PAPERBACK
$56.00
Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 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.) 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
iVehicles 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.
83
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.
Unresolvecl 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
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.
OCR for page 83
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 tuel 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-
OCR for page 83
POTENTIAL MODIFICATIONS OF AND ALTERNATIVES TO CAFE
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 Consiclerecl
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-
85
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
tonics 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.
OCR for page 83
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, classes:
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
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.
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
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.
OCR for page 83
POTENTIAL MODIFICATIONS OF AND ALTERNATIVES TO CAFE
The Two-Fleet Rule Differentiating Between Domestic and
Imp ortecl 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. Averaging 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.
Tractable Fuel Economy Creclits
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.
Sunder 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-m~le
consumption.
87
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-Basecl 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.~°
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.
Aclopt Energy Demancl-Recluction 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.
Throughout 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.
i°Targets 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.
OCR for page 83
88
Gasoline Tax
EFFECTIVENESS AND IMPACT OF CORPORATE AVERAGE FUEL ECONOMY (CAFE) STANDARDS
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-ancl-Tracle 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 GovernmenVlnclustry
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 (PNGVJ
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."
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-
iil. Alston, EPA, Letter to the committee dated April 16, 2001.
OCR for page 83
OCR for page 83
OCR for page 83
OCR for page 83
OCR for page 83
OCR for page 83
OCR for page 83
OCR for page 83
OCR for page 83
OCR for page 83
OCR for page 83
OCR for page 83
OCR for page 83
OCR for page 83
OCR for page 83
OCR for page 83
POTENTIAL MODIFICATIONS OF AND ALTERNATIVES TO CAFE
ential between truck and car CAFE standards would elimi-
nate all CAFE-derived incentives for manufacturers to have
vehicles classified as trucks. 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.~3
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.~4
However, few of these vehicles ever use any fuel other
than gasoline. Estimates from the Energy Information Ad-
ministration (EIA) suggesti5 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-
i2This 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.
i3The 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 arid light-duty trucks.
i4E85 consists of 85 percent ethanol and 15 percent gasoline.
i5The data appear online at
9o
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
Tractable Fuel Economy Creclits
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. 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.~7
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.l8 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.
i6Applying 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.
i7In 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.
i8The 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
The tradable credits system would have another advan-
tage, especially if the sales price of tradable credits were
made public.~9 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 system2i 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
i9Regulations 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.
2iFor a more comprehensive study on feebates, see Davis et al. (1995).
91
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-Basecl 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 that
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
6.0- (~7 mpg)
5.5-
5.0-
. _
A
Q 3.5-
en
o
Is
CD
4.5-
4.0-
3.0-
2.5-
2.0-
EFFECTIVENESS AND IMPACT OF CORPORATE AVERAGE FUEL ECONOMY (CAFE) STANDARDS
(20 mpg)
~ . a'
. * .e,%
. .
a'
O — g~~tli ~ R
4 ~ 4~.
4 *by
(33 mpg) '4e~' ~
4 '444.
o .;
·4 R
4
(50 mp¢)
CAFE standard
27.5 mpg =
3.64 gals/100 miles
1.5-
1,500 2,500 3,500
Curb Weight
4,500 5,500
FIGURE 5-1 The operation of the current CAFE standards: pas-
sengers cars, gasoline engines only, 1999.
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
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.
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
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
an
2.D _
~ ' Hil' ~ ~ 'K. . _ _
~ 0 ~ ~ ~ OOO ~M O v
O ~ '~ ~ O ~ O ~
I, ~ . OF
~ 0~ ~
l
1500 2500 3500
Curb Weight
j
4500 5500
FIGURE 5-2 Fuel economy targets under the Enhanced-CAFE sys-
tem: cars with gasoline engines, 1999.
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.
Aclopt Energy Demancl-Recluction 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
93
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-ancl-Tracle 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
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 of] 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.
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 hi~h-
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 tee 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
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.
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.
101
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-
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
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, Of lice of Policy, February. Washing-
ton, D.C.: DOE.
103
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
Attachment 5A
Development of an Enhanced-CAFE Standard
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
6.0 -
5.0 -
. _
on
A
~ 4.0
Q
o
3.0-
2.0 -
1500
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
.
A
.
.
. .
·
L L L ~ A . - L
· .
.
.
~ \ gap
. ~ art.-. ,. ~ .
· A ~~ ^~ · ~
·4~4~ · ~ -
~^~ dA~ ~ —
~ ·4L/LA4J\~` T
CAFE line
27.5 mpg
(3.64 gal/100 miles
gap/ . . . ~~,,~ A
At. ^ ~~"
B ~
· ~~& · ~
· ·A
·`
HA
b
AL ,11.
1 1
2500 3500 4500 5500
Curb Weight (pounds)
FIGURE SA-1 Gallons used per 100 miles (cars only, gasoline engines only).
104
POTENTIAL MODIFICATIONS OF AND ALTERNATIVES TO CAFE
6.0
5.0 -
~ 4.0 -
CL
u)
a)
o
Cal
3.0-
2.0 -
/.
· ~
~ /
A —
_ ^
1 ~ ~ ~ . ~
· · ~ 4~ · I A
~ ~ ~ ~ 1
MA OF
· ~ ~ ~ AIL
a,` 4~. I ~3,000 pound
, cars
.
/
who
1500 2500 3500 4500 5500
Curb Weight (pounds)
FIGURE SA-2 Regression line through the car data in Figure SA-1 (passenger cars only, gasoline engines only).
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.) 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.
iWhy 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.
105
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
106
in
o
i
CD
7.0
6.0
.O
0 5.0
Q
- 4.0
3.0 -
2.0 - a
1500 2500 3500
Curb Weight (pounds)
~ ~~ .
I'd.
/ A
..'
FIGURE SA-3 Gallons to drive 100 miles, with regression lines (cars and trucks, gasoline engines only).
6
In
.F
lo
lo
In
o
a, 4 -
a
. _
Q
In
o
~ V
IL
5 -
/
~ Arm
EFFECTIVENESS AND IMPACT OF CORPORATE AVERAGE FUEL ECONOMY (CAFE) STANDARDS
Trucks = squares.
Dashed line is the
average truck.
.
~_, Ai
Cars = triangles.
Solid line is the
average car.
4500 5500
A\
/
/
X OK ~ CAFE: Trucks
· . ~ 8 -
0~ +
Xx
XX
~~
— We ~ - _ -
Premium Small
· Entry Small
~ Midsize
0 Near Luxury Midsize
· Luxury Midsize
0 Large
· Luxurv Larne
CAFE: Cars
+ Comp VAN
X Comp PU
>K STD PU
~~ Small SUV
:: Comp SUV
EB Luxury Comp SUV
· Large SUV
- CAFE Standard for Cars CAFE Standard for Trucks
—WB 2000
/
.
2- , __
2000 2500 3000 3500 4000
Curb Weight, lb
FIGURE SA-4 Gallons used per 100 miles (all vehicles).
4500 5000 5500
POTENTIAL MODIFICATIONS OF AND ALTERNATIVES TO CAFE
4
3 -
o
.~ -
An _
a
O 2 -
IL
c'
. _
c'
Q
cn
= 1 -
O-
2000
X
FAX
W ~ ~
C1 Premium Small
· Entry Small
/\ Midsize
o Near Luxury Midsize
· Luxury Midsize
o Large
· Luxury Large
Weight Specific Fuel Consumption
+ Comp VAN
x Comp PU
>< STD PU
~~ Small SUV
:: Comp SUV
s' Luxury Comp SUV
· Large SUV
><
· ~
.
2500 3000
3500
Curb Weight, lb
FIGURE SA-S 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
107
4000 4500 5000 5500
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.
PA/NC/ML ALTERNAT/VE: 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
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.
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
~7
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 ALTERNAT/VE 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
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
6.0
5.0
In
. _
o
o
4.0
Q
In
o
Id
CD
3.0
2.0
FIGURE SA-6 Enhanced CAFE targets.
4
In
. _
° 3-
s
. _
c'
. _
is, 2 -
>
o
t
o
CD 1 -
IL
U'
FIGURE SA-7 Enhanced CAFE targets in WSFC units.
109
·.
.
· do. - ~ · -!
\.~ ·~ · · "!
· ~ ·~. · ~ ~
..
.
· Trucks
Cars
Targets
1500 2000 2500 3000 3500 4000 4500 5000 5500
Curb Weight
X
Xx
X
w
A\
· ~ n '~' ^ O ~ .:: by- 83 ><
| More Stringent Standard For Vehicles |
I Greater Than 4000 lb. l
_ _ , , ~ ~~ , \ ~ =~ , ~
__~_L1__~-^-~ ~ - ~ ;-+--a- ~ ----i---
Trucks
Cars
[1 Premium Small
· Entry Small
/\ Midsize
0 Near Luxury Midsize
· Luxury Midsize
O Large
· Luxury Large
- - - Average
X
+ Comp VAN
Comp PU
STD PU
Small SUV
Comp SUV
~ Luxury Comp SUV
· Large SUV
)
o
2000 2500 3000 3500 4000 4500 5000 5500
Weight, lb
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.
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 analys* char-
acterized GVW as "a remarkably arbitrary figure." As cur-
rently determined, it lacks the objective reliability needed
when setting a regulatory standard.
4Personal communication from Eldert Bonteko, Environmental Protec-
tion Agency, Ann Arbor, to committee member Charles Lave.
Share
Research
Rights & Permissions
