Energy Policy ()

Contents lists available at ScienceDirect

Energy Policy
journal homepage: www.elsevier.com/locate/enpol

The role of CAFE standards and alternative-fuel

vehicle production credits in U.S. biofuels markets
Jarrett Whistance n, Wyatt Thompson
FAPRI-MU, 101 Park DeVille Drive, Suite E, Columbia, MO 65203, United States

H I G H L I G H T S

CAFE standards are estimated to have a moderate impact on RFS compliance costs.
Flex-fuel vehicle production incentives are estimated to have much less impact.
Level of auto manufacturer response to production incentives is uncertain.
Analysis suggests results are fairly sensitive to level of manufacturer response.

art ic l e i nf o

a b s t r a c t

Article history:
Received 31 March 2014
Received in revised form
3 September 2014
Accepted 4 September 2014

This paper investigates the impact of CAFE standards and alternative-fuel vehicle production incentives
on the biofuel market and mandate, in particular. The study develops a structural model of the domestic
light-duty vehicle sector, as well as reduced-form versions of domestic gasoline, diesel, and biofuel
markets. The results suggest that holding CAFE standards at the 2010 level could facilitate U.S. ethanol
market expansion, making it easier to meet the mandate. Alternative-fuel vehicle production incentives
appear to have small effects. However, there is uncertainty about the level of automaker response to
those incentives, and analysis indicates the model is fairly sensitive to the assumed level of response.
& 2014 Elsevier Ltd. All rights reserved.

Keywords:
CAFE standards
Alternative Motor Fuel Act
Renewable fuel standard
Flex-fuel vehicles

1. Introduction
The role of the U.S. government in supporting biofuels has been a
source of heated debate in recent years. Taken together, the passage
of both the Energy Policy Act (2005) (EPACT) of 2005 (PL 109-58) and
the Energy Independence and Security Act (2007) (EISA) of 2007
(PL 110-140) marked a turning point in biofuels policy. Those two
acts dene the Renewable Fuel Standard (RFS), which requires a
minimum amount of biofuel use each year at least to 2022, subject to
some exibility. Other forms of direct government intervention in
biofuels markets have included tax credits for ethanol and biodiesel
blending, tariffs on imported ethanol, and additional support for
cellulosic ethanol and other 2nd generation biofuel producers. In
light of these policies, extensive research investigates their effects,
and the general impacts of a larger biofuels market. Some authors
focus on broad topics such as social welfare of blend mandates
(de Gorter and Just, 2009) while others focus more narrowly on

Corresponding author.
E-mail addresses: WhistanceJL@missouri.edu (J. Whistance),
ThompsonW@missouri.edu (W. Thompson).

market impacts (Du and Hayes, 2009; Thompson et al., 2011b;

Whistance and Thompson, 2010). One result from the market
analysis literature is that a rising biofuel mandate must be reconciled
with the limits of ethanol use in the current nished motor gasoline
pool. Expansion beyond the market for E10, gasoline blended with
10% ethanol, requires more ethanol to be used in higher ethanol
blends. Some high-blend fuels, such as E85, can only be used by exfuel vehicles and might have to be distributed differently than
other fuels.
There are other government policies that affect the biofuel
industry less directly. This study primarily focuses on the Corporate Average Fuel Economy (CAFE) standards and the alternativefuel vehicle (AFV) production incentives contained within the
Alternative Motor Fuel Act (AMFA). In particular, we look at how
changes in the fuel economy standards might shape future motor
fuel demand, and thus biofuel demand, in the U.S. Furthermore,
we investigate how changes in AFV incentives, in conjunction with
the CAFE standards themselves, might also affect biofuel demand
in the U.S. Finally, we trace out how these effects interact with
biofuel use mandates of the RFS.
We develop and utilize a structural model of the light-duty vehicle
sector in the U.S. We use this model to simulate a forward-looking

http://dx.doi.org/10.1016/j.enpol.2014.09.004
0301-4215/& 2014 Elsevier Ltd. All rights reserved.

Please cite this article as: Whistance, J., Thompson, W., The role of CAFE standards and alternative-fuel vehicle production credits in U.S.
biofuels markets. Energy Policy (2014), http://dx.doi.org/10.1016/j.enpol.2014.09.004i

J. Whistance, W. Thompson / Energy Policy ()

baseline path for these markets assuming policies (i.e. RFS, CAFE,
AMFA incentives) are applied from now to 2025 as required by law or
as announced. We estimate the effects those policies can have over
the baseline period by comparing the baseline to counterfactual
scenarios in which the CAFE and AFV policies are modied.
This study is valuable because it sheds light on an often
overlooked set of policies as they relate to the biofuel industry
and biofuel policy. The CAFE standards are set to become much
more stringent over the next decade and beyond, and the AMFA
credits for exible-fuel vehicles (FFVs) capable of using high-level
blends of gasoline and ethanol such as E85 are to be phased out by
2020 under current policy. This research provides information
regarding these policy effects that is both timely and relevant to
policymakers and other interested stakeholders in the transportation and biofuels industries.

1




42 mpg:
0:5  1=25 0:5  0:15=19

1.1. Background
In this section, we provide a brief overview of the CAFE
standards as well as the AMFA credits. Then, we highlight some
of the previous studies that have focused on these policies, their
effects in the transportation sector, and, in the case of a few
particularly germane cases, the relationship between these policies and other biofuel-related policies.
The Energy Policy and Conservation Act (1975) (EPCA), passed
in 1975 (PL 94-163), contained the original set of CAFE standards.
In the wake of the rst major oil price shock in the U.S., the
standards were viewed as a stepping stone toward domestic
energy security in addition to other environmental goals. The
EPCA called for new, light-duty passenger vehicles to achieve an
average fuel economy of 18 miles per gallon (mpg) by 1978.
Beginning in 1982, light-duty cars and light-duty trucks were
treated separately. Each type of vehicle had its own fuel economy
targets to reach, and the targets for light-duty trucks have
continued to be less stringent. The required standards for model
year 2011 vehicles were approximately 30 and 24 mpg for passenger cars and light-duty trucks, respectively (National Highway
Transportation Safety Administration, 2012a). The current standards are attribute-based and vary according to the size of the
vehicle footprint. NHTSA estimates the standards in 2017 and
2025 to be 40 and 56 mpg for passenger cars and 29 and 40 mpg
for light-duty trucks based on their projections of vehicle production volume by type (Environmental Protection Agency, National
Highway Transportation Safety Administration, 2012). Manufacturers that fail to comply with the CAFE standards face a ne that
is a function of both the number of vehicles they sell and the
margin by which the manufacturer fails to meet the standard. The
original ne in EPCA was $50 per vehicle sold from that model
year per mpg below the standard, but in 1997 that amount was
raised to $55 (U.S. Government Accountability Ofce, 2007).
The Alternative Motor Fuel Act (1988) (AMFA), passed in 1988,
set the stage for many of the biofuel policies currently in place
(PL 100-494). AMFA created incentives for automobile manufacturers to produce AFVs that would, in turn, lead to the widespread
adoption of alternative fuels. The primary incentive is the formula
for calculating the fuel economy of AFVs. For example, the fuel
economy for an FFV using E85, MPGFFV, would be calculated as:
MPGFFV 

1
0:5 

denominator. The equation applies an additional weight of 0.15

to MPGE85. The implication is that the fuel economy of FFVs is
calculated exclusively on the gasoline content (assumed to be 15%
by volume) of E85.
To see the effect this incentive has on calculated MPGs for CAFE
compliance, one can compare the calculated MPG values of two
vehicles, one gasoline-only and the other ex-fuel capable, with
identical fuel economies. For the sake of the illustration, assume both
vehicles achieve 25 mpg using gasoline only. However, one gallon of
E85 has approximately 25% less energy than a gallon of gasoline, so
fuel economy of the FFV using E85 would be reduced by about 25%
relative to its gasoline fuel economy. In this case, the FFVs fuel
economy for E85 would be approximately 19 mpg (i.e. 25 mpg  0.75).
For CAFE compliance, the fuel economy of the gasoline vehicle would
be 25 mpg, but the calculated MPGFFV using Eq. (1), would be

1
MPGgas

0:15
0:5  MPG
E85

where MPGgas is the fuel economy when using gasoline fuel (with
no ethanol) and MPGE85 is the fuel economy when using E85. The
formula makes two key assumptions. First, it assumes that the
vehicle will use E85 half of the time and gasoline the other half.
Hence, each fuel type is weighted by a factor of 0.5 in the

The manufacturers overall CAFE is a share-weighted harmonic

average of all model lines, so the positive effect of the FFVs sold is
muted somewhat. Manufacturers can claim a maximum credit
resulting from FFVs sold of 1.2 mpg to apply toward their overall
CAFE requirement for the model year.
Of these two policies, the fuel efciency targets seem to have
garnered more attention in the academic literature. There are
numerous studies that investigate the impact of CAFE standards on
social welfare. At face value, the ambition of CAFE standards is to
improve social welfare through reduced fuel consumption and
greenhouse gas emissions. The actual welfare effects, however, are
debated. Some studies have reached the conclusion that tighter
CAFE standards result in reduced social welfare, mostly at the
expense of consumers (Austin and Dinan, 2005; Kleit, 2004; Small,
2012). However, other studies have reached the opposite conclusion that more stringent CAFE standards increase social welfare
under certain circumstances (Fischer et al., 2007). Conclusions of
reduced social welfare might stem from the unintended consequences of the CAFE standards that run counter to the original
goals. The so-called rebound effect of CAFE standards is a prime
example. The rebound effect is dened as the fraction of fuel
savings expected to result from an increase in fuel efciency
that is offset by additional vehicle use (National Highway
Transportation Safety Administration, 2010 p. 364). In other
words, CAFE standards that reduce fuel consumption through
better vehicle fuel economy also reduce the cost of driving on a
per-mile basis. That, in turn, increases the demand for vehicle
travel which results in a smaller reduction in overall fuel consumption. Other welfare effects might arise through changes in
trafc safety, congestion costs, vehicle prices, and air pollution
(Parry et al., 2007; Portney et al., 2003).
The rebound effect, as it relates directly to overall fuel consumption, is also relevant in a structural model such as the one
developed in this study. There are several studies that have
examined the effect more closely. In light of the denition above,
the rebound effect is measured as the elasticity of vehicle miles
traveled (VMT) with respect to driving cost per mile multiplied by
a factor of  1 (Small and Dender, 2007). The rebound effect is,
thus, expressed as the percent increase in VMT for a given decrease
in driving cost per mile. Most empirical studies have estimated
rebound effects in the range of 10% to 30% (Hymel et al., 2010;
National Highway Transportation Safety Administration, 2010;
Small and Dender, 2007).
CAFE standards, the rebound effect, the exibility provisions
provided by AMFA, and biofuel policies such as the RFS are all
interrelated. The interactions between them have received only
minimal attention in the literature to this point. An important nding
from this line of study is that the interactions among the policies

Please cite this article as: Whistance, J., Thompson, W., The role of CAFE standards and alternative-fuel vehicle production credits in U.S.
biofuels markets. Energy Policy (2014), http://dx.doi.org/10.1016/j.enpol.2014.09.004i

J. Whistance, W. Thompson / Energy Policy ()

might tend to reduce their overall effectiveness or increase their cost.

The AMFA credits that can be applied toward meeting the CAFE
standards have resulted in more AFVs being produced, but they
might limit the effectiveness of CAFE standards by allowing auto
manufacturers to produce vehicles that are more fuel inefcient than
the standards would otherwise allow (Collantes, 2008; Liu and
Helfand, 2009). However, those inefciencies may make the RFS
requirements easier to meet if overall motor fuel use becomes large
enough. Furthermore, a larger eet of FFVs might make the RFS
requirements easier to meet, but only if the larger eet translates to
more consumption of E85. Historically, distribution, pricing, and
consumer awareness issues have limited the expansion of E85 use.
On the other hand, tighter CAFE standards could reduce overall
fuel consumption and make it more difcult to meet the RFS
requirements unless there is an increase in the adoption of highlevel ethanol blends or, perhaps, drop-in replacements for
petroleum based fuels (Westbrook et al., 2014). This arises from
the fact that, within a compliance year, the RFS is a volume
requirement and can be sensitive to changes in fuel consumption.
At the same time, the AMFA credits are being phased out as part of
EISA 2007 and with them go one incentive to produce FFVs
capable of using high-level ethanol blends. These two effects are
important for the problem of getting through the blend wall, the
anticipated inelastic portion of the ethanol demand curve when
the E10 market is saturated and further expansion requires low
enough prices to coax consumers to buy high-blend fuels. The RFS
became more binding than ever before in early 2013 because the
E10 market was saturated by that point, with ethanol use already
equal to 10% of total nished motor gasoline use, and rising biofuel
mandates suggest that more ethanol will have to be used in highblend fuels in the future (Thompson and Whistance, 2013). The
challenges of meeting a greater mandated volume might only be
exacerbated by tighter CAFE standards that reduce total fuel use
and reduced AMFA credits that lead to fewer FFVs.
There is an uncertainty, however, surrounding future RFS
requirements. The proposed requirements for 2014 are based on
expected levels of fuel use. If that method is adopted, the relationship between CAFE and RFS could become more complementary.
For example, future RFS requirements might be set with CAFE
compliance already taken into account under the proposed
method. In that case, more stringent CAFE standards might not
make the RFS more difcult to meet, even if it decreased overall
ethanol use. Furthermore, the requirements for CAFE and RFS
focus on different segments of the transportation sector. CAFE
primarily targets light-duty vehicles, although requirements for
medium- and heavy-duty vehicles have been proposed, while RFS
is concerned with over-the-road transportation by all three vehicle
classes. Thus, changes to CAFE may not affect how non-light duty
vehicles contribute toward the RFS, at least for the time being.
There is also uncertainty related to the technological improvements automakers might use to meet CAFE requirements. For
example, gasoline engines could be redesigned to be more fuel
efcient by making them smaller with higher compression ratios.
Such engines would require higher-octane fuels for optimum
performance. If ethanol continues to be a cheap source of octane
for fuel blenders, mid-level ethanol blends like E30 (30% ethanol,
70% gasoline) might become more popular and, again, the RFS and
CAFE requirements could become more complementary in nature.

2. Methods
2.1. Conceptual framework
The model in this study consists of three related components
(Fig. 1). Two of the components are reduced-form representations of

Petroleum and
petroleum products

Light-duty vehicles

Price and
Quantity
Linkages

Biofuels

Fig. 1. Flowchart summarizing the relationship between the reduced-form versions

of the light-duty vehicles, biofuels and petroleum and petroleum products markets.

a structural model of the U.S. and international petroleum, petroleum

products, and biofuels markets.1 Those models have been used in
prior research to study the greenhouse gas consequences of U.S.
biofuel policies (Thompson et al., 2011b; National Research Council,
2013). The nal element, which represents the light-duty vehicle
(LDV) sector, was developed for this analysis to enrich the petroleum
and biofuels models and create an overall structural model of
petroleum and biofuels markets to estimate the effects of transportation policies, such as CAFE standards and AMFA FFV production
incentives. Similar approaches have been used to estimate the effects
of U.S. biofuel and energy policies on domestic natural gas markets as
well as domestic biomass markets (Whistance, 2012; Whistance and
Thompson, 2010; Whistance et al., 2010).
The core of the LDV model (Fig. 2) resembles the system of
equations developed by Small and Dender (2007). The core
equations estimate the share of light-duty vehicles that consume
diesel, the proportion of light-duty vehicles that can consume E85,
the proportion of light-duty vehicles that consume gasoline, the
average fuel economy of light-duty vehicles, the total miles
traveled by light-duty vehicles, the stock of light-duty vehicles,
and the average real price of a new car. The equations are
represented as follows:
LnDiesel share of LDVs
Max0;
Min0:9;
Intercept
b1  LnReal cost of driving gasoline
vehicle per mile
b2  LnReal cost of driving diesel
vehicle per mile
b3  LnLagDiesel share
error term

LnFFV share of LDVs

Max0;
Min0:9;
Intercept
b1  LnReal cost of driving gasoline
1
For a more complete representation of the model including estimated
parameters and elasticities, please see the Supplemental data section.

Please cite this article as: Whistance, J., Thompson, W., The role of CAFE standards and alternative-fuel vehicle production credits in U.S.
biofuels markets. Energy Policy (2014), http://dx.doi.org/10.1016/j.enpol.2014.09.004i

J. Whistance, W. Thompson / Energy Policy ()

b3  LnReal consumer expenditures

b4  LnLagTotal vehicle miles traveled
by LDVs
b5  Linear time trend
error term

Outputs from Petroleum and petroleum products

model

LDV fleet fuel

economy

New vehicle
price

LDV stock

Average real priceof new car

Intercept
b2  4  year moving average of CAFE


standards; Car=Truck combined

LDV miles
traveled

b3  Real consumer expenditures

b4  Linear time trend
error term:

Total fuel
demand

Inputs to Petroleum and petroleum products

model

Inputs to Biofuel model

Fig. 2. Flowchart summarizing the relationship between endogenous elements of

the light-duty vehicles model and inputs to and outputs from the biofuels and
petroleum and petroleum products markets.

vehicle per mile

b2  LnReal cost of driving
FFV permile
b3  Real value of potential
CAFE finesavoided per FFV sold
b4  LnLagFFV share
error term
Gasoline share of LDVs
1  FFV share  Diesel shareOther share

3
4

LnAverage fuel economy of LDV fleet

Intercept
b1  LnReal gasoline price
b2  LnTotal vehicle miles traveled by LDVs
b3  LnReal consumer expenditures
b4  6  year moving average of CAFE
standards; Car=Truck combined
b5  Linear time trend
error term
5
LnTotal stock of LDVs
Intercept
b1  Ln Real gasoline price=Average fuel
economy of LDV fleet
b2  LnReal price of new vehicle
b3  LnReal consumer expenditures
b4  LnLagTotal stock ofLDVs
b5  Linear time trend
error term
LnTotal vehicle milestraveled by LDVs
Intercept
b1  Ln Real gasoline price=Average fuel
economy of LDV fleet
b2  LnTotal stock of LDVs

The proportions of diesel and FFVs are estimated primarily as

functions of own- and cross-costs of driving per mile. The equation
for FFV share of LDVs also includes the effect of avoided CAFE
penalties that arise through the AMFA credit. The proportion of
gasoline vehicles is a residual category that is also adjusted for an
exogenous share of light-duty vehicles that include electric, fuel
cell, and other alternative technologies.
The eet-average fuel economy is a function of gasoline prices,
vehicle miles traveled, income (as measured by consumer expenditures), the CAFE requirements, and a linear time trend that
captures the effect of changes in the mix of vehicle ages, sizes, and
technology over time. In this model, the CAFE requirements are
represented by the combined CAFE standard based on NHTSA
estimates2 in the nal rules for 20122025 (Environmental Protection Agency, National Highway Transportation Safety Administration, 2012). The total stock of light-duty vehicles is estimated as a
function of the eet-average cost of driving, new vehicle prices,
income, a linear time trend, and lagged values of light-duty vehicle
stock. Vehicle miles traveled are also estimated as a function of the
eet-average cost of driving, income, a linear time trend, and
lagged vehicle miles traveled. In addition, vehicle miles traveled by
LDVs is a function of the total stock of light-duty vehicles. Table 1
summarizes the estimated effects of the independent variables on
the dependent variables in Eqs. (2)(8). Many of the relationships
are estimated in loglog form, which implies constant elasticity
(i.e. percent change in dependent variable, given a 1% increase in
the independent variable). In some cases, log-linear or linear
relationships are estimated when constant elasticity is not
expected (e.g. CAFE standard effects in Eqs. (5) and (8)).
Dividing vehicle miles traveled by average fuel economy and
adjusting for energy content, as in (9), provides an estimate of
overall fuel use3 by light-duty vehicles, which is the primary link
to the models of biofuels and petroleum and petroleum product
markets. To make that connection, LDV fuel use is disaggregated
into the demands for gasoline, diesel, and E85 (Eqs. (9)(14)). This
study assumes the amount of each type of LDV fuel demand is
proportional to the estimated share of that type of vehicle in the
overall stock of LDVs. A limitation of this assumption is that it
would break down when applied to FFVs. E85 is consumed at a
much lower rate than the number of FFVs would indicate. This
model accounts for that difference by estimating the actual use of
E85, not just the potential if all FFVs use E85 all the time. The
2
As the EPA and NHTSA note in their rulemaking, the CAFE standards are
based on vehicle footprints, so actual CAFE obligations will differ as more complete
data are made available. A model that projects detailed vehicle sales by footprint
(or classes thereof) is beyond the scope of this study, so we make the assumption
that the EPAs estimates of CAFE obligations hold true throughout the period and
take them as given.
3
Measurements of fuel economy for CAFE compliance can differ substantially
from on-road fuel economy. The average fuel economy of the entire LDV eet is an
estimate of on-road fuel economy that includes older, less fuel-efcient models.

Please cite this article as: Whistance, J., Thompson, W., The role of CAFE standards and alternative-fuel vehicle production credits in U.S.
biofuels markets. Energy Policy (2014), http://dx.doi.org/10.1016/j.enpol.2014.09.004i

J. Whistance, W. Thompson / Energy Policy ()

Table 1
Elasticities calculated using average values for 201125 time period.
Denition

Elasticity

Term (e.g. with respect to)

Short-run
Diesel share of LDVs
Flex-fuel vehicle share of LDVs

Average fuel economy of LDV eet

Total LDV stock

Total vehicle miles traveled by LDVs

Long-run

 0.01
 0.99
1.45
0.20
0.07
0.75
 0.82
0.34
 0.01
 0.08
0.29
 0.07
0.17

 0.25
 2.68
3.91
0.50

 0.07
 0.52
1.85
 0.23
0.60

remaining portion of fuel consumed by FFVs is then added to lightduty gasoline consumption.
Total fuel use by LDVs
Total vehicle miles traveled by LDVs  5253; 000  1000
Average fuel economy by LDV fleet  42

LDV demand for gasoline

Gasoline share 1 Penetration rate of E85 use  FFV share  LDV fuel use
Energy content of gasoline adjusted for ethanol content over time

10
LDV demand for diesel

Diesel share of FFVs  LDV fuel use  1000; 000

5825; 000

Potential E85 consumption

FFV share  LDV fuel use  1000; 000

3985; 000

11
12

Penetration rate of E85 use

Intercept

b1  Implied retail ethanol price=Retail gasoline price

b2  Max 0; 0:67  Implied retail ethanol price=Retail gasoline price

error term

Actual E85 consumption

Penetration rate of E85 use  Potential E85 consumption:

13

14

The biofuels market is represented by an updated and much

simplied version of the FAPRI-MU biofuels model (Meyer and
Thompson, 2010). Here, the focus is on ethanol while biodiesel
remains exogenous. The estimation of ethanol production differs
from the FAPRI-MU version in that it does not rely on production
capacity and capacity utilization. Rather, an ethanol supply curve is
derived from the output of FAPRI-MUs stochastic baseline results
(Westhoff et al., 2012). For each of the 500 stochastic model
simulations over the baseline period, there are eleven observations
of ethanol production, ethanol input prices, and ethanol output
prices that can be used as data points. Ethanol production is
estimated as a function of the wholesale ethanol (output) price
averaged over the current and previous three years, an index of
natural gas (input) prices, lagged production, and a time trend.
This specication does not include explicitly corn and distillers
grains prices, the other two major input and output prices, but in a
simulation scenario their effects are implicit.
There are three components of fuel ethanol consumption in
this representation. Potential use of ethanol as a fuel additive is
determined primarily by the level of gasoline consumption by the
transportation sector, although ethanol price plays a small role.
The potential markets for both low-level and high-level blends are

Real cost of driving, diesel

Real cost of driving, E85
Real cost of driving, gasoline
Avoided penalty per FFV sold
Real cost of gasoline
Miles traveled
Real income
CAFE standards
Real cost of driving
Real price of new car
Real income
Real cost of driving
Real income

determined by the level of nished motor gasoline consumption,

as well, and the FFV share of light-duty vehicles plays a limiting
factor for E85 potential. The FFV share is determined, in part, by
the value of the AMFA incentive to auto manufacturers (described
below). All else equal, we expect the share of FFVs to increase as
auto manufacturers respond to the incentive. Furthermore, we
expect the increase of FFVs to come at the expense of gasolinepowered vehicles. This does not imply an equally offsetting switch
in fuel consumption. A market penetration rate for each type of
fuel blend determines the actual level of their use. The penetration rates depend on the relative prices of ethanol and gasoline4
(Eq. (13)), and it is structured so that the penetration rate becomes
more elastic once E85 is sold at a slight discount to gasoline in
energy equivalent terms. The discount necessary to motivate such
a response and the magnitude of the response are sources of
uncertainty, in part, because of heterogeneity among consumers
willingness to pay (Anderson, 2012).To the extent that the E85
penetration rate is less than 100%, an increase in the FFV share will
not result in a proportionate increase in actual E85 use.
In terms of both ethanol imports and exports, data from the
U.S. Energy Information Administration (2012) (EIA) indicate that
Brazil has historically been a main trading partner for the U.S. (U.S.
Energy Information Adiministration, 2014) The U.S. imports
sugarcane-based ethanol from Brazil in order to help meet the
part of the overall RFS requirement that is reserved for advanced
biofuels. In addition, Californias Low Carbon Fuel Standard could
provide an additional incentive to import sugarcane ethanol. The
U.S. also exports conventional (corn starch) ethanol to Brazil in
order for Brazil to satisfy its own demand for ethanol. Although
the Brazilian anhydrous-ethanol price is exogenous to this model,
the relative prices of advanced and conventional ethanol to the
Brazilian price determine U.S. ethanol imports and exports,
respectively.
The RFS requires minimum levels of biofuel use that are
scheduled to rise each year through 2022. The requirements are
comprised of four nested categories: overall, advanced, biomassbased diesel, and cellulosic. To qualify for the overall requirement,
biofuels must show at least a 20% reduction in greenhouse gas
(GHG) emissions relative to the 2005 level of the petroleum-based
fuel they replace. Conventional, or corn starch-based, ethanol
generally falls into this category, and can be used to meet part of
the 36 billion gallon overall RFS requirement by 2022. The
advanced biofuel requirement is nested within, and is less than,

4
The parameters for Eq. (13) were assumed to be the following: b1 (  2) and
b2 25.

Please cite this article as: Whistance, J., Thompson, W., The role of CAFE standards and alternative-fuel vehicle production credits in U.S.
biofuels markets. Energy Policy (2014), http://dx.doi.org/10.1016/j.enpol.2014.09.004i

J. Whistance, W. Thompson / Energy Policy ()

the overall requirement. It can be met with fuels that demonstrate

at least a 50% reduction in GHG emissions, including biomassbased diesel, cellulosic biofuels, and sugarcane ethanol from Brazil.
Nested within the advanced requirement are the biomass-based
diesel and cellulosic biofuel requirements. Eligible biomass-based
diesel must demonstrate at least a 50% reduction in GHG emissions, and eligible cellulosic biofuel must show GHG emission
reductions of at least 60%.
Renewable identication numbers (RINs) are tradable credits
used by obligated parties to show compliance with the RFS
requirements. The FAPRI-MU biofuels model employs a set of
equations to estimate RIN supply and demand. Those equations
have been the basis for several previous studies that have
examined the RIN markets more closely (Thompson et al., 2009
2011a). The present model employs the same general structure of
the RIN markets, but biodiesel and cellulosic RINs are held
exogenous (so their respective sub-mandates are omitted) and
RIN stocks are ignored. The two elements of the mandate that are
endogenous are (a) compliance with both the overall mandate for
all qualifying biofuels, including conventional ethanol, and (b) the
sub-mandate for advanced biofuels that includes imported sugarcane ethanol that meets a higher GHG reduction target. Here, the
overall and advanced RFS requirements are modeled as perfectly
inelastic lower bounds on domestic biofuel use. In practice, each
obligated party faces a requirement that is a percentage of their
expected production for the compliance year. The present model
assumes the minimum level of use is met, in aggregate. A technical
point that is retained is that the advanced mandate is a submandate, so extra advanced RINs generated with advanced biofuel
use beyond its mandate can be used to help to meet the overall
mandate, but the reverse is not true.
RIN prices are an indicator of the degree to which the RFS
mandates are binding. If market conditions are such that the
equilibrium demand for ethanol in absence of the mandate is less
than the mandated volume, then the mandate is considered
binding. In this context, binding implies that obligated parties
must use more ethanol than they would choose otherwise in order
to meet the minimum use requirement. To obtain that quantity,
fuel blenders must pay a higher price to ethanol producers, and to
sell that quantity, blenders must charge a lower ethanol price to
retail customers. The difference between the two prices is the
core RIN value that estimates mandate compliance costs per
gallon, excluding transaction costs (Thompson et al., 2010). In that
case, RINs become more valuable as they are needed for compliance, so the market for RINs becomes tight, and their prices rise
(sometimes dramatically as in the case of 2013). If the opposite is
true and equilibrium ethanol demand is higher than the RFS
requirement, the mandate is considered non-binding and the core
RIN value is zero. These possibilities are captured in the representation of the mandate used here. The RIN values are added to
the value of ethanol for transportation fuels to estimate the
wholesale prices that, in turn, drive domestic and import supplies
of conventional and advanced ethanol, respectively.
The LDV and biofuel models are linked to reduced-form
versions of the gasoline and diesel market models based on
Thompson et al. (2011b), with some updates. Within this framework, transportation demands for gasoline and distillate fuels are
disaggregated into demand by LDVs, which is determined in the
LDV model, and demand by other vehicles. Biofuel prices and
quantities also play a role in petroleum product markets. Transportation demands for gasoline and distillate fuel include biofuel quantities on an energy equivalent basis, and their prices
faced by consumers are share weighted averages of the biofuel and
petroleum product prices, plus average federal and state level fuel
taxes and the cost of mandate compliance (Thompson et al.,
2011b).

2.2. Data
Many of the historical data cover the period from 1970 to 2010.
Data regarding LDVs including fuel economy, vehicle miles traveled, and vehicle stock were obtained from the Highway Statistics
series published by the Federal Highway Administration (U.S.
Federal Highway Administration, 2012).
The average fuel economy of the light-duty vehicle eet
increased only slightly before the CAFE standards were enacted.
After that, fuel economy rose steadily along with the CAFE
requirements until the early 1990s. Along with the policy requirements, high oil prices and the high cost of driving, as a result, also
contributed to the desire for greater fuel economy. The growth in
average LDV fuel economy slowed in the late 1980s and early
1990s as the CAFE standards plateaued and cheaper oil made
driving less costly. Fuel use increased in spite of higher fuel
economy as drivers traveled more. The rebound effect from higher
fuel economy might have played some role in increased travel, but
the low cost of driving, more total drivers, and larger incomes also
played important roles.
We calculate the average value of an FFV to automakers
associated with the AMFA provisions. Automobile manufacturers
in violation of CAFE standards must pay a penalty per vehicle sold
per mpg they fall short of the standard. The penalty increased
from $50 to $55 in 2001. The AMFA provisions allow automakers to
reduce the potential nes they face by increasing their calculated
CAFE up to 1.2 mpg through the production of FFVs. Thus, the
1.2 mpg credit has a value to auto manufacturers that can be
expressed in terms of the penalty avoided per FFV sold. Various
reports by the NHTSA and a report by the Departments of
Transportation and Energy as well as the EPA provide the size of
the credit, if any, claimed by automakers from 1993 to 2011
(National Highway Transportation Safety Administration, 2001,
2002, 2012b; U.S. Department of Transportation, U.S. Department
of Energy, U.S. Environmental Protection Agency, 2002). The gross
monetary value of the credit per FFV sold is calculated each year
from 1995 forward using vehicle sales data of the companies that
claim the credit. In 2010, for example, the four automakers
claiming at least some credit sold a combined 5 million vehicles
in the U.S., of which 1.5 million were FFVs. The average credit
claimed was approximately 1 mpg to be applied toward each
companys calculated CAFE. The total potential nes that could
have been collected had they fallen short of the requirement by
the size of the credit claimed would have been:
$55  1 mpg credit  5 million vehicles sold $275 million
and the gross value per FFV sold would have been:
$275 million=1:5 million FFVs sold $183=FFV sold
We assume additional marginal cost of producing an FFV
relative to a non-FFV is $100 per vehicle. This is the low end of
the $100$200 range used by Anderson and Sallee (2011) in their
estimation, but within the full range of values they cited from the
literature. In addition to the baseline calculation, we use projected
vehicles sales data to estimate the value of the credit if the policy
was extended (Fig. 3).

3. Results and discussion

This study investigates, in a forward looking manner, the
potential market effects of CAFE standards and AMFA credits in
three separate scenarios. The scenarios are as follows:

Constant 2010 CAFE requirements: The car/truck combined CAFE

standard in the projection period, 2011 to 2025, is held xed at

Please cite this article as: Whistance, J., Thompson, W., The role of CAFE standards and alternative-fuel vehicle production credits in U.S.
biofuels markets. Energy Policy (2014), http://dx.doi.org/10.1016/j.enpol.2014.09.004i

J. Whistance, W. Thompson / Energy Policy ()

Fig. 3. Calculated value of AMFA credit per FFV sold.

Sources: National Highway Transportation Safety Administration; Energy Information Administration.

the 2010 level of 26 mpg. The AMFA credit is phased out as

scheduled.
AMFA credits extended: The AMFA credit remains at 1.2 mpg
throughout the projection period. CAFE standards rise in
accordance with current policy.
Constant 2010 CAFE requirements and AMFA credit extension: A
combination of the two previous shocks.

Each scenario is compared to a baseline5 projection from the

year 2011 to 2025. The baseline assumes current or announced
policies are in place. The CAFE standards for the years 2017 to
2025, which were nalized in 2013, are represented in the baseline. The AMFA credit is phased out by 2020 and the combined
CAFE fuel economy requirement rises to 49 mpg. The RFS is
assumed to be in effect, although this model focuses on the part
of the RFS applicable to conventional ethanol. The EISA allows the
amount of conventional ethanol eligible for the RFS to rise to 15
billion gallons in 2015 and then held the amount constant.
Advanced ethanol is assumed to account for 3.5 billion gallons of
the RFS in 2022 and held constant at that level through 2025.
Biodiesel is assumed to make up another 1.5 billion gallons of
advanced biofuels, in ethanol equivalent terms, and the cellulosic
mandate is assumed for the sake of simplicity to continue to be
waived and production to remain quite small. Our analysis
precedes EPAs recent proposal for the 2014 standards which, if
adopted, would result in lower volumes for ethanol and maintain a
higher volume for biodiesel. We also assume the RFS volumes are
extended to 2025 as current law suggests could happen, although
this assumption ignores the potential for other EPA waivers, the
EPA review required in the near future, and the possibility of
legislated changes. The scenario results are summarized in Table 2
as well as Fig. 4.
In the baseline, the overall and advanced mandates are binding
in all years, as indicated by RIN prices greater than zero. The
baseline projection, including the FFV adoption rate, is calibrated
to the reference case of the EIA 2013 Annual Energy Outlook that
also assumes the CAFE standards for 2017 to 2025 are in effect and
the AMFA credits are phased out as scheduled (U.S. Energy
Information Administration, 2013). As a further check on model
validity, we compare our estimate of the rebound effect using
ordinary least squares (OLS) to the OLS estimate from Small and
Dender (2007). Their estimate of the vehicle miles elasticity with
respect to the real cost of driving per mile was  0.085 (i.e. an 8.5%
effect) in the short run and  0.34 in the long run. The estimates in
this study are  0.07 in the short run and  0.23 in the long run.
5
Selected baseline results are presented in Table A.1 of the Supplementary
materials.

In the rst scenario, CAFE standards in the projection period are

held constant at the 2010 level of 26 mpg in contrast to the
announced increase in the CAFE standard that is applied in the
baseline. The AMFA credit is assumed to be phased out in this
scenario, as planned. The estimated changes that occur relative to
the baseline are summarized for the 15-year projection period
(Table 2). All else equal, this assumption reduces the incentive for
automakers to pursue gains in fuel economy. In the absence of
CAFE standards that grow over time, the growth in average fuel
economy slows and more fuel is required by light-duty vehicles.
The demand curves for gasoline, ethanol, and diesel shift out
(i.e. to the right), and their equilibrium prices increase. The
combination of lower average fuel economy and higher fuel prices
increases the cost of driving on a per mile basis. As a result, drivers
tend to travel fewer miles. Lower CAFE standards are easier for
auto manufacturers to achieve, and the lower vehicle production
costs are represented by lower nominal prices for new cars. As
vehicles are more affordable relative to the baseline, there is a
slight increase in stock of LDVs. The relative price changes, in
conjunction with the lower fuel economy of FFVs relative to
gasoline-powered vehicles, also result in a shift in the mix of
light-duty vehicles away from FFVs and toward gasoline-powered
vehicles.
The interaction of CAFE standards with the RFS requirements
occurs primarily through the shift in the blend wall. As fuel
demand shifts out in the rst scenario relative to the baseline,
there is more room for E10 use to expand and help meet the rising
RFS requirements. Although the penetration rate and use of E85
decrease, the expansion in the E10 market is enough to make the
overall RFS mandate non-binding. Ethanol prices rise and, in
response, ethanol production increases and ethanol exports
decrease. Conventional RIN prices fall to zero as blenders have
no trouble obtaining and submitting the requisite number of RINs
to meet the total RFS mandate in each of the compliance years. The
lower prices for advanced RINs show that the RFS requirement for
advanced biofuels becomes less binding in this scenario, as well,
due to greater ethanol demand. The cost of complying with the
total mandate (i.e. RIN prices multiplied by the amount required
for compliance) falls along with the falling RIN prices. Although
the compliance costs are reduced, the partial sum of related fuel
and vehicle costs increases relative to the baseline because the
increase in consumer fuel expenditures is greater than the reduction in vehicle expenditures.
In the second scenario, the CAFE standard rises in line with
current policy, as in the baseline, but the AMFA credit is no longer
phased out. Beyond 2014, automakers are allowed to continue
claiming a credit of up to 1.2 mpg to apply toward meeting the
CAFE requirements. The extension of the AMFA credits effectively
increases the potential nes automakers can avoid by producing
and selling FFVs. The production of FFVs becomes more desirable
and the FFV share of the total light-duty vehicle stock increases
slightly at the expense of gasoline-powered vehicles (Table 2).
Because of the greater number of FFVs with the AMFA credit
extension, the gasoline demand curve for light-duty vehicles is
shifted back (i.e. to the left).
The expansion of FFVs increases the potential for E85 consumption and eases the blend wall to a small extent, making the
RFS requirement less binding. The small relative price movements of ethanol and gasoline imply the penetration rate of E85
remains virtually unchanged. E85 consumption increases a little.
E10 use declines slightly, but it does not quite offset the increase
in E85 consumption. Domestic use of ethanol increases very
slightly, and the lower prices for both conventional and advanced
RINs indicate the RFS mandate is less binding than in the
baseline. As obligated parties face lower compliance costs, the
savings are assumed to be passed on to gasoline consumers

Please cite this article as: Whistance, J., Thompson, W., The role of CAFE standards and alternative-fuel vehicle production credits in U.S.
biofuels markets. Energy Policy (2014), http://dx.doi.org/10.1016/j.enpol.2014.09.004i

J. Whistance, W. Thompson / Energy Policy ()

Fig. 4. Selected scenario results compared to each other and the baseline.

in the form of slightly lower retail prices than might occur

otherwise.
The nal scenario combines the shocks of the two previous
scenarios, so the CAFE standards are held at 26 mpg and the
AMFA credits are extended. In a way, the two shocks supplement
one another and reinforce certain effects. For example, the
combined shocks allow for increases in both gasoline and E85
demand resulting in higher prices for gasoline and ethanol. The
RFS requirements are even less binding, although the increased
fuel expenditures still lead to an increase in related costs relative
to the baseline.
3.1. Sensitivity analysis
A key uncertainty in this study is the responsiveness of automakers to the AMFA credits. Although there are many other
sources of uncertainty in these markets, the impact of AMFA
credits might be the most relevant given that they affect how
the elimination of this credit affects FFV share. The credits give
automakers an incentive to produce more FFVs, so the marginal
effect of AMFA credits on FFV share is expected to be positive. The
magnitude of that effect is less clear. There have been many
studies that investigated the relationship between vehicle choice
at both the producer and consumer level, particularly as it relates
to fuel economy and CAFE standards. Examples of such studies
include Goldberg (1998), Turrentine and Kurani (2007), and Shiau

et al. (2009). However, there have been fewer empirical studies

that examined those behaviors in light of the AMFA credits.
Examples include Rubin and Leiby (2000) and Liu and Helfand
(2009, 2012) to name just a few. One such study estimated that
AMFA credits would be responsible for one-half of the alternativefueled vehicles sold in the following decade (Rubin and Leiby,
2000). Given the extent to which the biofuel and auto industries
have changed since that paper was published, it is unclear
whether or not the same relationship would still hold. Moreover,
the present study focuses on the share of FFVs within the lightduty vehicle stock. The relationship found by Rubin and Leiby is
somewhat less applicable to the share representation as vehicle
sales and scrappage are implicit. 6
As the model was being developed, the marginal effect of AMFA
credits on FFV share was uncertain. Data availability was limited,
and the small sample made it difcult to identify and correct for
the potential bias. In absence of a better estimate, the baseline
coefcient in the FFV share equation was assumed to be 0.002 for
an elasticity of approximately 0.2.7 In other words, a $10 increase
in the potential penalty avoided per FFV sold would increase the

This list is not intended to be exhaustive.

The FFV share equation is estimated in log-level form with respect to an
independent variable representing the monetary value of potentially avoided CAFE
nes per FFV sold and adjusted for ination. The marginal effect is calculated as %
y 100  (exp(0.002)  1) 0.2.
7

Please cite this article as: Whistance, J., Thompson, W., The role of CAFE standards and alternative-fuel vehicle production credits in U.S.
biofuels markets. Energy Policy (2014), http://dx.doi.org/10.1016/j.enpol.2014.09.004i

J. Whistance, W. Thompson / Energy Policy ()

Table 2
Scenario results, average changes relative to baseline.

Fuel economy of light-duty vehicles, miles per gallon

Miles traveled by light-duty vehicles, billion miles
Light-duty vehicles, million vehicles
Gasoline share of light-duty vehicles, percentage points
Flex-fuel share of light-duty vehicles, percentage points

CAFE constant at 26 mpg

(20112025 average)

Extend AMFA credit

(20152025 average)

Combined scenario
(20152025 average)

Level

Level

Level

Percent (%)

Percent (%)

Percent (%)

 2.3
 81.9
2.5

9.8
 2.9
1.1

0.0
1.3
0.0

0.0
0.1
0.0

 3.0
 110.5
3.2

 12.4
 3.8
1.4

0.9
 0.9

1.0
 14.5

 1.3
1.3

 1.0
20.0

 1.1
1.1

 1.3
16.1

6.2
7.6

32.9
23.7

0.0
0.0

Gasoline use by transportation sectora, million gallons

Gasoline use by light duty vehicles, million gallons

8380
8857

Total domestic ethanol use, million gallons

E85 use, million gallons

385.7
 477.1

2.7
 31.2

0.1
9.2

0.0
0.6

911.7
 134.2

6.1
 6.7

0.35
0.64
 1599

10.0
26.4
 6.1

 0.01
0.19
0

0.0
7.9
0.0

0.47
0.96
 2054

13.6
41.8
 7.7

 0.53
 0.67
 9.4

 99.4
 64.4
 95.1

 0.22
 0.22
 4.1

 41.8
 21.5
 41.3

 0.64
 0.95
 11.8

 100.0
 81.0
 97.3

Retail unleaded gasoline price, $/gallon

Implied retail ethanol price (conventional), $/gallon
New car price (nominal), dollars
Conventional RIN price, $/RIN
Advance RIN price, $/RIN
RFS compliance cost (ethanol portion), billion dollars
Gasoline expendituresb, billion dollars
Ethanol expenditures, billion dollars
Vehicle expendituresc, billion dollars
Sum of cost estimates, billion dollars
a
b
c

81.5
10.4
 12.8
79.1

18.3
29.7
 3.1
8.8

0.3
2.7
0.1
3.2

0.1
7.9
0.0
0.4

10,836
10,970

104.1
17.2
 40.3
81.1

8.1
9.7

23.3
50.5
 9.1
8.8

For consistency, gasoline use by transportation sector matches the EIA denition which includes all fuel ethanol.
We assume the ethanol portion of RFS compliance costs is passed on to all gasoline consumers as part of the retail gasoline price.
Assumes 6% of LDV stock turns over each year.

share of FFVs by 2.0%.8 This is a key assumption in this model, and

the rest of this section discusses how the results of an AMFA credit
extension might differ if the FFV share with respect to potential
penalty avoided is more or less responsive to the credit value than
in the baseline case.
The sensitivity analysis tested two alternative elasticity
assumptions: one alternative to test a much lower elasticity; the
other to test a much higher elasticity. The elasticity in the rst test,
relative to the baseline, was assumed to be lower by a factor of 10,
so a $10 increase in the potential penalty avoided would increase
FFV share by 0.2%. In the second test, the elasticity was assumed to
be higher by a factor of 10 relative to the baseline, so a $10 increase
in the potential penalty avoided would increase FFV share by 20%.
In each case, a baseline projection was estimated in which AMFA
credits were phased out. Each model was then run with the AMFA
credit extension in place. The results comparing the differences of
each scenario relative to their respective baselines are summarized
for the 2015 to 2025 projection period (Table 3).
The effects of extending the AMFA credits are muted in the
low-elasticity scenario. Automakers respond only slightly to the
increased incentive to produce FFVs. The response is strong
enough to induce a very small increase in E85 consumption, which
eases the blend wall problem only slightly. On average, the
conventional RIN price falls by about $0.03 which equates to
average yearly savings of approximately $500 million per year in
mandate compliance costs.
If the elasticity is higher than in the base case, the effects of
extending the AMFA credits become more pronounced. The share
of FFVs increases by a larger amount, on average, in response to
the AMFA incentives. The use of ethanol in the form of both lowand high-level blends increases enough to render the overall RFS
8
The share of FFVs is already measured in percentage terms (e.g. 10% of all
LDVs), so a $10 increase in the avoided penalty would not increase the share of FFVs
by 2 percentage points (i.e. an increase from 10% to 12%). Rather, it would increase
the percent share by 2% (i.e. an increase from 10.0% to 10.2%).

requirement much less binding, with the conventional RIN price

falling by an average of $0.56. Mandate compliance costs are
substantially reduced during this period, as well.
3.2. Limitations
The results of this study are relevant in that they shed light on
an often overlooked set of policy relationships. Although the
results indicate the CAFE standards could have more of an impact
on the degree to which the RFS requirements are binding than the
AMFA credits, there are uncertainties that still exist, related to onthe-road fuel economy, the proposed RFS rules going forward,
and the response of automakers to the AMFA credit. For the
purposes of CAFE compliance, vehicle fuel economy is measured
with a controlled, standard testing procedure. These are theoretical measures of fuel economy, but the actual fuel economy can be
as much as 20% lower in real-world driving environments (Environmental Protection Agency, National Highway Transportation
Safety Administration, 2012). In that case, the link between CAFE
standards and RFS compliance may be somewhat weaker than
presented above because actual fuel consumption might be
declining less than strict CAFE compliance would indicate.
Uncertainty also surrounds the RFS requirements beyond 2014.
The proposed rules indicate volumes that are based on expected
levels of biofuel use. Reduced fuel consumption through CAFE
compliance would, in theory, be taken into account when setting
the RFS requirement for a given year. If the proposed rules are
adopted, CAFE compliance might not make the RFS more difcult
to meet.
The response of automakers to the AMFA credit presents
another key uncertainty. The literature provides little indication
of the appropriate magnitude of that elasticity, but the results of
the sensitivity analysis show that the response matters. In addition, the model in this study takes a somewhat simplied view of
biofuel, RIN, gasoline, and diesel markets. Biodiesel and cellulosic
biofuels are not included in this study, and exibility provisions in

Please cite this article as: Whistance, J., Thompson, W., The role of CAFE standards and alternative-fuel vehicle production credits in U.S.
biofuels markets. Energy Policy (2014), http://dx.doi.org/10.1016/j.enpol.2014.09.004i

J. Whistance, W. Thompson / Energy Policy ()

10

Table 3
Sensitivity of AMFA credit effects, 20152025 average changes relative to the baseline.
Low ( 0.02)

Base ( 0.2)

 0.2
0.2

 1.3
1.3

 6.6
6.6

Gasoline use by transportation sectora, million gallons

Gasoline use by light duty vehicles, million gallons

4.2
3.0

32.9
23.7

112.7
 414.1

Total domestic ethanol use, million gallons

E85 use, million gallons

0.0
1.2

0.1
9.2

426.8
526.3

Retail unleaded gasoline price, $/gallon

Implied retail ethanol price (conventional), $/gallon

0.00
0.02

 0.01
0.19

0.00
0.67

 0.03
 0.5

 0.22
 4.1

 0.56
 10.4

Gasoline share of light-duty vehicles, percentage points

Flex-fuel share of light-duty vehicles, percentage points

Conventional RIN price, $/RIN

RFS compliance cost (ethanol portion), billion dollars
a

High ( 2.0)

For consistency, gasoline use by transportation sector matches the EIA denition which includes all fuel ethanol.

the RFS also are not considered. The resolution of these uncertainties remains an important extension along this line of research.

4. Conclusions and policy implications

The ndings suggest there are some notable relationships
between transportation and biofuel policies. The CAFE standards,
as they currently are written, will play a large role in determining
the amount of fuel consumed by light-duty vehicles in the future.
Vehicles with better fuel economy will tend to consume less fuel,
both renewable and non-renewable, even after the rebound effect
is taken into account. At the same time, the RFS might require
increasing amounts of renewable fuels to be used. In the baseline
projection, in which the CAFE standards and RFS mandates are in
full effect, the requirements of both policies must be met. The RINs
and mandate compliance costs (i.e. RIN prices multiplied by the
amount required for compliance) indicate the costs of meeting the
RFS in this context. Holding the CAFE standards constant at the
2010 level appears to relax the blend wall constraint, looking
forward, as fuel use expands relative to the baseline. The RFS
requirements would be easier to achieve, in that case, as indicated
by a $9.4 billion reduction in average mandate compliance costs.
Gasoline and ethanol fuel expenditures, which include RFS compliance costs, increase by nearly $90 billion on average.
Alternative-fueled vehicles, including ex-fuel vehicles capable
of using E85, currently receive favorable treatment in the calculation of their fuel economy. Automakers that produce FFVs can take
advantage of that favorable treatment by claiming AMFA credits to
help meet the CAFE standards, subject to a 1.2 mpg limit that
declines to zero by 2020 under current policy. AMFA credits
incentivize the production of FFVs, to some extent, by reducing
the potential nes an auto manufacturer faces if it fails to meet the
CAFE requirement. Under the assumed elasticity, an extension of
the maximum AMFA credit limit of 1.2 mpg through the projection
period makes it somewhat easier to meet the RFS requirements as
indicated by an estimated $4.1 billion decline in average compliance costs. However, the effects are modest and compliance costs
are reduced moderately when compared to the baseline. However,
the exact effect of the AMFA on FFVs marketed is uncertain. The
results of this study indicate that the AMFA credit could reduce
RFS compliance costs by as little as $0.5 billion if producer
sensitivity to the credit is low, or the reduction could be as large
as $10.4 billion if producer sensitivity is high.
This study nds that the changes to the CAFE standard make the
legislated targets of the RFS more difcult to meet. By reducing fuel
use and eliminating the credit for ex-fuel vehicle production, the
new standards reduce the potential market for E85, making it costly

to expand ethanol use. While CAFE standards are estimated to lower

overall expenditures on vehicles and fuel, including on biofuels, by
about $80 billion per year, the cost of RFS compliance rises by nearly
$12 billion, on average. However, this study is not all-inclusive, and
in particular questions relating to fuel and vehicle use, such as
greenhouse gas emissions, congestion, energy security, and road
maintenance, are not addressed. While the ndings are relevant for
decision making and policy assessment, they are not intended to
promote or argue against any particular policy.

Acknowledgments
This material is based upon work supported by the National
Institute of Food and Agriculture and the Ofce of the Chief
Economist, U.S. Department of Agriculture, under Agreement no.
and 58-0111-13-002. Any opinions, ndings, conclusions, or
recommendations expressed in this publication are those of the
authors and do not necessarily reect the view of the U.S.
Department of Agriculture.

Appendix A. Supporting information

Supplementary data associated with this article can be found in
the online version at http://dx.doi.org/10.1016/j.enpol.2014.09.004.
References
Alternative Motor Fuel Act, 1988. Public Law 100-494.
Anderson, S.T., 2012. The demand for ethanol as a gasoline substitute. J. Environ.
Econ. Manage. 63, 151168.
Anderson, S.T., Sallee, J.M., 2011. Using loopholes to reveal the marginal cost of
regulation: the case of fuel-economy standards. Am. Econ. Rev. 101, 13751409.
Austin, D., Dinan, T., 2005. Clearing the air: the costs and consequences of higher
CAFE standards and increased gasoline taxes. J. Environ. Econ. Manage. 50,
562582.
Collantes, G., 2008. Biofuels and the Corporate Average Fuel Economy Program: The
Staute, Policy Issues, and Alternatives. Belfer Center for Science and International Affairs. Harvard Kennedy School.
de Gorter, H., Just, D.R., 2009. The economics of a blend mandate for biofuels. Am.
J. Agric. Econ. 91, 738750.
Du, X., Hayes, D.J., 2009. The impact of ethanol production on U.S. and regional
gasoline markets. Energy Policy 37, 32273234.
Energy Independence and Security Act 2007. Public Law 110-140.
Energy Policy Act, 2005. Public Law 109-58.
Energy Policy and Conservation Act, 1975. Public Law 94-163.
Environmental Protection Agency, National Highway Transportation Safety Administration, 2012. 2017 and Later Model Year Light-Duty Vehicle Greenhouse Gas
Emissions and Corporate Average Fuel Economy Standards; Final Rule. Federal
Register 77, 6262363200.
Fischer, C., Harrington, W., Parry, I.W.H., 2007. Should automobile fuel economy
standards be tightened? Energy J. 28, 129.

Please cite this article as: Whistance, J., Thompson, W., The role of CAFE standards and alternative-fuel vehicle production credits in U.S.
biofuels markets. Energy Policy (2014), http://dx.doi.org/10.1016/j.enpol.2014.09.004i

J. Whistance, W. Thompson / Energy Policy ()

Goldberg, P.K., 1998. The effects of the corporate average fuel efciency standards
in the U.S. J. Ind. Econ. 46, 133.
Hymel, K.M., Small, K.A., Dender, K.V., 2010. Induced demand and rebound effects
in road transport. Transp. Res.: Part B: Methodological 44, 12201241.
Kleit, A.N., 2004. Impacts of long-range increases in the fuel economy (CAFE)
standard. Econ. Inquiry 42, 279294.
Liu, Y., Helfand, G.E., 2009. The Alternative Motor Fuels Act, alternative-fuel
vehicles, and greenhouse gas emissions. Transp. Res.: Part A: Policy Pract.
43, 755764.
Liu, Y., Helfand, G.E., 2012. A hedonic test of the effects of the Alternative Motor
Fuels Act. Transp. Res.: Part A: Policy Pract. 46, 17071715.
Meyer, S., Thompson, W., 2010. FAPRI-MU U.S. Biofuels, Corn Processing, Distillers
Grains, Fats, Switchgrass, and Corn Stover Model Documentation. FAPRI-MU
Report #09-10.
National Highway Transportation Safety Administration, 2001. Automotive Fuel
Economy Program: Annual Update Calendar Year 2001.
National Highway Transportation Safety Administration, 2002. Automotive Fuel
Economy Program: Annual Update Calendar Year 2002.
National Highway Transportation Safety Administration, 2010. Corporate Average
Fuel Economy for MY 2012MY 2016 Passenger Cars and Light Trucks: Final
Regulatory Impact Analysis.
National Highway Transportation Safety Administration, 2012a. Fuel Economy
Performance With and Without AMFA.
National Highway Transportation Safety Administration, 2012b. Summary of Fuel
Economy Performance.
National Research Council, 2013. Effects of U.S. Tax Policy on Greenhouse Gas
Emissions. The National Academies Press.
Parry, I.W.H., Walls, M., Winston, H., 2007. Automobile externalities and policies.
J. Econ. Lit. 45, 373399.
Portney, P.R., Parry, I.W.H., Gruenspecht, H.K., Harrington, W., 2003. Policy watch:
the economics of fuel economy standards. J. Econ. Perspect. 17, 203217.
Rubin, J., Leiby, P., 2000. An analysis of alternative fuel credit provisions of US
automotive fuel economy standards. Energy Policy 28, 589601.
Shiau, C.-S.N., Michalek, J.J., Hendrickson, C.T., 2009. A structural analysis of vehicle
design responses to Corporate Average Fuel Economy policy. Transp. Res.: Part
A: Policy Pract. 43, 814828.
Small, K.A., 2012. Energy policies for passenger motor vehicles. Transp. Res.: Part A:
Policy Pract. 46, 874889.
Small, K.A., Dender, K.V., 2007. Fuel efciency and motor vehicle travel: the
declining rebound effect. Energy J. 28, 25.

11

Thompson, W., Meyer, S., Westhoff, P., 2009. Renewable identication numbers are
the tracking instrument and bellwether of U.S. biofuel mandates. Identikationsnummern fr Erneuerbare Energien als Mittel der Rckverfolgbarkeit und
Indikator fr US-amerikanische Biokraftstoffverpichtungen Les numros
didentication des carburants renouvelables sont les instruments et baromtres de suivi des obligations lgales concernant les biocarburants aux tatsUnis. EuroChoices 8, 4350.
Thompson, W., Meyer, S., Westhoff, P., 2010. The new markets for renewable
identication numbers. Appl. Econ. Perspect. Policy 32, 588603.
Thompson, W., Meyer, S., Westhoff, P., 2011a. What to conclude about biofuel
mandates from evolving prices for renewable identication numbers? Am.
J. Agric. Econ. 93, 481487.
Thompson, W., Whistance, J., Meyer, S., 2011b. Effects of U.S. biofuel policies on U.S.
and world petroleum product markets with consequences for greenhouse gas
emissions. Energy Policy 39, 55095518.
Thompson, W., Whistance, J., 2013. Implied RIN Prices for E85 Expansion and the
Effects of a Steeper Blend Wall. FAPRI-MU Report #03-13.
Turrentine, T.S., Kurani, K.S., 2007. Car buyers and fuel economy? Energy Policy 35,
12131223.
U.S. Department of Transportation, U.S. Department of Energy, U.S. Environmental
Protection Agency, 2002. Effects of the Alternative Motor Fuels Act CAFE
Incentives Policy.
U.S. Energy Information Administration, 2012. Annual Energy Outlook 2012.
U.S. Energy Information Administration, 2013. Annual Energy Outlook 2013.
U.S. Energy Information Adiministration, 2014. Petroleum Supply Monthly (various
issues).
U.S. Federal Highway Administration, 2012. Highway Statistics.
U.S. Government Accountability Ofce, 2007. Vehicle Fuel Economy.
Westbrook, J., Barter, G.E., Manley, D.K., West, T.H., 2014. A parametric analysis of
future ethanol use in the light-duty transportation sector: can the US meet its
Renewable Fuel standard goals without an enforcement mechanism? Energy
Policy 65, 419431.
Westhoff, P., Bineld, J., Gerlt, S., 2012. U.S. Baseline Brieng Book. FAPRI-MU Report
#01-12.
Whistance, J., 2012. Domestic energy policy effects on the US biomass market.
Biomass Bioenergy 46, 133144.
Whistance, J., Thompson, W., 2010. How does increased corn-ethanol production
affect U.S. natural gas prices? Energy Policy 38, 23152325.
Whistance, J., Thompson, W., Meyer, S., 2010. Ethanol policy effects on U.S. natural
gas prices and quantities. Am. Econ. Rev. 100, 178182.

Please cite this article as: Whistance, J., Thompson, W., The role of CAFE standards and alternative-fuel vehicle production credits in U.S.
biofuels markets. Energy Policy (2014), http://dx.doi.org/10.1016/j.enpol.2014.09.004i