, would be an increase in the federal excise tax on gasoline from its current level of $0.184/gal.25 Every $0.10/gal increase in the gasoline tax would increase the price of gasoline by almost as much.26 Increasing the gasoline tax would encourage consumers to drive more efficient vehicles. This would indirectly pro- vide incentives to the manufacturers to increase the fuel effi- ciency of their vehicles. In addition, a gasoline tax would have an immediate broad impact on gasoline consumption: It would encourage consumers not only to buy more effi- cient new vehicles but also to drive all vehicles less. If the policy goal is to reduce gasoline consumption and the envi- ronmental and oil market impacts of gasoline consumption, then a gasoline tax increase would broadly respond to that goal. Gasoline taxes, however, have faced significant opposi- tion. Critics point out that gasoline taxes fall particularly hard on rural families and those in more remote locations, where long-distance driving is a normal part of life. It is often as- serted that gasoline taxes are regressive and impact the low- est income families the most, even though urban poor and wealthy people typically spend a smaller portion of their in- come on gasoline than do middle-class families. If the fed- eral gasoline tax were increased, Congress could make the tax revenue neutral or could take other measures to ensure that the change would not cause undue harm. Carbon Taxes/Carbon Cap-and-Trade Systems To address problems of global greenhouse gas release, the United States could (1) impose a carbon tax or (2) adopt a carbon cap-and-trade system, under which the total annual emissions of carbon dioxide would be capped or limited to some policy-determined level. In a system of carbon taxes, each fossil fuel would be FIGURE 5-2 Fuel economy targets under the Enhanced-CAFE sys- tem: cars with gasoline engines, 1999. 2.0 3.0 4.0 5.0 6.0 1500 2500 3500 4500 5500 Curb Weight G al s pe r 10 0 m ile s 25In addition, state excise taxes average $0.20/gal, according to the En- ergy Information Administration (2000). 26The price of gasoline would increase by slightly less than the increase in the gasoline tax because the imposition of the tax would reduce oil de- mand, which in turn would reduce crude oil price and would reduce the per- gallon earnings of refiners and marketers. However, the price and earnings reductions would be small.

94 EFFECTIVENESS AND IMPACT OF CORPORATE AVERAGE FUEL ECONOMY (CAFE) STANDARDS taxed, with the tax in proportion to the amount of carbon (or CO2) released in its combustion. Thus, coal would have the highest tax per unit of energy, petroleum and petroleum prod- ucts would have a lower tax, and natural gas would have the lowest tax. In a carbon cap-and-trade system, an energy-producing or -importing firm would be required to possess carbon credits equal to the amount of carbon emitted from its prod- ucts.27 A limit would be set on the overall U.S. emissions of CO2 or of greenhouse gases. Credits could be auctioned off to firms or could be given to firms in proportion to their historical sales of products, and these rights would be trad- able among firms. Either way, the price of energy would increase, with the increases on a fuel-by-fuel basis roughly in proportion to the amount of CO2 released from combus- tion of the fuels.28 A carbon tax, or carbon cap-and-trade system, would have all the advantages of a gasoline tax, and it would extend them to other sectors as well: The rise in power costs would encourage consumers to buy more efficient furnaces, air con- ditioners, and appliances and would encourage them to use their existing furnaces, air conditioners, and appliances more efficiently. Carbon taxes or trading would provide a broad- based incentive to use less of all fossil fuels, especially those that are particularly carbon-intensive, extending the principle of “least cost” to the entire economy. Implementing such a plan would prevent one energy-using sector from hav- ing marginal costs of carbon reduction grossly out of line with those of the other sectors and thus would be an eco- nomically efficient means of promoting the reduction of car- bon emissions. Pursue Cooperative Government/Industry Technology Strategies The final class of strategies would attempt to advance technologies that create dramatic changes in available auto- motive technologies, changes that could, for the indefinite future, greatly alter fuel economy or types of fuels used. The committee has not evaluated these technology incentive strategies (such an evaluation is beyond the scope of its charge), but it believes such strategies would best be viewed as complementary to the other policy directions, at least once major technological successes had been achieved: A suc- cessful technology strategy could greatly reduce the costs of increasing fuel economy. However, in the near term, aggres- sive requirements to increase fuel economy could divert R&D expenditures away from such technology development efforts and could lead to short-term modest increases in fuel economy at the expense of long-term, dramatic reductions in fuel use. Partnership for a New Generation of Vehicles A particular ongoing example of such strategies is the Partnership for a New Generation of Vehicles (PNGV) pro- gram. PNGV is a private-public research partnership that conducts precompetitive research directed at new automo- tive technology. One of its goals is to create marketable pas- senger cars with fuel economies up to 80 mpg. Each partici- pating company has developed a concept car that would approach the fuel economy goal but that, because of the re- maining technological challenges and high production costs, is unlikely to be marketable in the near term. This program has led to advances in technologies such as compression ig- nition engines, hybrid vehicles, batteries, fuel cells, light engine and vehicle body materials, and advanced drive trains. Technology Incentives During the committee’s open information-gathering meet- ings, representatives of the automobile manufacturers pro- posed that the government provide tax credits to consumers who purchase vehicles that embody new, high-efficiency technology, especially hybrid electric vehicles. Such rebates would strengthen incentives for manufacturers to pursue ad- vanced technology research and then bring those new tech- nologies to market and for consumers to buy these products. However, the committee has not evaluated the particular policy instruments; that evaluation is beyond the scope of this study. ANALYSIS OF ALTERNATIVES Dimensions for Assessing Alternatives In analyzing the alternatives to CAFE, it is important to assess several issues associated with each particular policy instrument as well as the trade-offs among these issues: • Fuel use responses encouraged by the policy, • Effectiveness in reducing fuel use, • Minimizing costs of fuel use reduction, • Other potential consequences —Distributional impacts —Safety —Consumer satisfaction —Mobility —Environment —Potential inequities, and • Administrative feasibility. These are discussed in varying depth below. 27See, for example, Kopp et al. (2000). 28Since all prices would further adjust in response to the changing de- mand conditions, the final equilibrium price increases would not be exactly in proportion to the amount of CO2 released but would depend on the vari- ous supply and demand elasticities.

POTENTIAL MODIFICATIONS OF AND ALTERNATIVES TO CAFE 95 Alternative Policies and Incentives for Fuel Use Responses The various policy measures encourage at least seven quite different fuel use responses. The measures differ in terms of which possible responses they will motivate: those related to the fuel economy of new vehicles, to the usage of vehicles, or to fleet turnover. Three fuel use responses are directly related to the fuel economy of new vehicles: • The number of cars sold in each weight class deter- mines the average weight of the new vehicle fleet. All else equal, lighter vehicles use less fuel. This adjust- ment is referred to as “weight.” • Engine power, other performance characteristics, and vehicle utility can vary. All else equal, greater perfor- mance implies more fuel consumed. This adjustment is referred to as “performance.” • More energy-efficient technologies can be incorpo- rated into the engine, drive train, tires, and body struc- ture, often at increased costs, to reduce fuel use with- out changing weight or performance. This adjustment is referred to as “technology.” Three fuel use responses are directly related to the usage of vehicles: • The total vehicle miles traveled (VMT) by the entire fleet of vehicles—new and old—is fundamental to fuel consumption. All else equal, fuel consumption is di- rectly proportional to total VMT. This adjustment is referred to as “VMT.” • Driving patterns can be altered—driving during con- gested times, relative usage frequency of various ve- hicles, speed. Such adjustment possibilities are re- ferred to as “existing vehicle use.” • Maintenance of existing vehicles can influence fuel use. All else equal, a well-maintained vehicle will use less fuel. This adjustment is referred to as “vehicle maintenance.” One possible fuel use response is directly related to ve- hicle fleet turnover: • The rate of retirement of older, generally less fuel- efficient vehicles influences the rate at which the over- all fuel economy of the vehicle fleet changes, espe- cially if new vehicles are more fuel-efficient than old vehicles. This adjustment is referred to as “vehicle retirements.” Effectiveness in Reducing Fuel Use The various policy instruments differ in which responses they motivate and therefore differ in terms of their effective- ness in reducing fuel use. Table 5-1 summarizes fuel use responses caused by various policy instruments. It provides a very rough indication of whether a particular policy instru- ment would lead to a larger adjustment than would other instruments. Each cell of the table indicates changes relative to what would be the case absent the particular policy being considered, with all else held equal. It should be noted that for these comparisons, the magnitude of responses associ- ated with any particular policy instrument depends on how aggressively that instrument is implemented—the severity TABLE 5-1 Incentives of the Various Policy Instruments for Seven Types of Fuel Use Response Fuel Use Adjustment Existing Vehicle Vehicle Policy Instrument Average Weight Performance Technology VMT Vehicle Use Maintenance Retirements Current CAFE, with or without two-fleet rule CAFE with reformed car/truck differentiation Current CAFE targets with tradable credits Current CAFE but with E-CAFE weight- based targets Tradable credits E-CAFE weight-based targets UPI for each manufacturer Feebates Carbon taxes/ carbon trading Gasoline taxes Decrease weight Reduce weight of heavy vehicles; increase weight of light vehicles Decrease weight Reduce performance Incorporate more fuel- efficient technology Small increase in VMT Decrease in VMT No impact on existing vehicle use Shift to reduce fuel use No impact on vehicle maintenance Improve maintenance Insignificant impact on vehicle retirements Increase retirements of low fuel economy vehicles

96 EFFECTIVENESS AND IMPACT OF CORPORATE AVERAGE FUEL ECONOMY (CAFE) STANDARDS and speed of the intervention—and on the choices the manu- facturers and the consumers make in responding to the imple- mentation. Table 5-1 incorporates an assumption that sever- ity and implementation speed of the interventions are similar for each instrument. Table 5-1 indicates that neither CAFE nor the alternative instruments directed at fuel economy (feebates, tradable credits, weight-based targets, tradable credits with weight- based targets, UPI for each manufacturer) have any signifi- cant gasoline-reducing impacts through VMT, existing ve- hicle use, vehicle maintenance, or old vehicle retirements. Only the two broad-based demand reduction policies— carbon taxes/trading or gasoline taxes—provide incentives for people to drive less, to use their more fuel-efficient ve- hicle when possible, to maintain existing vehicles, or to re- tire old fuel-inefficient vehicles. Any of the instruments directed toward new car fuel economy would result in small increases in VMT, usually referred to as the rebound effect, because increasing the fuel economy of new vehicles decreases the variable cost per mile of driving, encouraging consumers to drive more. As dis- cussed in Chapter 2, empirical studies suggest that every 10 percent increase in gasoline price leads to a 1 or 2 percent reduction in vehicle miles. Similarly, a 10 percent increase in fleet economy can be expected to lead to a 1 or 2 percent increase in VMT. The net impact would be an 8 or 9 percent reduction in fuel use for every 10 percent increase in fuel economy of the entire vehicle fleet. Table 5-1 indicates that all the policy instruments will motivate changes in technology that reduce gasoline use, and they all motivate reductions in performance. Each instru- ment, other than those including proportionate weight-based targets, may reduce fuel use through reductions in the aver- age vehicle weight. Proportionate weight-based targets, on the other hand, are unlikely to motivate weight reductions, since reducing vehicle weights would not bring the manu- facturer closer to the targets. The broad demand-reduction policies differ from the other policies in the way they motivate changes in weight, power, and technology. CAFE and the alternative instru- ments directed at fuel economy are felt directly by the manu- facturer and give it a direct incentive to offer lighter, lower- performance cars with more energy-efficient technologies. Gasoline taxes and carbon taxes motivate consumers to shift their new vehicle purchases, directing them toward lighter, lower-performance cars with more energy-efficient technologies, which, in turn, provides a stimulus to the manufacturers. In principle, a direct incentive for manufacturers to sell vehicles with increased fuel economy and an equally strong incentive for consumers to buy vehicles with increased fuel economy should have similar fuel economy impacts. Manu- facturers would change the characteristics of their vehicles in anticipation of changes in consumer choices caused by higher fuel taxes. Manufacturers facing a direct incentive would respond to that incentive and would pass the financial incentives on to the consumers, who would modify their choice of vehicles. In practice, however, considerable uncertainty is involved in calibrating a gasoline tax or carbon tax so that it would have the same impact as a direct incentive for manufactur- ers. If consumers discounted future gasoline costs over the entire vehicle life, using identical discount rates, a gasoline tax and a direct manufacturer incentive would change fuel economy to the same extent if the manufacturer incentive were equal to the discounted present value of the gasoline- tax incentive. However, if some new vehicle buyers count, say, only 3 years of gasoline costs, the discounted present value of the gasoline tax cost over 3 years would provide an equivalent manufacturer incentive. The results of the two calculations differ by a factor between two and four, depend- ing on the discount rate used,29 so there is significant uncer- tainty about how large a gasoline tax would, in fact, be equivalent to a direct manufacturer incentive. Minimizing the Costs of Fuel Economy Increases A fuel economy policy may produce an increase in manu- facturing cost through changes that require more costly ma- terials, more complex systems, more control systems, or en- tirely different power train configurations. A fuel economy policy may also impose nonmonetary costs on consumers through reductions in vehicle performance or interior space. Performance reductions may make vehicles less attractive for towing trailers, for merging quickly onto freeways, or for driving in mountainous locations. Next, the committee examines whether particular policy instruments would minimize the overall cost of whatever fuel reductions are achieved. This is a matter both of the structure of the policy instruments and of the aggressiveness of policy implementation, including the lead time manufacturers have to make the requisite adjustments. Also examined is whether the policy instruments can be expected to reach an appropri- ate balance between higher vehicle costs and the savings from reduced gasoline use (including external costs of fuel use). Each issue of cost-minimizing policy and cost-minimiz- ing manufacturer and consumer response can be examined based on most likely costs. However, no one has perfect 29Consider a gasoline tax intended to embody the same incentive as a tradable credit with a credit price of $0.26/gal over a 150,000-mile lifetime. Under tradable credits, a vehicle with a fuel economy of 25 mpg when the standard was 30 mpg would need 150,000 × [(1/25) – (1/30)] credits, or 1,000 credits, with a value of $260. Assume the buyer discounts gasoline use over only the first 3 years, at a 10 percent discount rate, driving 15,000 miles per year. A gasoline tax of $1.04/gal would give a discounted tax difference of $260 between a 30-mpg and a 25-mpg car, the same as the value of the credits. Thus, for such a consumer, a manufacturer incentive of $0.26/gal for each of 150,000 miles is economically equivalent to a $1.04/ gal gasoline tax.

POTENTIAL MODIFICATIONS OF AND ALTERNATIVES TO CAFE 97 knowledge about all costs and no one can predict with confi- dence changes in technology or economic conditions. There- fore, the discussion focuses on whether various policy in- struments allow manufacturers, consumers, and policy makers to respond to uncertainty in future conditions. This section provides a conceptual framework for exam- ining these questions. Minimizing the Total Cost of Fixed Economy Increases The issue of minimizing the total cost to society of what- ever average fuel economy increases are achieved depends critically on the observation that vehicles differ significantly from one another in size, weight, body type, and features, and that some manufacturers offer a full line of such ve- hicles and others offer only a limited one. To minimize the overall cost to society of a given reduction in fuel use, the marginal cost of reducing gasoline use must be the same (or approximately the same) across all vehicles.30 Manufactur- ers must be able to meet this condition for all vehicles, in- cluding domestic and import fleets of one manufacturer. Thus, under ideal conditions, policy options should not in- duce significant differences in the marginal costs among manufacturers. Policies that do induce significant differences will not minimize the total cost to society of whatever level of average fuel economy increase is achieved. Balancing the Costs and Benefits of Reducing Fuel Use Two components, private costs (and benefits) and exter- nal costs, must be considered when evaluating costs and ben- efits of changing fuel use. The private costs of the vehicle and its operation, including the fuel costs, can be estimated by the new vehicle purchaser through reading magazines, government reports, or examining the stickers on the car or truck being considered for purchase. It is in the purchaser’s own interest to take into account these private costs, not just for the time he or she expects to own the vehicle but also for periods after it is sold as a used vehicle, since the resale value can be expected to reflect these gasoline purchase costs. The manufacturer can adequately balance this portion of costs against the manufacturing costs of reducing fuel use. Accounting for external costs is more difficult. Absent policy intervention, it is not in the owner’s own interests to account for these costs. These costs must be incorporated into financial incentives facing consumers or manufacturers to balance the costs and benefits of reducing fuel use: The efficient policy intervention would require the manufacturer or consumer to face a financial incentive that added a cost equal to the external marginal cost of additional gasoline used. If the incentive were substantially greater than the marginal external cost, then the intervention would be too severe—the fuel savings would cost more than they were worth. If the incentive were substantially smaller than the marginal external cost, then the intervention would be too weak—the additional fuel savings would cost substantially less than they were worth. Thus, given the magnitude of ex- ternal costs, one can determine the appropriate severity of incentives. This task is complicated by the uncertainty facing the policy process. Although future gasoline prices can be esti- mated, they cannot be predicted with certainty (EIA, 2000). Estimates of external costs of gasoline use have wide error bands. In addition, estimates of the marginal cost of fuel use reductions will remain quite uncertain until these costs are engineered into vehicles. Therefore, in evaluating any par- ticular policy instrument, it is important to assess how the instrument would operate given real-world uncertainties. The task is also complicated by the existing gasoline taxes. The consumer currently faces $0.38/gal combined state and federal excise taxes on gasoline as part of the gaso- line price. These taxes are costs from the perspective of the consumer but are transfer payments from the perspective of the nation. For consumer decisions about how much to drive, these gasoline taxes can be seen as user fees covering the costs of building and maintaining roads, highways, and as- sociated infrastructure and providing services to motorists. However, for consumer decisions about fuel economy of new vehicles, the user service interpretation is inappropriate. As discussed above, the committee has identified a total exter- nal marginal cost associated with the fuel economy of new vehicles of about $0.26/gal, based on CO2 emissions and international oil market impacts. But the price the consumer faces in purchasing gasoline already includes an average of $0.38/gal, an amount larger than the committee’s estimate of those external costs that vary with fuel economy of new ve- hicles. Thus, if existing gasoline taxes are included, con- sumers are already paying tax costs larger than the external costs seen as justifying intervention to increase fuel economy above the market-determined levels. If existing gasoline taxes are excluded, perhaps because they are used primarily as a mechanism to collect revenue for states and for the fed- eral government, then the $0.26/gal remains as a reason for intervening so as to increase fuel economy beyond market- determined levels. 30This equal-marginal-cost rule is easiest to see by considering an out- come for which different vehicles had very different marginal costs of re- ducing fuel use. For example, assume that the cost for vehicle A of reducing lifetime fuel use by 200 gallons is slightly more than $300 for the next 200 gallons reduction or slightly less than $300 for the previous 200 gallons (that is, about $1.50/gal) and assume that the cost for vehicle B of reducing lifetime fuel use by 200 gallons is slightly more than $500 for the next 200 gallons or slightly less than $500 for the previous 200 gallons (that is, about $2.50/per gal). In that situation, reducing vehicle A lifetime gasoline use by 200 gallons and increasing vehicle B lifetime gasoline use by 200 gallons would not change the total lifetime fuel use for the two cars but would reduce total costs by about $200, the difference between the saving for vehicle B and the additional cost for vehicle A.

98 EFFECTIVENESS AND IMPACT OF CORPORATE AVERAGE FUEL ECONOMY (CAFE) STANDARDS Analysis of the Costs of Various Policy Instruments With that framework in place, the policy instruments can be analyzed in terms of cost minimization for a given fuel economy increase and balancing of marginal costs of fuel economy increases with marginal benefits. Table 5-2 pro- vides a summary. Several of the policy instruments satisfy all rules for mini- mizing the cost for obtaining a given mean increase in fuel economy: feebates, gasoline taxes, carbon taxes, and tradable credits with attribute-independent or with proportionate31 weight-based targets. Therefore, each of these instruments would minimize the cost of obtaining a given mean fuel economy increase. Furthermore, gasoline taxes and carbon taxes would roughly minimize the overall cost of reducing fuel use, not simply of increasing fuel economy, since they also include incentives to reduce VMT. In addition, each of these instruments (feebates, tradable credits, gasoline taxes, or carbon taxes) could efficiently balance the costs and ben- efits of fuel economy even under uncertain or changing eco- nomic conditions, if the incentives in these are chosen cor- rectly. This would require the marginal incentives to equal the marginal value of the external costs (in our example, $0.26/ gal of gasoline). At the other end of the spectrum, two of in- struments—the current CAFE targets and UPI for each manu- facturer—satisfy none of the conditions unless the two-fleet rule is eliminated and the distinction between passenger cars and trucks is appropriately reformed. Elimination of the two- fleet rule and elimination or appropriate repair of the distinc- tion between passenger cars and trucks would improve both systems. Even if the two-fleet rule is eliminated and the car/ truck distinction appropriately repaired, neither of these in- struments can be expected to lead to an efficient allocation of reductions among manufacturers or to an economically effi- cient overall level of fuel economy increases. Performance Trade-offs for the Various Policy Instruments The discussion above makes it clear that there are trade- offs among the various performance objectives important for the various policies. In particular, there may be trade-offs between three separate performance objectives: • Flexibility of choice for manufacturers and consumers to allow appropriate fuel economy increases, while maintaining consumer choice; • Certainty, or predictability, in the magnitude of total fuel consumption reductions, or at least certainty in the magnitude of new vehicle fuel economy increases; and • Effectiveness in motivating the broadest range of ap- propriate consumer and manufacturer fuel use re- sponses, and possibly the greatest fuel use reductions. Table 5-3 summarizes these trade-offs. For each policy instrument it provides a summary of performance of the in- strument in terms of these three objectives and compares them where possible. Each summary of the performance characteristics is stated relative to the current CAFE system. Since none of the policy instruments provides certainty about the magnitude of fuel use reductions, primarily because TABLE 5-2 Issues of Cost Minimization for the Various Policy Instruments Can Policy Assure For Manufacturer, Balance of Costs For Manufacturer, Is Marginal Cost Is Marginal Cost and Benefits with Is Marginal Cost of of Fuel Reduction of Fuel Uncertain or Fuel Reduction Same Across Reduction Same Changing Same Across Fleet Domestic Fleet and Across All Economic Policy Instrument of Cars and Trucks? Import Fleet? Manufacturers? Conditions? Current CAFE, with two-fleet rule Unlikely Current CAFE, without two-fleet rule No. Probably significant Yes No. Probably significant differences differences No CAFE with reformed car/truck Only if two-fleet rule differentiation removed Current CAFE targets with tradable Yes Yes Yes Yes credits Current CAFE but with E-CAFE No. Probably significant Only if two-fleet rule Better than current CAFE No weight-based targets differences removed Tradable credits, E-CAFE weight- No Yes Yes Yes based targets UPI for each manufacturer Same as CAFE Same as CAFE No No Feebates Yes Yes Yes Yes 31The addition of the enhanced-CAFE weight-based targets, even with tradable credits, would give different incentives for increasing fuel economy for heavy and light cars.

POTENTIAL MODIFICATIONS OF AND ALTERNATIVES TO CAFE 99 none can directly control total vehicle miles, the second col- umn of Table 5-3, “How certain will be the magnitude of new car fuel economy increases?”, summarizes the degree of cer- tainty in the magnitude of new vehicle fuel economy increases not in the magnitude of total fuel use reductions. The third column, “How broad a range of fuel use re- sponses would be motivated?”, summarizes the material from Table 5-1. Broader ranges of responses tend to lead to larger responses for a given magnitude of incentives and thus larger expected reductions in fuel use. Table 5-3 underscores the observation that choice among instruments will necessarily involve trade-offs among those three performance objectives. In particular, there is an inher- ent trade-off between flexibility and certainty. Generally, policy instruments that give the most certainty of response are the ones that allow the least flexibility of choice. In addition, certainty of response should not be confused with magnitude of response. For equivalent magnitudes, the broad energy de- mand reduction policies—gasoline taxes and carbon taxes— have less certainty of magnitude but can be expected to lead to the greatest reductions in fuel use. In addition to the trade-offs highlighted in Table 5-3, the various policy options will have additional impacts and therefore additional trade-offs. The following section of this chapter points out some of the additional impacts of the vari- ous policy options. Table 5-3 incorporates an implicit assumption that the se- verity and implementation speed of the interventions are simi- lar for each instrument. However, this analytic requirement for consistency is not a requirement for policy making. For example, it is quite possible to decide whether one would pre- fer to implement a system of tradable credits with aggressive fuel economy targets or to increase the targets of the CAFE system less aggressively. Or, one could choose between a moderate carbon tax system or an aggressive gasoline tax in- crease. Further analysis of the trade-offs involved in such choices is a necessary part of the policy-making process. Other Potential Consequences In addition to issues discussed above, there may be other consequences, some unintended, associated with these in- struments. Six classes of potential consequence have been identified: distributional impacts, automotive safety, con- sumer satisfaction, mobility of the population, employment, and environmental impacts. These will be discussed in what follows. Distributional Impacts of the Instruments In addition to the economic efficiency impacts of the vari- ous policy instruments, there are also impacts on the distri- bution of income/wealth between automotive companies, consumers, and the government. Carbon taxes and gasoline taxes could have the greatest distributional consequences: Vehicle owners would pay taxes to the federal government TABLE 5-3 Performance Trade-offs for the Various Policy Instruments Will Manufacturers and Consumers Have Flexibility How Broad a Range of Choice to Allow How Certain Will Be the of Fuel Use Responses Economically Efficient Fuel Magnitude of New Car Would Be Policy Instrument Economy Increases? Fuel Economy Increases? Motivated? Current CAFE, with two-fleet rule No Broad range of fuel economy Current CAFE, without two-fleet rule No, but more flexibility Much certainty responses than current CAFE CAFE with reformed car/truck differentiation Similar to current CAFE More certainty than current CAFE Current CAFE targets with tradable credits Yes Less certainty than current CAFE Broader than current CAFE Current CAFE but with E-CAFE weight-based More flexibility than current targets CAFE Tradable credits, E-CAFE weight-based targets Yes UPI for each manufacturer Similar to current CAFE Similar to current CAFE Broader range than current CAFE Feebates Less certainty than current CAFE Broader range than current CAFE Gasoline taxes Yes Broader than current CAFE, Considerably less certainty than plus VMT responses Carbon taxes current CAFE Like gasoline taxes, plus economy-wide responses

100 EFFECTIVENESS AND IMPACT OF CORPORATE AVERAGE FUEL ECONOMY (CAFE) STANDARDS even if the fuel economy of their cars or trucks exceeded the fuel economy targets. For example, with a $0.26/gal gaso- line tax, the consumer driving a 30-mpg vehicle over a 150,000-mile lifetime would pay $1,300 in gasoline taxes during the life of the vehicle. The payment to the govern- ment would allow other taxes to be reduced or would allow additional beneficial government spending. Current CAFE standards (or UPI), with or without the two-fleet rule, with or without a redesign of the distinction between cars and trucks, and with or without attribute-based targets, would not lead to any financial transfer among manu- facturers who meet or exceed the targets. However, if, to meet the targets, those manufacturers need to increase the price of low-mpg vehicles and decrease the price of high- mpg vehicles, then this pricing strategy would have a differ- ential impact on consumers. In addition, those manufactur- ers who fail to meet the targets would pay financial penalties to the government. Like the current CAFE systems, feebates and tradable fuel economy credits cause no financial transfers for those man- ufacturers just meeting the target. Like the current CAFE systems, both systems would lead to payments from those manufacturers who failed, on average, to meet the targets. A difference is that, unlike the current CAFE system, those manufacturers exceeding the fuel economy targets receive a financial transfer and that transfer provides the motivation for further increases in fuel economy. Feebates would require manufacturers with fuel economy lower than the targets to pay money to the federal govern- ment, while those with higher fuel economy would receive payments from the government. Although money would pass through the government, on net there would be a transfer of payments between the two groups of manufacturers. If the mean fuel economy of the entire new vehicle fleet exceeds the mean target, there would be a financial transfer from the government to the automotive industry as a whole; con- versely, there would be a financial transfer from the industry to the government if the mean economy is lower than the average target. Thus, whether the system adds net tax rev- enues or subtracts from tax revenues would depend on the average fuel economy of the new cars sold. Tradable credits, either with or without weight-based tar- gets, could lead to financial transfers between automobile manufacturers. However, the transfers would be smaller (on a per-vehicle basis) than the transfers associated with an equivalent gasoline tax or a carbon tax. In the example above, with a 30-mpg target and a $0.26/gal price of credits, a manu- facturer whose mean economy was 35 mpg would receive $185 per vehicle from selling credits, and the manufacturer of a 25-mpg vehicle would need to pay $260 per vehicle for purchasing credits. These could be compared with the $1,114 per-vehicle consumer payment of gasoline taxes (at $0.26 per gallon) for the 35-mpg vehicle and the $1,560 per- vehicle consumer payment of gasoline taxes for the 25-mpg vehicle. Safety As discussed in Chapters 2 and 4, over the last several decades, driving continued to become safer, as measured by the number of fatalities and severe injuries per mile of driv- ing. This trend has resulted from a combination of factors, including improvement in highway design; enforcement of traffic laws, including alcohol restrictions; and improved design of vehicles. Although most of these factors are apt to be unaffected by changes in CAFE standards, one of them, the design of vehicles—in particular, their weights and sizes—may well be influenced by changes in the form and severity of CAFE standards. The relationships between weight and risk are complex and have not been dependably quantified. However, in gen- eral, it appears that policies that result in lighter vehicles are likely to increase fatalities (relative to their historic down- ward trend), although the quantitative relationship between mass and safety is still subject to uncertainty. Chapter 4 estimated that about 370 additional fatalities per year could occur for a 10 percent improvement in fuel economy if downweighting follows the pattern of 1975– 1984. The downweighting by itself would improve fuel economy by about 2.7 percent for cars and 1.0 percent for trucks, which would reduce gasoline consumption by about 4.3 billion gallons. Thus, if this relationship accurately de- scribed future downweighting, and if for policy analysis pur- poses a benefit of $4 million is assigned to every life, then the safety costs would be $1.4 billion, or $0.33 in costs of lost lives per gallon of gasoline saved by downweighting.32 If, on the other hand, weight reduction was limited to light trucks, the net result could be a reduction in fatalities, with a safety benefit. Note that there is much uncertainty in the es- timate of value of lives and in the number of fatalities. Higher or lower values of these figures would increase or decrease the estimate in safety costs or benefits proportionately. As discussed above, proportionate weight-based targets would eliminate motivation for weight reductions and thus would avoid any adverse safety implications. Inclusion of the enhanced-CAFE weight-based targets would eliminate any motivation for weight reductions for small vehicles and would concentrate all weight reductions in the larger vehicles, in- cluding most light-duty trucks, possibly improving safety. 32The committee has not determined what fraction of this cost consum- ers already take into account in their choices of vehicles. If consumers are already taking into account consequences of vehicle purchases for their own safety, these safety consequences help explain consumer preferences for larger cars. If consumers do not take this cost into account, the $0.33 per gallon cost of downweighting overwhelms the $0.26 per gallon reduction in external costs that would result from downweighting. Finally, if consumers already fully include consequences of car purchases for their own safety but ignore all consequences of car purchases for the safety of other drivers with whom they may collide, then these costs imposed on others would represent another external cost, one associated strictly with choice of new vehicle weight. The committee has not quantified this possible source of external costs.

POTENTIAL MODIFICATIONS OF AND ALTERNATIVES TO CAFE 101 All other policy instruments could be expected to reduce vehicle weights and thus, all else equal, could be expected to have unintended safety impact. Any policy instrument that encourages the sale of small cars beyond the level of normal consumer demand may have adverse safety consequences. Consumer Satisfaction Reductions in weight or performance of vehicles below that desired by consumers could reduce vehicle purchaser satisfaction by reducing the utility of the vehicles available to consumers.33 Reduction in satisfaction might reduce pur- chases of new vehicles and thus adversely impact the auto industry and employment in that industry. This issue may be particularly important for policies that implicitly or explic- itly restrict automotive firms from selling vehicles designed to appeal to consumers whose preferences differ from those of typical customers. For example, some consumers expect to use vehicles for towing boats or trailers, for farming, or for construction, activities in which vehicle power—at rela- tively low speeds—is more important than fuel economy. Others consumers may desire luxury features that sacrifice fuel economy. Some may use their vehicle for carpooling several families or transporting youth soccer teams, Girl Scout troops, or school groups and thus may need the ability to safely seat many people. All of the policy instruments provide strong incentives to reduce performance, and many also provide incentives to reduce weight. Such reductions could have unintended nega- tive consequences. Policy instruments that provide no moti- vation for weight reduction are less likely to have such nega- tive consequences. Mobility Personal mobility is highly valued in American society. People living in suburban and rural locations often have no alternatives to light-duty vehicles for personal mobility. Large numbers of people cannot afford to live close to their work, and many families include two wage-earners who may work a significant distance from one another. Thus, any poli- cies that reduce the mobility of these people may create un- intended hardships. Mobility may be reduced by policies that greatly increase the cost of driving. Large enough gasoline taxes or carbon taxes could have this impact. Conversely, any policy that, all else equal, increases fuel economy could reduce the cost of driving and could increase options for mobility. Employment Employment in the U.S. economy is linked primarily to monetary and fiscal policies pursued at the federal level and to regional policies that allow a range of employment oppor- tunities throughout the United States. None of the policy options can be expected to significantly impact monetary and fiscal policies. However, policies that reduce the num- ber of vehicles manufactured in the United States or rapidly and significantly reduce the scale of an industry central to a regional economy could have at least temporary regional employment impacts. The committee believes that none of the policies dis- cussed here would have such negative consequences if implemented wisely. If implemented too aggressively, any of the policies could greatly increase the cost of vehicles or their use and thus have the potential for harming employ- ment. In the remainder of this chapter, it is assumed that no policy will be implemented so aggressively that there would be such employment impacts. One particular issue was raised in the committee’s open sessions. It was suggested that maintenance of the two-fleet rule, requiring each manufacturer’s domestic fleet and im- ported fleet of passenger cars to separately meet the CAFE standards, was important for avoiding job losses in the United States, particularly in the automotive industry. How- ever, the committee found no evidence nor was it offered any evidence or analysis to support that contention. In addi- tion, the current rule does provide a strong motivation for manufacturers to reduce domestic content of some vehicles, particularly the larger vehicles, to keep them in the import- fleet category. Therefore, while at some time the two-fleet rule may have protected domestic production, the committee sees no reason to believe it continues to play this role at all. Therefore, elimination of the two-fleet rule is not expected to have net adverse impacts on employment in the U.S. auto- motive industry. Environment Environmental impacts can be viewed in two categories, depending on whether the impacts are closely related to the amount of gasoline used or are independent of gasoline use but instead are dependent on the VMT and on the character- istics of individual vehicles. In the first category are the environmental consequences of the release of CO2, which is directly proportional to the amount of gasoline consumed. These direct environmental externalities, discussed previously, are a major reason for market intervention. In the second category are the environmental conse- quences of criteria pollutants emitted from cars and trucks. For a given vehicle, the more it is driven, the greater will be the amount of emissions released and the greater the envi- ronmental impact. Thus, policies that reduce VMT—gaso- 33The effect of weight reduction on consumer satisfaction is compli- cated, however, because structural redesign and the use of lightweight ma- terials can allow weight reduction without changes in structural integrity or handling.

102 EFFECTIVENESS AND IMPACT OF CORPORATE AVERAGE FUEL ECONOMY (CAFE) STANDARDS line taxes and carbon trading/taxes—can lead to additional environmental benefits. New cars emit far less criteria pollutants per mile driven than do cars of older vintage. Thus, policies that reduce re- tirements of old cars (by, for example, significantly increas- ing the costs of new vehicles) would increase the average emissions per vehicle mile traveled and the total emissions of criteria pollutants. Potential Inequities The issue of equity or inequity is subjective. However, one concept of equity among manufacturers requires equal treatment of equivalent vehicles made by different manufac- turers. The current CAFE standards fail this test. If one manufacturer was positioned in the market selling many large passenger cars and thereby was just meeting the CAFE standard, adding a 22-mpg car (below the 27.5-mpg stan- dard) would result in a financial penalty or would require significant improvements in fuel economy for the remainder of the passenger cars. But, if another manufacturer was sell- ing many small cars and was significantly exceeding the CAFE standard, adding a 22-mpg vehicle would have no negative consequences. This differential treatment of identical vehicles charac- terizes the current CAFE system with or without the two- fleet rule, with or without reclassifications of trucks, and without weight-based targets. With the enhanced-CAFE weight-based targets, this differential treatment would con- tinue to exist. Another notion of equity involves whether manufacturers are rewarded, treated neutrally, or punished for incorporat- ing fuel-economy-enhancing technologies when they are not required by law to do so. Rewards or neutral treatment seem equitable; punishment seems inequitable. Uniform percentage improvements would operate in this inequitable manner. Consider two initially identical manu- facturers initially selling identical fleets of vehicles, both just meeting current CAFE standards. Suppose that one manu- facturer of its own volition introduces improved technolo- gies that increase fuel economy and the other does not; sup- pose further that some years later the government adopts a UPI regulation. The first manufacturer would then be re- quired to achieve a higher fuel economy than the second. But the first manufacturer would already have used the low- cost fuel-saving technologies and would be forced to use higher-cost technologies for the further improvement. That manufacturer would be significantly punished for having improved fuel economy beyond what was required by law. A final concept of equity among manufacturers is more global. New policies that would impose costs disproportion- ately on particular manufacturers, who themselves have re- mained in compliance with existing law and policies, seem inequitable because they would impose unequal costs on oth- erwise similarly situated manufacturers. A policy decision to simply increase the standard for light-duty trucks to the same level as for passenger cars would operate in this inequitable manner. Some manufac- turers have concentrated their production in light-duty trucks while others have concentrated production in pas- senger cars. But since trucks tend to be heavier than cars and are more likely to have attributes, such as four-wheel drive, that reduce fuel economy, those manufacturers whose production was concentrated in light-duty trucks would be financially penalized relative to those manufac- turers whose production was concentrated in cars. Such a policy decision would impose unequal costs on otherwise similarly situated manufacturers. Administrative Feasibility Perhaps the easiest policy to implement would be in- creases in gasoline or other motor fuel taxes. There is al- ready a system of gasoline taxation in place, and implemen- tation of the policy would simply involve increasing the tax rate. No additional administrative mechanisms would be required. It is also relatively easy to enforce the current CAFE stan- dards. These standards, with or without the two-fleet rule, involve reporting the sales of the various models of vehicles and reporting the fuel economy of each model. The fuel economy data are already developed for reporting to con- sumers. The sales data and the associated mathematical calculations could be self-reported. Accounting for carry- forward and penalties is straightforward. Thus, CAFE stan- dards involve little or no administrative difficulty except in setting the standards. However, setting target levels requires significant economic and technological information, so the levels should be revised periodically. Similarly, weight- based targets require only one additional piece of readily available data, the vehicle curb weight. Thus, the adminis- trative issues would be virtually no different from those of the current CAFE standards. Feebates require financial transfers between the auto manufacturers and the government and thus require a report- ing and collecting function. This function could be integrated with other taxation functions but would require administra- tive efforts. Tradable credits, with or without weight-based targets, require a new administrative mechanism. Perhaps the easi- est approach would involve an extension of the reporting mechanisms required under the current CAFE standards. At the end of the year, the manufacturer would report the number of cars sold and their fuel economies (based on the standard testing procedure) and their weights, if weight- based targets are to be used. They would also report the number of fuel economy credits purchased or sold, the names of the other companies involved in the transactions, and sales prices of credits. The government agency might require public reporting of the credit sales prices to allow

POTENTIAL MODIFICATIONS OF AND ALTERNATIVES TO CAFE 103 use of this information for further rule making. Reporting on the names of the other companies involved in transac- tions would allow cross-checking to assure that every re- ported purchase of credits had a corresponding reported sale. The government agency would also need a mecha- nism to sell credits to any firm that decided to buy them. In general, however, the administrative requirements would be modest. Introducing a carbon tax or carbon trading mechanism would require a system that extended well beyond just the automobile manufacturers. Discussion of how one might set up such a system is well beyond the scope of this report. REFERENCES Davis, W.B., M.D. Levine, and K. Train. 1995. Effects of Feebates on Ve- hicle Fuel Economy, Carbon Emissions, and Consumer Surplus. DOE/ PO-0031. Department of Energy, Office of Policy, February. Washing- ton, D.C.: DOE. Delucchi, M.A., D.L. Greene, and M. Wang. 1994. “Motor Vehicle Fuel Economy: The Forgotten Hydrocarbon Control Strategy?” Transporta- tion Research-A 28A: 223–244. EIA (Energy Information Administration). 2000. A Primer on Gasoline Prices (an Update). Available online at <http://www.eia.doe.gov/pub/ oil_gas/petroleum/analysis_publications/primer_on_gasoline_prices/ html/petbro.html>. EIA. 2001. Alternatives to Traditional Transportation Fuels. Table 13: Es- timated Consumption of Alternative Transportation Fuels in the United States, by Vehicle Ownership, 1997, 1999, and 2001 (Thousand Gaso- line–Equivalent Gallons). Available online at <http://www.eia.doe.gov/ cneaf/alternate/page/datatables/atf1-13_00.html>. EMF (Energy Modeling Forum). 1982. World Oil–Summary Report. EMF Report 6, Vol. 1, February. Stanford, Calif.: EMF, Stanford University. Kopp, R., R. Morgenstern, W. Pizer, and M. Toman. 2000. A Proposal for Credible Early Action in U.S. Climate Policy. Washington, D.C.: Re- sources for the Future. Available online at <http://www.weathervane. rff.org/features/feature060.html>. Leiby P.N., D.W. Jones, T.R. Curlee, and L. Russell. 1997. Oil Imports: An Assessment of Benefits and Costs. ORNL-6851. Oak Ridge, Tenn.: Oak Ridge National Laboratory. Stoffer, H. 2001. “CAFE Can Discourage U.S. Employment,” Automotive News, No. 5921:2.

104 This attachment develops the Enhanced-CAFE (E-CAFE) standard, an alternative to the current fuel economy regula- tions. This alternative has a number of advantages: It has the potential to decrease fuel consumption, reduce the “gaming” of the fuel economy standards, and increase the safety of the overall vehicle fleet. The committee views the new system as a serious alternative to the current CAFE standards. Be- cause of limitations of time and data, it has only been able to do an approximate calibration of the effects of the new sys- tem. Thus, although the E-CAFE standard is highly promis- ing, some additional analysis will be required. TARGETS UNDER THE CURRENT CAFE STANDARD Figure 5A-1 shows the general relationship between fuel consumption and vehicle weight for passenger cars, based on 1999 data. Fuel consumption, the vertical axis, is ex- pressed as number of gallons needed to drive 100 miles. Each point in the graph is a single car model, e.g., a four-cylinder Accord. (Car models that sell fewer than 1,000 cars per year in the United States are not shown.) The current CAFE standard sets a passenger car target of 27.5 mpg (3.64 gallons per 100 miles of driving) for each manufacturer. Compliance is determined by averaging gal- lons per mile across the manufacturer’s entire fleet of cars. With averaging, the manufacturer can produce some cars that get low mpg if it balances them with enough cars that get high mpg. The horizontal line in Figure 5A-1 shows the CAFE target. Point A in Figure 5A-1 represents a car model that con- sumes more fuel than is allowed by the CAFE standard. The gap between point A and the horizontal CAFE line is the amount of excess fuel consumption. Point B is a car that con- sumes less fuel than the CAFE standard. The gap between point B and the CAFE line is not as large, so the manufacturer who makes As and Bs will have to sell approximately two point B cars to offset the high fuel consumption of one point A car. Manufacturers have an incentive to sell more of the lighter Attachment 5A Development of an Enhanced-CAFE Standard FIGURE 5A-1 Gallons used per 100 miles (cars only, gasoline engines only). 2.0 3.0 4.0 5.0 6.0 1500 2500 3500 4500 5500 Curb Weight (pounds) G al lo ns u se d pe r 10 0 M ile s gap gap B A CAFE line 27.5 mpg (3.64 gal/100 miles

POTENTIAL MODIFICATIONS OF AND ALTERNATIVES TO CAFE 105 cars, which may lead to an increase in traffic fatalities. In ad- dition, there are equity problems: Those manufacturers who specialize in making large, heavy cars have a harder task meet- ing CAFE targets than those who specialize in making small, light cars. These disadvantages of the current CAFE system motivate the search for an alternative. THE RELATIONSHIP BETWEEN WEIGHT AND FUEL CONSUMPTION Figure 5A-2 fits a regression line through the car data— the upward-sloping straight line. There is strong relationship between weight and fuel consumption.1 Figure 5A-3 adds in the data for light-duty trucks (gasoline engines only; models that sold fewer than 1,000 vehicles per year in the United States are not shown). Again, there is a strong relationship between weight and fuel consumption, though with some- what more outliers than in the car graphs. A regression through the truck data was computed and is shown as a dashed line. It is nearly parallel to the car line. If all manufacturers exactly met the weight-based targets shown by the two regression lines, the total car fleet would average 28.1 mpg, and the total truck fleet would average 20.1 mpg, a difference of 8 mpg. But the two regression lines in Figure 5A-3 are only about 2.5 mpg apart. The rea- son for this apparent difference (8 mpg instead of 2.5 mpg) is that the regression lines estimate fuel consumption while holding weight constant. The 8-mpg car–truck gap occurs because the average car is being produced on the left-hand, low-weight end of the technology curve, while the average truck is being produced on the high weight end. Analyzing the components of the 8-mpg gap: 2.5 mpg of the gap is technological—trucks have more aerodynamic drag, and in general their drive trains are not as technologically advanced. And 5.5 mpg of the gap occurs because trucks are designed to be heavier than cars. To gain a better sense of the characteristics of specific vehicles, Figure 5A-4 shows a sample of 33 trucks and 44 cars that are representative of cars, vans, SUVs, and pickup trucks. For analytic purposes it is sometimes more conve- nient to normalize the vertical scale of Figure 5A-4: Divide the fuel consumption of each point by the curb weight. This ratio is the weight-specific fuel consumption (WSFC). Figures 5A-3, 5A-4, and 5A-5 show that fuel consump- tion is roughly proportional to the weight of the vehicles or, equivalently, that the weight-specific fuel consumption is roughly constant across the various weights. That is, the most significant variable explaining fuel consumption is weight. This suggests the possibility of basing fuel economy stan- dards on the weight of the vehicle. For example, use the 1Why do the points in Figure 5A-2 scatter? Imagine a vertical line drawn at the 3,000-lb point on the weight axis, and consider the cars that fall along that 3,000-lb line. The cars do not all have the same fuel consumption be- cause they do not all have the same powertrain technology, aerodynamic efficiency, and rolling resistance. The point where the 3,000-lb line crosses the sloping line represents the average technology of 3,000-lb cars. The sloping line is derived from a sales-weighted regression fit and therefore gives more importance to high-volume vehicles. For this reason it puts most weight on the most often used technologies within each weight class. FIGURE 5A-2 Regression line through the car data in Figure 5A-1 (passenger cars only, gasoline engines only). 2.0 3.0 4.0 5.0 6.0 1500 2500 3500 4500 5500 Curb Weight (pounds) G al lo ns u se d pe r 10 0 M ile s 3,000 pound cars

106 EFFECTIVENESS AND IMPACT OF CORPORATE AVERAGE FUEL ECONOMY (CAFE) STANDARDS 2.0 3.0 4.0 5.0 6.0 7.0 1500 2500 3500 4500 5500 Curb Weight (pounds) G al lo ns u se d pe r 10 0 M ile s Cars = triangles. Solid line is the average car. Trucks = squares. Dashed line is the average truck. FIGURE 5A-3 Gallons to drive 100 miles, with regression lines (cars and trucks, gasoline engines only). 2 3 4 5 6 2000 2500 3000 3500 4000 4500 5000 5500 Curb Weight, lb F ue l C on su m pt io n, g al lo ns / 1 00 m ile s Premium Small Comp VAN Entry Small Comp PU Midsize STD PU Near Luxury Midsize Small SUV Luxury Midsize Comp SUV Large Luxury Comp SUV Luxury Large Large SUV CAFE Standard for Cars CAFE Standard for Trucks WB 2000 CAFE: Cars CAFE: Trucks FIGURE 5A-4 Gallons used per 100 miles (all vehicles).

POTENTIAL MODIFICATIONS OF AND ALTERNATIVES TO CAFE 107 0 1 2 3 4 2000 2500 3000 3500 4000 4500 5000 5500 Curb Weight, lb W ei gh t- S pe ci fic F ue l C on su m pt io n Premium Small Comp VAN Entry Small Comp PU Midsize STD PU Near Luxury Midsize Small SUV Luxury Midsize Comp SUV Large Luxury Comp SUV Luxury Large Large SUV Weight Specific Fuel Consumption FIGURE 5A-5 Weight-specific fuel consumption (gal/100 miles/ton). sloped line in Figure 5A-2 as the target baseline. This con- trasts to the current CAFE system, which computes the gap between each vehicle and the horizontal CAFE line in Fig- ure 5A-1. Instead, use the sloped line in Figure 5A-2 and compute the gap between each vehicle and the sloped line. WEIGHT-BASED TARGETS VERSUS CURRENT CAFE TARGETS A regulatory system using weight-based targets would remove the intense incentives for manufacturers to down- weight their small cars, thereby reducing the potential neg- ative safety effects of the current system. It would also produce greater equity across manufacturers—under CAFE, manufacturers who make a full range of car sizes have a harder time meeting the standards. Weight-based targets also have three major disadvan- tages. First, because they are weight-neutral, the principal lever for influencing vehicle fuel economy is lost. Second, they remove most of the incentive behind the current re- search programs that are pursuing the use of lightweight materials to substitute for the steel in vehicles. Such pro- grams have the potential to reduce vehicle weight while pre- serving vehicle size, reducing fuel consumption while pre- serving safety. Third, and most important, weight-based standards could result in higher fuel consumption. Unlike with CAFE, there is no cap on the fleet average, so the average vehicle could move to the right (upweight) on the curve. Is this likely? Note that car weights and truck weights have been increas- ing over the past decade despite strong counteracting pres- sure from CAFE. Furthermore, the profit margin associated with large vehicles has traditionally been much higher than that associated with small ones. Thus there are substantial market incentives for manufacturers to increase vehicle weights and no restraints on their doing so once CAFE is removed. With these advantages and disadvantages in mind, the weight-based standard could be modified to become the Enhanced-CAFE standard discussed in the next section. The committee recommends that serious consideration be given to this alternative as a substitute for the current CAFE system. PRINCIPAL ALTERNATIVE: THE ENHANCED-CAFE STANDARD It is possible to combine the CAFE system with weight- based targets to preserve most advantages of each while eliminating most disadvantages. In particular, the combined system should improve safety, so it is called the Enhanced- CAFE (E-CAFE) system. The E-CAFE system is a way to restructure the current regulatory system. It creates a differ- ent kind of baseline for measuring compliance and hence creates different incentives for manufacturers—incentives that move the regulatory system toward some highly desir- able goals. One possible set of targets is illustrated below,

108 EFFECTIVENESS AND IMPACT OF CORPORATE AVERAGE FUEL ECONOMY (CAFE) STANDARDS but the actual targets would be determined by the legislative and regulatory process. It is possible to have separate E-CAFE baselines for cars and for trucks. However, there would be a substantial advan- tage to using a single baseline that applies to all light-duty vehicles—it would eliminate the “gaming” possibilities in- herent in a two-class system with different standards for each class. The horizontal line in Figure 5A-5 shows such a fuel consumption target: a single baseline used to measure per- formance deviations for both cars and trucks. For vehicles that weigh less than 4,000 lb, the target is sloped upward like the weight-based targets.2 For vehicles that weigh more than 4,000 lb, the target is a horizontal line like the current CAFE standard. This approach in effect uses a more stringent target for vehicles above 4,000 lb, creating incentives to use ad- vanced technology to improve power-train efficiency, reduce aerodynamic and rolling resistance losses, and reduce acces- sory power. E-CAFE creates a strong set of incentives to improve the fuel economy of the heaviest vehicles. Under current CAFE, if a manufacturer wishes to offset the excess fuel consump- tion of a large vehicle, it can do so by selling a light vehicle: The vertical gap of the large vehicle (“A”) in Figure 5A-1 is offset by the vertical gap of the small vehicle (“B”). But if the baseline is changed to E-CAFE (Figure 5A-6), the small vehicle does not generate a large credit because it is on the sloped portion of the baseline and its gap is measured with respect to the slope, not with respect to the horizontal line. For our illustrative example, the horizontal line is set at 20.7 mpg, the current CAFE standard for light-duty trucks. Each manufacturer is judged on its entire fleet of cars and trucks: Vehicles that use less fuel than the targets can bal- ance vehicles that use more. The committee recommends that a system of tradable credits, such as that described ear- lier in this chapter, be made part of the regulation. The E-CAFE targets can also be expressed in terms of the weight-specific fuel consumption (WSFC) of the vehicles, which is fuel consumption per ton of vehicle weight used in 100 miles of driving. This normalized measure is shown in Figure 5A-7. How would this proposal affect the different manufactur- ers? A fleetwide compliance measure was computed for each of the Big 3 manufacturers plus Honda and Toyota to mea- sure their position with respect to the illustrative E-CAFE targets. Compliance ranged from 3 percent below the targets to 6 percent above the targets. None of the major manufac- turers begins with a large compliance deviation. It is a rela- tively fair starting point. The system has a single set of targets for all vehicles. This eliminates any concerns about arbitrary truck/car distinctions and their possible manipulation, since all such distinctions would be eliminated. There would be a small incentive for lightweight vehicles to be made heavier and a large incentive for vehicles weigh- ing more than the cutoff weight to be made lighter. Thus, the variance in weight across the combined fleet should be lower. This reduction in weight variance would improve safety in car-to-car collisions. The present position of the lines could serve as the initial baseline under the E-CAFE system. It produces a combined car and truck fuel economy of 24.6 mpg.3 To improve the overall fleet fuel economy in subsequent years, the horizon- tal portion of the baseline would be lowered, while simulta- neously reducing the slope of the lower portion of the baseline. The slope of the lower portion could also be ad- justed to reflect the most cost-effective use of technology (see Chapter 4). If the E-CAFE system is adopted, there should be a phase-in period associated with the new stan- dards: Manufacturers have already made plans based on the existing CAFE standards and must be given time to analyze the implications of the new standards and to redo their prod- uct plans. AN ALTERNATIVE ATTRIBUTE SYSTEM Instead of basing the E-CAFE standards on curb weight, they might be based on some measure of the vehicle’s load- carrying capacity, such as gross vehicle weight (GVW). Thus, vehicles capable of carrying more load would be given more liberal fuel consumption targets. This concept has some potentially useful features, as described in Attachment 3A. For regulatory purposes, however, it would have some seri- ous problems. Passenger vehicles rarely travel under full-load condi- tions. For example, data on vehicle occupancy from the Na- tionwide Personal Transportation Survey show that the large seating capacity of these vehicles is typically unused. The average van carries only 2.1 people, the average SUV car- ries 1.7 people, the average pickup carries 1.4 people, and for “other trucks” the average was 1.1 people. All these occupancy figures need to be compared with the average automobile, which carries 1.6 people. That is, if GVW were used instead of curb weight, heavier vehicles would be al- lowed to consume more fuel all the time because they might carry a full load. There are no measured data available that would allow 2This figure is only an illustration of one possible implementation of E- CAFE. The equation for this part of the targets is: GPM100 = –0.409 + 1.31 times weight, where GPM100 is gasoline consumption in gallons per 100 miles and weight is the vehicle weight in thousands of pounds. 3This was calculated by assuming that every manufacturer complied with the E-CAFE standards. Compliance is measured by computing the gap be- tween the E-CAFE baseline and each vehicle’s estimated fuel-consump- tion, multiplied by the number of vehicles of that model that were sold— some would be positive numbers, some negative. These would be summed over all the models made by a given manufacturer.

POTENTIAL MODIFICATIONS OF AND ALTERNATIVES TO CAFE 109 2.0 3.0 4.0 5.0 6.0 1500 2000 2500 3000 3500 4000 4500 5000 5500 Curb Weight G al lo ns p er 1 00 m ile s Trucks Cars Targets 0 1 2 3 4 2000 2500 3000 3500 4000 4500 5000 5500 Weight, lb W S F C , G al lo ns /T on o f V eh ic le W ei gh t - 1 00 M ile s Premium Small Comp VAN Entry Small Comp PU Midsize STD PU Near Luxury Midsize Small SUV Luxury Midsize Comp SUV Large Luxury Comp SUV Luxury Large Large SUV Average Cars Trucks More Stringent Standard For Vehicles Greater Than 4000 lb. FIGURE 5A-6 Enhanced CAFE targets. FIGURE 5A-7 Enhanced CAFE targets in WSFC units.

110 EFFECTIVENESS AND IMPACT OF CORPORATE AVERAGE FUEL ECONOMY (CAFE) STANDARDS the determination of slopes and cutoffs under a GVW stan- dard. EPA does not measure fuel consumption under those extra weight conditions. It has adjustment formulas capable of estimating the change in fuel consumption for small changes in vehicle weight, but the uncertainty of the predic- tions becomes larger and larger as the prospective weight change increases. To do an accurate analysis to set standards under the GVW criterion, the EPA would have to rerun the tests across all the vehicles in the fleet. Recalibrating to a regulatory standard based on GVW would take a long time. 4Personal communication from Eldert Bonteko, Environmental Protec- tion Agency, Ann Arbor, to committee member Charles Lave. Finally, GVW is a rating, not a measure, and it is deter- mined by each manufacturer, using that manufacturer’s own judgment of carrying capacity. A senior EPA analyst4 char- acterized GVW as “a remarkably arbitrary figure.” As cur- rently determined, it lacks the objective reliability needed when setting a regulatory standard.

Next: 6 Findings and Recommendations »
Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards Get This Book
×
Buy Paperback | $59.00 Buy Ebook | $47.99
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

Since CAFE standards were established 25 years ago, there have been significant changes in motor vehicle technology, globalization of the industry, the mix and characteristics of vehicle sales, production capacity, and other factors. This volume evaluates the implications of these changes as well as changes anticipated in the next few years, on the need for CAFE, as well as the stringency and/or structure of the CAFE program in future years.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!