TRANSPORTATION
RESEARCH
RECORD
1416
115
Improving
Fuel Economy: A Case
Study
of
the
1992
Honda
Civic
Hatchbacks
JoNATHAN
G.
KooMEY,
DEBORAH
E.
SCHECHTER;
AND
DEBORAH
GORDON
Since the early 1980s, U.S. automobile makers and policy makers
have resisted policies to increase automobile fuel economy, ar-
guing
in
part that such increases were neither technically feasible
nor economically justified. Such assertions for the
1992
Honda
Civic hatchbacks are analyzed. With the
1992
Honda Civic model
line, an automobile maker has, for the first time, produced cars
that are virtually identical to the previous year's models
in
size,
vehicle amenity, engine power, and performance, but that offer
substantially increased fuel economy and improved safety. The
cost of improving fuel economy
is
assessed using actual retail
prices, after correcting for differences in cosmetic features. Cal-
culations indicate that the efficiency of the
1991
Civic DX
was
improved
by
56
percent from 1991to1992 at a cost per conserved
liter of gasoline that
is
$0.20/L ($0.77/gal), or
30
percent less than
the levelized gasoline price without externalities or taxes.
In
ad-
dition, a comparison of two other
Civic
models indicates that fuel
economy
was
improved in the
1992
versions at no additional cost.
Virtually
all
of the efficiency increases described here were achieved
through measures that do not affect safety or vehicle size, such
as engine modifications, transmission alterations, and drag
reduction.
Since
the
early 1980s, U.S. automobile makers and some
analysts
(J) have argued that policies to increase automobile
fuel economy were neither technically feasible nor econom-
ically justified. This
paper
applies Kenneth Boulding's first
law
("anything that exists is possible") to analyze such as-
sertions in the case
of
the 1992
Honda
Civic hatchbacks. With
the new
Hondas,
an automobile maker has, for the first time,
produced cars
that
are virtually identical to the previous year's
models in size, vehicle amenity, engine power, and perfor-
mance, but
that
offer substantially increased fuel economy
and improved safety.
This
paper
[which is a summary
of
a more detailed analysis
contained elsewhere (2)] describes the characteristics
of
the
1991
and 1992
Honda
Civics and demonstrates their equiva-
lence in vehicle amenity.
It
presents the fuel economy tech-
nologies that
Honda
_used
to
improve the efficiency
of
the
Civic by
more
than 50 percent.
It
describes the methodology
for estimating the cost of conserved energy (CCE) for these
efficiency improvements
and
presents the results
of
our
CCE
calculations.
The
paper
concludes by discussing
the
potential
impact
of
gasoline taxes
and
"feebate"
policies
on
both con-
sumer and manufacturer behavior related to energy efficiency
choices for these vehicles.
J.
G.
Koomey, Energy Analysis Program, Energy
and
Environment
Division, Lawrence Berkeley Laboratory,
Building
90-4000, Univer-
sity
of California, Berkeley,
Calif.
94720.
D.
E.
Schechter
and
D.
Gordon, Union of Concerned Scientists, 2397 Shattuck Ave.,
Suite
203, Berkeley,
Calif.
94704.
CHARACTERISTICS
OF
HONDA CIVICS
This section describes the level
of
vehicle amenity of the
1991
Civic
DX
and the 1992 Civic
DX
and VX. Koomey
et
al. (2)
also describe a similar comparison between the
1991
Civic
base-model hatchback and the 1992 Civic CX Hatchback.
Examination
of
the specifications
of
these vehicles and actual
test drives reveal that fuel economy gains were achieved with
negligible impact
on
performance, driveability, and comfort.
We can conclude from the results
of
this section that the cars
deliver equivalent consumer utility.
General Description
The
1992 model year
Honda
Civics represent a
"new
gen-
eration"
of
Civics.
Honda
completely redesigned the engine,
body style, suspension, aerodynamics, and
other
major fea-
tures
of
this model
but
kept
total interior space constant while
improving performance. In addition,
Honda
added a new
hatchback, the VX, to its Civic line.
The
VX
is
similar to the
mid-cost Civic
DX
hatchback, except that the VX
is
optimized
for fuel economy.
Table 1 presents specifications and features of the
1991
and
1992 Civic
DX
and VX hatchbacks
(3-6;
J. Leestma, personal
communication).
The
major
difference among the
1991
Civic
DX,
the 1992
DX,
and the 1992
VX
is
the improved fuel
economy
of
the 1992 vehicles.
The
1992
DX
is
about
13
percent more fuel efficient than the
1991
model, whereas
the 1992 VX has 56 percent higher efficiency (this estimate
for the VX
is
for the "49-state" VX sold in all states
but
California).
The
1992
DX
and
VX
are slightly larger than the
1991
DX,
as shown by the interior and exterior dimensions given in
Table
1. In addition, the 1992 models are equipped with. a
driver-side air bag, resulting in improved safety over the
1991
DX.
The
fuel tank
of
the VX
is
more than 7 L (1.9 gal),
or
16
percent, smaller than those
of
the
1991
and 1992
DX.
However, the improved fuel economy of the VX means that
a
VX
owner would still have to refuel less often than an
identical
DX
owner.
Performance
Other
than fuel economy differences, operational and perfor-
mance variations among the three cars are minimal.
The
1992
VX
and the
1991
DX
are both rated
at
92
horsepower. How-
116
TRANSPORTATION
RESEARCH
RECORD
1416
TABLE
1 Specifications/Features
of
Honda
Civic Models
Specifications/Features
1991 DX
1992
DX
1992 vx
Fuel
Economy
Unadjusted liters per 100
km
(city/hwy)
6.916.0
(34/39)
6.0/5.3 (39/44)
4.4/3.9 (53/61) (a)
Adjusted liters per
100
km
(city/hwy)
7.6/6. 7 (31/35)
6.7/5.9 (35/40)
4.9/4.3 (48/55) (a)
Adjusted liters per
100
km
(composite)
7.2 (32.7)
6.3
(37.1)
4.6 (50.9)
Engine, Drive
Train
Horsepower(@ rpm)
92
@6000
102@
5900
92@
5500
Torque (Newton-meters @ rpm)
121
(89) @ 4500
133 (98) @ 5000
132 (97) @ 4500
Valve train SOHC, 16-valve
SOHC, 16-valve
VTEC-E
Fuel induction (b)
DP Fuel Injection
MP Fuel Injection MP Fuel Injection
Drive-train type Front-wheel Drive
Front-wheel Drive Front-wheel Drive
Transmission
5-Speed Manual
5-Speed Manual
5-Speed Manual
Final drive train ratio 3.89
4.06
3.25
Exterior
Dimensions
Wheelbase (cm)
250 257 257
Overall Length (cm) 399 407 407
Overall
Width (cm) 168
170 170
Curb weight (kg) 979 (2158) 988 (2178) 950 (2094)
Coefficient
of
drag 0.33 0.32 0.31
Interior
Dimensions
Headroom front/rear (cm)
97.0/93.0
98.0/93.0 98.0/93.0
Legroom front/rear (cm)
98.0/93.0 108177.5 108177.5
Cargo volume (cu. m) 0.48 0.38 0.38
Passenger
volume (cu. m)
2.1
2.2 2.2
Fuel capacity
(1)
45.0 (11.9) 45.0 (11.9)
37.9
(10)
Power
features
Steering
no no no
Windows no
no no
Safety
features
Driver airbag
not available
standard
standard
Cost
(1992 $)
Invoice/dealer cost
8171 (c)
8663
9258
MSRP (b)
9563 (c)
10140
10840
Performance
Seconds
to
go from 0
to
100 kph
NA
10.2
10.5
Source:
Reference (2).
English units given in parentheses. Fuel economy: mi/gal; torque: ft-lbs; curb weight: lbs; fuel
capacity: gal.
a.
Fuel economy is for the 49-State version
of
the VX. The California version is less efficient.
b.
DP = Dual-point; MP = Multi-point; MSRP = Manufacturer's suggested retail price.
c. 1991 costs adjusted
to
1992 $assuming 4% inflation.
ever, maximum horsepower
is
achieved at 5 ,500 rpm in the
VX and at
6,000 rpm in the
1991
DX. Thus, the. VX engine
provides slightly more power at engine speeds up to 5
,500
rpm, which
is
the range in which most drivers operate.
The
1992
DX
reaches a maximum horsepower of
102
at 5,900 rpm.
However, in comparison with the VX, the horsepower dif-
ference
is
likely to go unnoticed unless one drives at engine
speeds greater than
5,500 rpm (which
few
drivers ever do). The
time required to go from
0 to
100
kph
(62
mph)
is
also related
to horsepower. There
is
little difference between the
1992
DX
and the
1992
VX
in this area: the
1992
DX
takes 10.2 sec to
reach
100 kph, whereas the 1992 VX takes 10.5 sec.
the
VX both provide slight torque improvements over the
1991 DX. The 1992
DX
supplies
133
N-m (98 ft-lb)
at
5,000
rpm, whereas the
1991
DX
supplies
121
N-m (89 ft-lb) at
4,500 rpm. The VX
is
likely to have the best
"pickup"
at
engine speeds comparable with those encountered in everyday
driving, since it attains
132
N-m (97 ft-lb) of torque
at
only
4,500 rpm (J. Keebler, personal communication).
Driveability
The
comparison
of
features and specifications has focused on
the differences between the three vehicles
on
paper. How-
ever, before one can conclude that the Civic hatchbacks are
identical in terms
of
the service they provide,
one
must also
Another important indicator of vehicle performance
is
torque.
High torque allows quicker acceleration at low engine rpm
(e.g., when accelerating from a stoplight). The 1992
DX
and
Koomey
et
al.
evaluate the cars on the road. A series
of
drivers who test-
drove the
VX
found that, in general, it handled well and
performance was impressive. Some drivers found that they
had
to adjust their driving styles to
take
advantage
of
the
taller gearing
of
the VX (R. Maio, personal communication;
K.
Passino, personal communication). Taller gearing results
in lower engine speeds
than
those typically experienced in a
given gear. Some drivers also noted occasional engine
"stum-
ble,"
or hesitation, during quick acceleration in lean-burn
operation (7). This hesitation occurs as
the
engine adjusts to
a lower air/fuel ratio. All
but
one
Automotive News reviewer
believed that this effect would not adversely influence the
average driver's perception
of
the
vehicle's performance, and
the
reviewer who found the stumble unacceptable was a driver
who preferred high-performance vehicles (J. Keebler, per-
sonal communication).
For
typical Civic drivers (who prob-
ably do
not
seek high power), we can conclude from these
reviews that the performance and driveability
of
the
VX
are
equivalent to those
of
the 1991 and 1992
DX.
Comfort
and
Amenities
Although
the
primary specifications
and
performance
of
the
Civic models are essentially identical, minor differences exist
in the cosmetic features
of
the
DX
and
VX
hatchbacks. These
features and their estimated costs are described by Koomey
et
al. (2).
The
1991
and 1992
DX
models are both equipped
with an adjustable steering column,
rear
cargo cover, rear
windshield wiper, and bodyside molding, whereas the VX
lacks these features but has a tachometer and lightweight alloy
wheels.
The
cargo
area
cover adds utility to the
DX
models
because it hides any cargo and makes it
appear
that
the
vehicle
has a trunk.
The
lightweight alloy wheels on the
VX
are
cosmetic in
that
they look
"sportier,"
but
they also affect fuel
economy because
of
their lighter weight.
Safety
The
safety
of
the 1992 Civic models was improved significantly
by the addition
of
a driver's side air bag in
both
the
DX
and
the VX.
The
1991
DX
does not have a driver's side air bag.
The
added safety provided by the air bag
is
reflected in re-
duced insurance premiums.
For
example,
the
United Services
Automobile Association
(USAA)
Casualty Insurance Com-
pany reduces the premium for medical payments coverage
(MPC) by
60
percent
compared
with
the
1991
DX
for owners
of
the 1992
DX
or
VX
(V. Blackstone, personal communi-
cation).
There
is
no difference in
the
premium for
MPC
for
the 1992
DX
and VX, which indicates
that
professional risk
assessors
of
at least
one
major
insurance company believe
that
the
slight difference in weight
of
these two vehicles has
a negligible effect on safety.
Furthermore,
because the
VX
is
lighter
than
the
DX,
its use imposes less risk
on
other
vehicles.
There are currently no crash test
data
with which to further
compare the safety
of
these vehicles.
Emissions
As described by Koomey
et
al. (2),
CO
and
HC
emissions
from the 49-state version
of
the VX are comparable with those
117
from
the
1991
and 1992
DX.
NOx emissions are slightly higher
in
the
49-state
VX
than
in the
1991and1992
DX
models, and
carbon dioxide emissions are lower in direct relation to
the
efficiency
of
the vehicles. All
of
these automobiles
meet
cur-
rent emissions standards in the states in which they are sold.
FUEL ECONOMY TECHNOLOGIES
As discussed above, the 1992 VX provides a
56
percent im-
provement
in efficiency over the
1991
DX.
This
is
achieved
by
the
use
of
technological improvements
that
increase
the
efficiency
of
converting fuel energy to usable work and reduce
the
amount
of
work required to move the vehicle.
The
technological differences responsible for
the
improved
fuel economy in the
VX
include
•
VTEC-E
engine with lean-burn,
• Changes in axle and gear ratios,
•Multipoint
fuel injection,
•Decreased
vehicle weight,
•Improved
aerodynamic characteristics,
• Low rolling resistance tires,
•Reduced
idle speed, and
• Shift indicator light.
Table 2 summarizes these technologies and presents estimated
contributions to fuel efficiency and costs (in 1992 dollars)
associated with each approach
(8-10; T. Harrington, personal
communication).
More
details on particular technologies are
provided by Koomey
et
al. (2) and Bleviss (11).
The
largest percentage improvements come from trans-
mission/gearing and engine modifications: This fact
is
note-
worthy because changing engine and transmission character-
istics do not affect safety
or
vehicle size. Only weight reduction
may have an effect on safety, depending on where
the
weight
is
removed.
The
weight changes in the
VX
are small
(3
to
4
percent), so they are unlikely to significantly affect safety.
Capital Costs
of
Fuel Economy Improvements
The
costs
of
the technologies listed previously are not readily
available and vary widely depending on the source
of
the
estimate.
The
process
of
estimating costs
is
further compli-
cated by the fact
that
several
of
the technologies may overlap.
For
example,
the
variable valve feature
of
the
VTEC-E
engine
permits the use
of
lean-burn technology and changes in drive
ratio. Thus, an estimate
of
the
cost
of
variable valve timing
may also include the cost
of
lean-burn technology
and
drive
ratio changes. Despite these complications, we provide esti-
mated
costs
of
fuel economy techn.ologies in Table
2.
The
total estimated costs
of
these technologies range from $448
to $1,084.
Applicability of Civic
VX
Improvements to
Other
Vehicles
Not
all technologies used to improve
the
efficiency
of
the
Civic can currently be transferred to
other
new cars. We focus
118
TRANSPORTATION
RESEARCH
RECORD
1416
TABLE
2 Technologies Used To
Increase
Efficiency in the 1992 VX
Efficiency
Improvement(%)
Cost (a)
Technology '91 DX to
'92 VX (1992 $/car)
Multi-point fuel injection
1.5
56-162
Low rolling resistance tires 1
21-22
VTEC-E engine
variable valve timing
2.5
108 -164
lean
bum
5 -10 150 - 500
reduced friction
1.5
35-65
roller cam followers
1
19-54
Weight reduction 2.5
37 - 78 (b)
Aerodynamic improvements 1.5
22 - 39 (c)
Gearing and drive ratio changes
21
NIA
(d)
Reduced idle speed/rpm
3
NIA
(d)
Shift indicator light
5
NIA
(d)
Total
45.5 -
50.5 (e) 448 - 1084
Source: Reference (2).
a.
All costs represent retail costs to the consumer. Most cost estimates adjusted from 1988 and
1990
$based
on
4.1
% implicit price deflator for GNP for 1989 and assumed 4% annual deflator
for
1990 to 1992.
b. Cost estimate from Greene and Duleep based on
$0.50llb reduced (1988$). Estimate from
SRI based on 5% weight reduction.
c.
Cost based on 10% aerodynamic improvement
d.
NI A = not available.
e.
Totals based on simple addition do not add to 56% due to synergistic effects
of
fuel
economy technologies (e.g., variable valve timing allows gearing changes and use
of
lean bum).
in particular
on
the applicability of the
lean-bum
engine.
D_e-
tails on how
other
efficiency options might apply to different
portions
of
the
tJ
.S. automobile fleet
are
given by
Ledbetter
and Ross (12).
Keebler (7) reports that
"heavy vehicles have poor drive-
ability when calibrated with lean-bum fuel strategies," which
implies that this strategy, as currently implemented, may
not
be directly transferable to the larger cars in the U.S. fleet.
Because
of
increasingly strict
NOx
emissions standards, lean-
burn technology may not be viable in some vehicles until
improved
NOx
catalysts are developed. According to Sanger
(13),
Honda
engineers believe it will be "several years
...
before they can transfer the technology to larger, less efficient
engines." However, it has been reported that
Honda
plans
to use lean-bum technology on its larger Accord model as
early as the 1994 model year
(14). Research
on
this issue
is
proceeding elsewhere as well. Recently, a company in Mas-
sachusetts announced the development
of
a new lean-bum
engine that combines high efficiency and low
NOx
emissions
for an additional cost of
$100 to $200
per
car (15).
METHODOLOGY
The purpose
of
the calculations in the next two sections
is
to
estimate the costs and benefits
of
improving the fuel economy
of
the
1991
Civic
DX
to the level of the 1992 Civic
DX
and
VX models. Actual retail prices are used to estimate the cost
of
improving fuel economy, whereas projections of
motor
gasoline prices are used to estimate the levelized fuel price.
Definition of Cost-Effectiveness
By cost-effective, we mean that the costs of investing in au-
tomobile efficiency are lower than the costs avoided by this
investment. The cost
of
an efficiency improvement is usually
assessed by calculating the
CCE.
The
costs avoided by the
efficiency investment include the direct cost
of
the unused
fuel and whatever social
or
external costs are associated with
the consumption
of
gasoline
that
are not included in the gas-
oline price. Whenever the
CCE
is
lower than the avoided
direct costs plus external costs (in dollars
per
gallon), we can
say that an efficiency investment
is
cost-effective.
Cost Perspective
We adopt the perspective
of
the buyer of a new car who will
use the vehicle over its entire lifetime. This simplifying
as-
sumption
is
also roughly comparable with the societal per-
spective without externalities (assuming that the discount rate
used reflects social and not individual preferences).
CCE
The
CCE
(in dollars
per
liter) is calculated using Equation
1:
CCE
capital cost ($) X
[l
_
(l
d + d)-n]
annual energy savings (liters)
(1)
Koomey
et
al.
where
d = discount rate,
n = lifetime of the automobile, and
dl[l
-
(1
+
d)-"]
=
the
capital recovery factor.
The
numerator in the right-hand side of Equation 1
is
the
annualized cost
of
the conservation
or
efficiency investment.
Dividing annualized cost by annual energy savings yields the
CCE,
which
is
independent
of,
but
can be compared with,
the levelized price
of
fuel (in dollars
per
liter). More details
on such calculations are given by Meier et al.
(16) and Koomey
et
al. (17).
Consumer Choice Models
There
is
some controversy over the procedure that consumers
actually use to choose
the
efficiency level
of
the automobiles
they purchase.
Greene
(18), in a review
of
such decision al-
gorithms, summarizes this controversy.
The
main issue
of
contention concerns the multifaceted nature
of
the purchase
decision.
Usually, the choice between vehicles
is
based
on
many decision criteria, most of which are unrelated to
the
efficiency of
the
vehicle.
The
use
of
a
CCE
model (or, equiv-
alently, a life cycle cost model) to describe such choices is
problematic in
that
it
is
a simple measure that does not address
the complexity
of
the purchase decision.
Whereas this issue
is
important in assessing consumer choices
over a broad range
of
vehicle types, it does not significantly
affect our analysis. We have, to a first approximation, created
a comparison between vehicles that have different fuel econ-
omy but are otherwise equivalent in terms
of
size, features,
performance,
and
safety.
For
this reason, we believe that it
is
appropriate to discuss choices between these vehicles as
if consumers were actually using a discount rate in a
CCE
calculation.
Discount Rate
The
discount
rate
in
our
calculations
is
7 percent real. This
value roughly corresponds to the current cost of capital for
consumers seeking an automobile loan
(11
to
12
percent with
inflation). We also perform a sensitivity analysis using real
discount rates
of
3, 10,
and
30 percent.
The
results
of
the
sensitivity analysis
are
described by Koomey
et
al. (2).
Miles Driven
We use an estimate
of
16 400 km (10,200 mi) traveled
per
year for a typical U.S. automobile in 1988 [Davis and Morris
(19)].
The
source cited by Davis and Morris
is
the U.S.
De-
partment
of
Energy's Residential Transportation
Energy
Consumption Survey.
Rebound Effect
Greene (20) suggests, after reviewing the literature,
that
con-
sumers will increase
their
vehicle miles traveled by 0.05 to
119
0.15 percent in response to a 1 percent decrease in the fuel
cost
per
mile of their vehicles. We omit this factor in calcu-
lating the
CCE,
because if consumers use their vehicles more,
the increased mobility must be worth more to them than the
increased expenditure on gasoline. Therefore,
our
per
unit
cost-effectiv~ness
calculation
is
unaffected by such rebound.
If
one
is
interested in calculating total energy savings from
a given policy affecting many such vehicles, this correction
factor must be included. We do not make such a calculation
here. In any case, the correction is a small one.
Vehicle Lifetime
We use an estimate
of
automobile lifetime of 13.3 years de-
rived from a retirement curve for vehicles presented Davis
and Morris
(19). This curve applies to vehicles purchased
between 1987 and 1989. We assume that the fuel economy
improvement technologies used in the VX will not affect the
vehicle lifetime.
Rated Fuel Economy
Fuel economy estimates based
on
the
EPA
test procedure
have been found to diverge from actual performance. This
divergence was significant enough to
induce·
EPA
to reduce
the sticker fuel economy relative to the test procedure values
to
better
account for real-world driving conditions. Beginning
in 1985,
EPA
reduced the city fuel economy estimates from
the test procedure by
10
percent and reduced the highway
estimates by
22
percent to calculate the fuel economy rating
on
the sticker. This correction
is
important, because if actual
miles
per
gallon (mpg)
is
lower than the rated mpg, using the
rated mpg to calculate .gasoline savings will underestimate
those savings in absolute terms.
We use the city and highway fuel economy as listed
on
the
EPA
sticker for each car, which includes the preceding cor-
rection factors. We weight the city and highway fuel economy
sticker values to estimate composite fuel economy for
our
cost-effectiveness calculations. This weighting assumes that
55
percent
of
driving
is
city driving and
45
percent
is
highway
driving, as specified in Section 503
of
the Energy Policy and
Conservation
Act
passed in 1975.
Consistency
of
Comparison
All fuel prices and capital costs are in 1992 dollars. We use
a real discount rate (without inflation) to levelize the prices
and the same real discount rate
to
calculate the CCE.
The
comparison between the initial capital expense and the lev-
elized fuel price
is
therefore consistent.
Fuel Prices
Average
motor
gasoline prices are taken from the
Annual
Energy Outlook (21) and are levelized using a 7 percent real
discount rate [using
the
method
of
Kahn (22)]. According to
the forecast, the retail price of motor gasoline will be
$0.34/L
120
($1.27/gal) in 1992
and
$0.43/L ($1.61/gal) in 2005 (in 1992
dollars, calculated assuming 4 percent inflation for
1990 to
1992). Levelized over this period (which corresponds to the
lifetime
of
our
Honda
Civic purchased in 1992), the price
of
gasoline
is
$0.37/L ($1.40/gal).
This price includes roughly
$0.07 to $0.08
per
liter ($0.25
to
$0.30
per
gallon)
of
state and federal gasoline taxes, which
are used primarily
to
fund highway construction and main-
tenance. Society does
not
avoid the construction and upkeep
of
roads if automobiles are more efficient,
so
a societal cost
comparison should
not
include these costs in the avoided cost
of
fuel. This price also does not include the external costs
associated with gasoline combustion, many of which are re-
duced by a more fuel-efficient car.
·
We show comparisons with the levelized fuel price with
and
without taxes.
The
case with taxes provides an understandable
reference point and represents the situation in which avoid-
able external costs roughly equal the level of state and federal
taxes.
The
case without taxes represents the situation in which
external costs are assumed to be equal to zero.
Operation
and
Maintenance Costs
We assume that operation and maintenance
(O&M) costs for
the
VX
are
unaffected by the technologies used to achieve
improved fuel economy. Thus, we assume that lifetime
O&M
costs for the
1991
and 1992
DX
models and for the 1992
VX
are identical.
Invoice Cost Versus Manufacturer's Suggested Retail
Price
Invoice cost
is
also known
as
dealer cost.
It
is
the average
price charged to the
dealer
by the automobile manufacturer.
Manufacturer's suggested retail price
(MSRP)
is
also known
as sticker price and
is
supposed to represent the price
of
the
car to the consumer. In this analysis, we rely on
MSRP as an
"official" price.
The
invoice costs are reported by Koomey
et
al. (2).
The
invoice cost and MSRP are taken from
USAA
(5
,6)
and documents from a local
Honda
dealer.
The
invoice
cost and
MSRP for the 1991
DX
have been adjusted to 1992
dollars, assuming 4 percent inflation.
The
MSRP
is
somewhat arbitrary.
Good
bargainers have
been
known to purchase automobiles at
or
below the invoice
cost. Automobile manufacturers also give
"volume incen-
tives" to dealers that sell more than a target number of cars.
Therefore, invoice cost
and
MSRP based on the sale
of
a
single car may not actually reflect the true cost to the dealer.
Does Retail Price Reflect
True
Cost?
Automobile pricing
is
a complicated process, and the
market
price of a vehicle may have little to do with actual production
costs.
For
example, antilock braking systems
provic;ied
as an
option
on
many cars are currently underpriced on the vehicle
"sticker" to encourage the purchase of these safety-enhancing
mechanisms (L. Rinek, personal communication). Some of
the redesign costs for the new Civics are probably included
TRANSPORTATION
RESEARCH
RECORD
1416
in the MSRP, as are any savings from the redesign. Without
detailed manufacturer data, we cannot determine the extent
to which such cost changes are related to fuel economy im-
provements alone. We also
cannot
know whether
Honda
is
taking a loss on the VX because it wants to gain experience
with new technology in anticipation of growing demand for
efficiency in a more environmentally conscious world.
We do not have access to
Honda's
cost data, and we cannot
determine the manufacturer cost for improving the fuel econ-
omy in the
Honda
Civic hatchbacks. Nevertheless,
we
believe
that the
MSRP offers an approximate representation of
the_
actual cost of improving automobile fuel economy in these
vehicles.
Air Bags
The
1992 Civics both have airbags, whereas the
1991
DX
does
not. Except for a minor weight penalty, air bags are unrelated
to fuel economy, and their cost should not
be
included in
our
assessment
of
the incremental cost associated with improving
the efficiency of the
1991
DX.
The
MSRP cost of an air bag
is
$800 in a new
Honda
Civic and $1,200 to replace an air bag
that has been "blown" in a collision
(R.
Maio, personal com-
munication). We subtract
$800 from the
MSRP
cost of the
1992
DX
and VX to correct for this added cost.
Correction for Cosmetic Differences
The
cargo cover
is
available
as
an option on the VX for $159.
Costs for the other cosmetic differences can only be roughly
estimated on the basis
of
estimates by the parts department
of
a local
Honda
dealer.
We add the average cost
of
hatch cover, body side molding,
and rear wiper/washer to the cost
of
the
VX (no correction
is
made for the cost of the adjustable steering column, since
the
costs of replacing the steering columns in the
DX
and the
VX
are the same). We add the midrange cost of the tach-
ometer
and half the cost
of
the alloy wheels to the price
of
the
1991
and 1992
DX.
Only half
the
cost of the alloy wheels
is
added to the
DX
price because some fraction of their cost
is
attributable to their lower weight
and
the rest
is
attributable
to their
"sporty" appearance. We choose half arbitrarily, since
we had no way to separate these two attributes
of
the wheels.
These cosmetic differences result in an additional
MSRP
cost of
$365
on both
DX
models and $614
on
the VX. By
correcting for cosmetic differences
and
for the air bag,
we
have created a consistent comparison and can draw conclu-
sions about the actual cost to improve the efficiency of the
1991
DX
to the level
of
the 1992 VX.
These
corrections result
in what we refer to as
our
"full correction" case, which rep-
resents
our
best estimate for the retail price of the fuel econ-
omy improvements in the VX compared with the
DX.
Although these cost corrections
make
the comparison more
consistent, they should be viewed as approximate for three
reasons:
1. Actual costs for these options are speculative, since the
features available
on
the
VX
are not available on the
DX,
and vice versa.
Koomey
et
al.
2.
Actual production costs for standard features may be
quite different from the costs for installing such features as
options after
the
car
is
manufactured.
3.
Separating the cost
of
the alloy wheels attributable to
cosmetic differences from
that
attributable to fuel economy
is
problematic.
To
account for the fact
that
some options are not-available
on specific models in the showroom, we also show a com-
parison between the 1992
DX
and the
1992
VX
that only
corrects for
the
feature
that
is actually an
option-the
hatch
cover. We
refer
to this case as the
"as
available" comparison.
Comparison
of
Estimated Technology Costs with
Retail Cost Difference
When we
compare
the estimated costs of fuel economy tech-
nologies (Table 2) with
the
retail cost difference calculated
after making
the
corrections described above, we see
that
the
results are similar.
The
mean
of the engineering cost estimates
for the VX efficiency improvements (Table 2)
is
$766, whereas
the cost difference between the
1991
DX
and the 1992
VX
in
Table 3 (based
on
the "full correction" case)
is
$726.
In
view
of
the
rather
large range
of
error
to be expected in such a
comparison, we can conclude that the engineering costs and
our
retail cost calculation give roughly the same result, which
gives us confidence that
our
calculations are
of
the correct
approximate magnitude.
COST-EFFECTIVENESS
OF
FUEL ECONOMY
TECHNOLOGIES
Cost of Improving 1991
DX
Efficiency to 1992
DX
and
VX
Levels
121
In this calculation, we estimate the cost of improving fuel
economy
of
the
1991
Honda
Civic
DX
to the levels of the
1992 Civic
DX
and VX. This information can
be
used to
determine whether the fuel economy
of
a particular vehicle
can be improved substantially
at
a cost less than the cost
of
fuel, while keeping vehicle amenity constant and without re-
ducing safety.
Table 3 gives the results
of
this calculation.
The
MSRP cost
of
an air bag ($800) has
been
subtracted from the cost
of
the
1992 Civics, which makes the
13
percent efficiency improve-
ment for the 1992
DX
achievable at negative net cost. Engine
torque also increased relative to the
1991
DX
in this case.
This result implies
that
Honda
improved the fuel economy
and power
of
this vehicle while reducing its initial cost.
After
subtracting the cost
of
the air bag, the additional
incremental cost for the VX over the
1991
DX
is
$477.
The
correction for cosmetic differences increases the incremental
cost of moving from the
1991
DX
to the 1992
VX
by $249,
giving a total incremental cost for
the
VX of $726. This $726
cost translates to a
CCE
of
$0.20
per
conserved liter ($0.
77
per
conserved gallon), which
is
about
45
percent less than the
levelized price
of
gasoline with taxes and
30
percent less
than
the price without taxes. This
CCE
corresponds to a simple
TABLE
3
Cost
of
Conserved
Gasoline
for
1992
Honda
Civic DX
and
VX
Hatchbacks
Changes
in fuel economy, fuel use,
and
capital costs
Costs
Costs
Costs
Costs
Fully Corrected
Fully Corrected
As
Available
Fully Corrected
91
DX
92DX
92
vx
91DX to 92DX 91DX to 92VX
92DXto
92VX
92DXto
92VX
EPA fuel economy estimates
Adjusted liters/100 km (city)
7.6(31)
35
48 13%
55% 37%
37%
Adjusted liters/100 km (highway)
6.7 (35)
40
55
14% 57% 38%
38%
Adiusted liters/100 km (EPA composite)
7.2 (32.7) 37.1
50.9
13% 56%
37%
37%
Fuel used (liters/vear)
1181
1041
757
-140
-424 -284
-284
MSRP cost (92 $)
9563
10140 10840
577
1277
700
700
MSRP
cost adjusted for airbag+cosmetic diffs (92
$)
9928
9705
10654
-223
726
859
949
Annualized incremental
MSRP cost ($/year)
1170 1144 1256 -26 86
101
112
CCE based on
MSRP cost (92 $niter)
<0
0.20 (0.77)
0.36
(1.36) 0.40 (1.50)
Simple payback time-MSRP & gas price w/tax (yr)
<0
4.6
8.2
9.0
Simple payback time-MSRP & gas price w/o
tax
(yr)
<0
5.9
10.4 11.5
Other parameters
Real discount rate
7%
Distance driven/year (km)
16415 (10200)
Auto lifetime (years)
13.3
Capital recovery factor
11.8%
MSRP cost
of
airbag (1992 $)
800
City driving percentage
55%
Levelized cost
of
gasoline w/taxes (92$/liter)
0.37 (1.40)
Highway driving percentage
45%
Levelized cost
of
gasoline w/o taxes (92$/Iiter)
0.29 (1.10)
Source: Reference (2).
English units given in parentheses. Fuel economy: mi/gal; fuel used: gal; CCE: 92$/gal; distance driven: mi; cost
of
gasoline: 92$/gal.
122
payback time
of
about
4.6 years using MSRP
and
including
taxes in the gasoline price and to 5. 9 years when taxes are
omitted.
Cost
of
the Consumer's Choice: 1992
DX
Versus
1992
vx
We also investigate
the
actual efficiency choice available to
consumers
on
the showroom floor (1992
DX
versus 1992 VX).
We show two cases: (a)
the
"as
available" case, which corrects
only for the cost
of
the
hatch cover in the VX, and (b) the
"full correction" case, which uses the cost estimates for all
the cosmetic differences. Table 3 gives
the results
of
this cal-
culation, which indicates
that
the
CCE
relative to the 1992
DX
(based
on
MSRP)
is comparable with the levelized price
of
gasoline with taxes
and
roughly
25
to
30
percent higher
than the levelized price
of
gasoline without taxes. This
CCE
corresponds to a simple payback time of 8 to
11
years de-
pending
on
the treatment
of
taxes and cosmetic differences.
These paybacks are long enough to make consumers think
twice about spending
the
extra money for the VX.
Limitations
of
Cost-Effectiveness Calculations
These calculations were
done
without accounting for external
societal costs. External costs include all costs to society
that
are not included in the
market
price of gasoline, such as
increased health costs; costs arising from damage to agricul-
ture; costs resulting from damage to physical structures
due
to air pollution from automobiles; increased national security
costs from consumption
of
imported oil; and increased en-
vironmental damage from acid
rain, carbon dioxide emissions,
and
other
pollutants. In practice, exact numerical values for
these externalities are difficult to calculate
(23). Many authors
have attempted to assess these costs in monetary terms, and
in general they find that such costs are probably substantial
(24-29). We do not estimate these costs here but simply note
that accounting for
them
would improve the relative cost-
effectiveness of efficiency improvements compared with
the
consumption
of
gasoline alone.
POLICY IMPLICATIONS
The
range
of
issues surrounding policies designed to affect
vehicle efficiency choices
are
too complex to describe in detail
here and are described elsewhere
(30,31).
Our
purpose in this
section
is
to summarize the most important policy-related
conclusions emerging from
our
work. These conclusions are
described more fully by Koomey
et
al. (2).
Implications for Society
We have shown that improving the fuel economy of a partic-
ular vehicle (the
1991
Civic
DX)
was not only possible, it was
cost-effective from society's perspective. The efficiency im-
provements in the 1992 Civic
VX
were achieved at a
CCE
that
is
about
45
percent less
than
the levelized cost
of
gasoline
TRANSPORTATION
RESEARCH
RECORD
1416
with taxes and 30 percent less than
the
levelized cost of gas-
oline without taxes (relative to the 1991
DX).
This empirical
evidence indicates that, at least for small cars similar to the
Civic, improvements in fuel economy can be achieved at at-
tractive costs.
Implications for Consumers
The
1992 Civic
DX
and
VX
deliver comparable performance,
but
the
VX
delivers higher fuel economy
at
a
CCE
that
is
comparable with the avoided cost of fuel. A consumer decid-
ing between these two vehicles will have little, if any, direct
economic incentive to choose the VX, although the United
States as a whole might prefer the lower carbon dioxide emis-
sions and reduced use
of
imported oil
of
the
more
efficient
vehicle. According to estimates from
Honda
Corporation,
about
5 percent
of
1992 Civic sales were VXs (32).
Need for Public Policy
Because consumers have little economic incentive to purchase
the
more
fuel-efficient vehicle, public policy is required to
ensure that socially beneficial choices are
made
regarding fuel
economy. Policies such as gas taxes and fee bates (which im-
pose fees
on
purchases of gas guzzlers while providing rebates
for purchases of fuel-efficient vehicles) would
make
the
VX
more cost-effective relative to the
DX
for consumers. Thus,
consumers would have incentive to act in a
manner
that
ben-
efits society.
CONCLUSIONS
Honda
has demonstrated that modest efficiency improve-
ments (13 percent) can be achieved
at
negative
net
cost in its
1992 Civics. Efficiency improvements
of
56
percent can be
achieved
at a
CCE
that
is
45
percent less than the cost
of
the
saved gasoline with taxes and
30 percent less than the cost of
the saved gasoline without taxes. Virtually all
of
the fuel
efficiency improvements in the 1992 Civic
VX
were achieved
using technologies that do
not
change safety or vehicle amen-
ity.
These
results suggest
that
the difficulty and cost
of
im-
proving fuel economy in new compact and subcompact au-
tomobiles may be less than has been suggested by
U.S.
automobile makers.
ACKNOWLEDGMENTS
Review comments and valuable advice were provided by Chris
Calwell
(NRDC);
Lee Schipper, Rich Howarth,
Evan
Mills,
and
Alan
Sanstad (LBL); Amory Lovins (Rocky Mountain
Institute); and Marc Ross (University of Michigan and Ar-
gonne National Laboratory). Jack Keebler of
Automotive News
provided insight into the performance and driveability of the
1992 VX. Toni Harrington of
Honda
North America gave us
detailed information
on
fuel economy improvement technol-
ogies. Finally, thanks are due to Jim
Doten's
Honda
of
Berke-
ley, California, for allowing us to test drive the 1992 Civics
Koomey et
al.
on more than one occasion and for providing information on
consumer response and option costs.
REFERENCES
1. Shin, D. The Costs
of
Federally Mandated Policies To Promote
Energy Conservation: The Case
of
the Automobile Efficiency
Standard.
Draft
Working Paper. American Petroleum Institute,
June 1989.
2.
Koomey,
J.
G.,
D.
E.
Schechter,
and
D.
Gordon.
Cost-
Effectiveness
of
Fuel Economy Improvements in
1992
Honda
Civic Hatchbacks. Presented at 72nd Annual Meeting of the
Transportation Research Board, Washington,
D.C.,
1993.
3.
1991
Civic Specifications and Features. American Honda Motor
Co., Inc.,
1990.
4. 1992 Civic Specifications and Features. American Honda Motor
Co., Inc., 1991.
5.
USAA
Auto
Pricing Plus: 1991 Honda Civic. United States Au-
tomobile Association, 1991.
6.
USAA
Auto
Pricing Plus: 1992 Honda Civic Hatchback. United
States Automobile Association, 1991.
7. Keebler, J.
Honda
Civic VX: Experiment in Clean, Fuel-Efficient
Motoring Pays
Off-Almost.
Automotive News June
29
1992
p.
20. , ' ,
8.
Greene,
D.
L.
and
K.
G. Duleep. Costs and Benefits
of
Auto-
motive Fuel Economy Improvement: A Partial Analysis. ORNL-
6704. Oak
Ridge National Laboratory, March 1992.
9. Automotive Fuel Economy: How
Far
Should
We
Go? National
Academy Press, Washington,
D.C.,
1992.
10.
Potential
for
Improved Fuel Economy in Passenger Cars and Light
Trucks.
SRI
International, July 1991.
11.
Bleviss,
D.
L.
The New Oil Crisis and Fuel Economy Technol-
ogies. Quorom Books, New York, 1988.
12.
Ledbetter, M., and M. Ross. Supply Curves
of
Conserved Energy
for Automobiles. Lawrence Berkeley Laboratory, March
1990.
13. Sanger,
D.
E. Fuel Efficiency: New Japan Coup? New
York
Times, July 31, 1991, p.
Cl.
14.
Johnson,
R.
Next Accord May Offer New Engine. Automotive
News, Oct. 19, 1992, p.
30.
15.
Lean Burn Takes a Step Forward. Global Environmental Change
Report,
Oct. 23, 1992, p. 7.
16.
Meier,
A.,
J. Wright, and
A.H.
Rosenfeld. Supplying Energy
Through Greater Efficiency.
University
of
California Press,
Berkeley, 1983.
17.
Koomey,
J.,
C. Atkinson,
A.
Meier,
J.E.
McMahon,
S.
Bogh-
123
osian, B. Atkinson,
I.
Turiel,
M.
D. Levine, B. Nordman, and
P. Chan. The Potential for Electricity Efficiency Improvements in
the U.S. Residential Sector.
LBL-30477. Lawrence Berkeley Lab-
oratory, July 1991.
18.
Greene, D.
L.
A Note on Implicit Consumer Discounting
of
Automobile Fuel Economy: Reviewing the Available Evidence.
Transportation Research, Vol. 17B, No.
6.
1983, p.
491.
19. Davis,
S.
C.,
and
M.
D. Morris. Transportation Data Book: Edi-
tion 12.
Oak
Ridge National Laboratory, March 1992.
20. Greene,
D.
L.
Vehicle Use and Fuel Economy: How Big
is
the
"Rebound"
Effect? The Energy Journal, Vol. 13, No. 1, 1992,
p. 117.
21.
Annual
Energy Outlook, with Projections to 2010.
DOE/EIA-
0383(92). Energy Information Administration, U.S. Department
of
Energy, 1992.
22.
Kahn, E. Electric Utility Planning and Regulation. American
Council for an Energy-Efficient Economy, Washington,
D.C.,
1988.
23. Koomey, J. Comparative Analysis
of
Monetary Estimates
of
Ex-
ternal Environmental Costs Associated with Combustion
of
Fossil
Fuels. LBL-28313. Lawrence Berkeley Laboratory, July
1990.
24. MacKenzie, J.
J.,
R. C. Dower, and
D.
T. Chen. The Going
Rate: What It Really Costs To Drive. World Resources Institute
Washington;
D.C.,
1992. '
25. Hall, D. C.
Oil and National Security. Energy Policy, 1991.
26. Cannon, J.
S.
The Health Costs
of
Air
Pollution, Third Edition
1984-1989. American Lung Association,
1990.
27. DeLuchi,
M.A.,
D. Sperling, and R.
A.
Johnston. A Compar-
ative Analysis
of
Future Transportation Fuels. UCB-ITS-RR-87-
13. Institute
of
Transportation Studies, University
of
California,
Berkeley, 1987.
28.
Broadman, H. G. The Social Cost
of
Importing Oil. Energy Pol-
icy, Vol. 14, No. 3, 1986, p. 242.
29. Bohi,
D.,
and W. D. Montgomery. Oil Prices, Energy Security,
and Import
Policy. Resources for the Future, Washington,
D.C.,
1982.
30. Gordon, D. Steering a New Course: Transportation, Energy, and
the Environment. Island Press, Washington,
D.C.,
1991.
31. Griffin, J. M., and H. B. Steele. Energy Economics and
Policy.
Academic Press College Division, Orlando, Fla. 1986.
32.
Pollack,
A.
Running Half-Engine To Save Fuel. New York Times,
Nov. 18, 1992, p. C16.
The views expressed here do not necessarily represent the views
of
the
Lawrence Berkeley Laboratory.
Publication
of
this paper sponsored by Committee on Energy Con-
servation and Transportation Demand.
