Content uploaded by Roger Bezdek
Author content
All content in this area was uploaded by Roger Bezdek on May 29, 2019
Content may be subject to copyright.
A reprint from
American Scientist
the magazine of Sigma Xi, The Scientific Research Society
This reprint is provided for personal and noncommercial use. For any other use, please send a request to Permissions,
American Scientist, P.O. Box 13975, Research Triangle Park, NC, 27709, U.S.A., or by electronic mail to perms@amsci.org.
©Sigma Xi, The Scientific Research Society and other rightsholders
132 American Scientist, Volume 93 © 2005 Sigma Xi, The Scientific Research Society. Reproduction
with permission only. Contact perms@amsci.org.
In the fall of 1973, a new phrase en-
tered the American lexicon: “energy
crisis.” On October 17 of that year, the
Organization of Petroleum Exporting
Countries (OPEC) slapped an embargo
on oil exports, hoping to punish the
United States for its support of Israel
in the Six Days’ War.
Although the embargo lasted only
five months, its effect on the U.S. econ-
omy was profound. Long lines of cars
became commonplace at filling sta-
tions, and gas rationing reared its head
for the first time since World War II. No
less profound was the impact of the
embargo on the popular psyche. For
the first time, most Americans awak-
ened to the fact that they were depen-
dent on oil from abroad—and not just
from anywhere, but from one of the
most politically volatile regions in the
world. Recognizing the danger, Con-
gress passed a variety of belt-tightening
laws, including the Energy Policy and
Conservation Act of 1975, which set up
mandatory fuel-efficiency standards
for automobile manufacturers.
Although the nation has become
much more efficient in its use of energy
since the 1970s, it nonetheless imports
60 percent of its oil, twice as large a
share as thirty years ago. Much of that
oil still comes from the Middle East, a
region that has become no more stable.
The toppling of the Shah of Iran in
1978 precipitated a second energy cri-
sis that winter. In 1981, the cost of gas
at the pump reached its highest levels
ever (nearly $3.00 in inflation-adjusted
dollars). Prices moderated over the fol-
lowing decades, but the terrorist at-
tacks of September 11, 2001, and the
ensuing wars in Afghanistan and Iraq
have once again called into question
the security of U.S. oil supplies.
Each of these episodes has led to
talk of toughening the fuel-economy
standards enacted in 1975. Most re-
cently, a bill that senators John Kerry
and John McCain sponsored in 2002,
which would have raised the overall
gas efficiency of American cars and
trucks by nearly 30 percent over the
next 10 years, failed to pass when 19
Democrats joined 43 Republicans in
voting against it. Whereas environ-
mentalists asserted that higher fuel-
economy standards were needed to
reduce gasoline consumption and
emission of greenhouse gases, the auto
industry and labor unions argued that
these requirements would drastical-
ly increase the price of new cars and
put hundreds of thousands of people
out of work. The labor unions’ argu-
ments—and their well-attended ral-
lies—proved to be particularly persua-
sive for the Democratic legislators.
Although fuel-economy standards
have not changed significantly in
20 years at the federal level, another
player recently entered the debate.
This fall, California, the only state
with legal authority to regulate vehicle
emissions, instituted new rules that
would require auto manufacturers to
reduce the output of greenhouse gases
(chiefly carbon dioxide) by 30 percent
before 2016. Unlike carbon monoxide
and other pollutants, carbon dioxide
cannot be eliminated by add-on tech-
nology like catalytic converters; it can
be reduced only by cutting fuel con-
sumption. The fate of California’s rules
will probably be played out in court,
as auto manufacturers argue that the
state has tried to bring in fuel-economy
standards through the back door.
How much fuel would the nation
save by enacting standards like the ones
in California or the ones voted down in
2002? Would the stiffer requirements
harm the economy? Would they really
cost thousands of workers their jobs?
Over a period of 30 years, first with the
Department of Commerce and then
as independent consultants, we have
developed a quantitative model of the
economy that is ideally suited to answer
such questions. We have employed this
model in the past for analyzing the cost
and benefit of national acid-rain legisla-
tion, and for gauging the contribution
of the environmental industry to the
U.S. economy, among others uses. Our
latest application of this model brings
some good news: Fuel efficiency can go
hand in hand with job growth.
CAFE: One Lump or Two?
When the energy crisis hit America in
1973, the fuel efficiency of the average
U.S. passenger car had fallen to less
than 13 miles per gallon (mpg). The
Energy Policy and Conservation Act of
1975 instituted a new Corporate Aver-
age Fuel Economy (CAFE) program,
which required automobile manufac-
turers to more than double the efficien-
cy of the cars they sold. The increase
was phased in over several years: For
the 1978 model year, the standard for
passenger cars was set at 18 mpg, and
Fuel Efficiency and the Economy
Input-output analysis shows how proposed changes to automotive
fuel-efficiency standards would propagate through the economy
Roger H. Bezdek and Robert M. Wendling
Roger H. Bezdek is president of Management
Information Services, an economic-research firm.
He received his Ph.D. in economics from the Uni-
versity of Illinois at Urbana-Champaign in 1971
and has worked in academia and for the federal
government. Robert M. Wendling is vice president
of Management Information Services. He received
a master’s degree in economics from George Wash-
ington University in 1977. He has served as senior
economist at the U.S. Department of Commerce,
program manager at the U.S. Department of Ener-
gy and director of the Department of Commerce's
STAT-USA office. Address: Management Informa-
tion Services, Inc., 2716 Colt Run Road, Oakton,
VA 22124. Internet: info@misi-net.com
2005 March–April 133
www.americanscientist.org © 2005 Sigma Xi, The Scientific Research Society. Reproduction
with permission only. Contact perms@amsci.org.
it gradually increased to 27.5 mpg by
1985, the same level as today.
The new regulations had exactly
the intended effect. The fuel efficiency
of new passenger cars rose rapidly
during the late 1970s and reached a
plateau in the early ’80s. Around this
time, car manufacturers had gotten
close enough to the target of 27.5 mpg
that they could concentrate their ef-
forts instead on improving engine per-
formance. Beginning in 1982, while the
mileage leveled off, the average accel-
eration time from 0 to 60 mph (a mea-
sure of performance) began to improve
steadily. The manufacturers also used
advances in technology to “buy” addi-
tional vehicle weight. That is, instead
of continuing to increase the mileage of
their fleet, they kept the fuel efficiency
just above the legal requirement and
gave the American car market what it
really wanted: bigger, faster cars.
However, in recent years an unex-
pected thing has happened: The aver-
age fuel economy for all new vehicles
has declined, from a peak of 26.2 mpg
in 1987 to 24.7 mpg for model year
2004. The reason is a loophole in the
CAFE regulations, coupled with a dra-
matic shift in the tastes of car buyers.
The CAFE standards treated “light
trucks” differently, and more leniently,
than passenger automobiles. Such ve-
hicles were considered to be primarily
for commercial use (though even by the
late 1970s, two-thirds or more of them
served as passenger carriers). Therefore
the CAFE standard for light trucks was
set at 20.7 mpg, where it has remained
through the 2004 model year. (It will
increase to 22.2 mpg by 2007.)
In 1976, shortly after the passage
of the CAFE legislation, sales of light
trucks amounted to less than 20 percent
of all light vehicle sales. But thanks to
the boom in sport-utility vehicles in the
1990s, the light truck category—which
includes pickups, minivans, SUVs,
mini-SUVs and even certain “cross-
over” vehicles like the Chrysler PT
Cruiser—now accounts for nearly 50
percent of all new sales. Manufacturers
have also taken advantage of the strict
wording in the law. By simply mak-
ing the rear seats removable in what
most of us would consider a personal
passenger vehicle, a manufacturer can
reclassify the vehicle as a light truck,
thus exempting it from the stricter stan-
dards for passenger cars.
Not Beyond the Horizon
In 2002, the National Research Council
(NRC), an arm of the National Acad-
emy of Sciences, published a landmark
study of the CAFE standards. The NRC
analyzed technical, safety and related
aspects of the CAFE requirements and
estimated how a variety of feasible
technologies would affect vehicle costs.
We have relied heavily on that NRC re-
port to develop realistic scenarios for
toughening the CAFE standards. Al-
though over the years there have been
other studies of the effects of CAFE on
the national economy, ours is the first
major research effort that was able to
draw on such a detailed and thorough
investigation of the costs of potential
improvements in gas mileage.
The most important finding in the
NRC report is that the technology to
achieve major increases in fuel efficiency
is not somewhere over the rainbow, nor
is it dependent on future research break-
throughs. It exists today. The NRC iden-
Figure 1. Long lines at the gas pump and odd- and even-numbered-day gas rationing were unpleasant facts of life for Americans in the winters
of 1973–74 and 1978–79. These energy crises motivated Congress to pass legislation requiring better fuel mileage for new vehicles, standards
that got progressively tougher from 1978 to 1985. But requirements have not changed since then, in large part because some fear that the impo-
sition of stricter standards would cost jobs. The authors shed light on this issue by modeling various scenarios and predicting their economic
consequences for the future.
Associated Press
134 American Scientist, Volume 93 © 2005 Sigma Xi, The Scientific Research Society. Reproduction
with permission only. Contact perms@amsci.org.
tified more than two dozen technologies,
many of which it labeled as “production
intent.” That is, they are already avail-
able, are well known to manufacturers
and their suppliers, and can be incor-
porated into vehicles as soon as a deci-
sion is made to use them. Some of these
technologies have been used in certain
makes for years. The NRC also identified
a smaller number of “emerging technol-
ogies,” ones that are beyond the research
phase and could be produced widely
within 10 to 15 years. Only two of the
technological improvements discussed
by the NRC (fuel-cell vehicles and series
hybrids) lie beyond the 15-year horizon,
and we did not assume any contribution
from them in our economic model. For
each technology, the NRC identified its
likely cost to the consumer and estimat-
ed the percentage improvement in fuel
economy that it could achieve. There is
no free lunch here: Everything on the
menu comes at a price, from as little as
$8 for low-friction lubricants (a 1-percent
improvement in fuel mileage) up to as
much as $560 for a “camless engine” (an
emerging technology that would save 5
to 10 percent in fuel mileage).
Would it help Americans overall if
the government required the higher fuel
efficiency these technologies make pos-
sible? To estimate the consequences of
changes in CAFE standards, we mod-
eled the economic outcomes of three
scenarios, corresponding to different
levels of governmental intervention. We
based these three hypothetical futures
on the NRC report, which ensured that
they were feasible in terms of technol-
ogy, economics and timing. Indeed, we
were motivated in part by the appear-
ance of the NRC study, as well as our
desire to provide hard data to inform
the contentious CAFE debate.
The base-case scenario (“business as
usual”) assumes no change in CAFE
standards and no increase in fleet gas
mileage. It retains the current distinc-
tion between cars and “light trucks.”
Under this scenario, we assume that
average fleet fuel economy remains
constant through 2030.
Our second version of the future,
which we call the “moderate” sce-
nario, assumes that CAFE-mandated
fuel economy increases by 20 percent
as of 2010, for cars from 27 mpg to 33
mpg and for light trucks from 20.7
mpg to 24.8 mpg. These increases are
phased in from 2005 to 2010 and re-
main at those levels through 2030. We
assume that the lowest-cost, currently
available fuel-efficiency technologies
are implemented (an economically
reasonable assumption) and that the
average vehicle prices increase by
$700 (3 percent).
Finally, our third version, the “ad-
vanced” scenario, pushes the envelope
on the fuel-efficiency gains possible
from current or upcoming technolo-
gies. It assumes that “emerging tech-
nologies” likely to be available by 2015
are implemented. CAFE standards are
increased by 30 percent in 2010—for
cars from 27.5 to 35.8 mpg and for light
trucks from 20.7 mpg to 26.9. By 2015,
the standards rise by 50 percent, to
41.3 mph for cars and 31 mpg for light
trucks. The changes are phased in be-
tween 2005 and 2015 and remain at the
same levels through 2030. In this sce-
nario, average vehicle prices increase
by about $2,700 (12 percent).
We believe that these sorts of gains
in fuel economy are feasible and cred-
ible. They derive from published en-
gineering studies and rely on innova-
tions that are either currently available
or well into research and development.
Neither of our scenarios requires de-
velopment of fundamentally “new”
vehicles or exotic technologies. Even
our advanced scheme is less ambitious
than the original CAFE timetable,
which mandated a 53-percent increase
(from 18 mpg to 27.5 mpg) in just sev-
en years, from 1978 to 1985.
Our two scenarios are comparable
with the proposals that were debated
in Congress in 2002, as well as with the
target set this year by the California Air
Resources Board (CARB). The moder-
ate scenario is less ambitious than all of
Figure 2. Fuel economy is a multidimensional problem, involving trade-offs of vehicle weight, engine performance and vehicle type. The introduc-
tion of the Corporate Average Fuel Economy (“CAFE”) program produced rapid gains in fuel economy of cars and trucks through 1985 (left, top).
Since then, the mileage of both classes has stayed level and slightly above the federal targets (light blue). However, since 1987 the combined mileage
for cars and trucks (green) has declined because of the increasing popularity of sport-utility vehicles (SUVs, right), which are classified as light trucks.
Since the early 1980s, automobile manufacturers have also chosen to use advances in engine technology to improve engine performance (lower left,
green) and to power heavier cars (lower left, blue) instead of improving gasoline mileage. (Source: U.S. Environmental Protection Agency).
2005 March–April 135
www.americanscientist.org © 2005 Sigma Xi, The Scientific Research Society. Reproduction
with permission only. Contact perms@amsci.org.
these regulatory solutions, whereas the
advanced scenario is more ambitious.
For example, Ernest Hollings, the
South Carolina Democrat who chaired
the Senate Commerce Committee, pro-
posed raising the CAFE standard for
passenger cars and light trucks to 37
mpg by 2014; John McCain, the rank-
ing Republican on that committee,
proposed 36 mpg by 2016; and McCain
and Kerry’s bipartisan proposal called
for a standard of 35 mpg by 2015. Al-
though the CARB regulations only set
a target for the emission of greenhouse
gases, they would effectively force au-
tomakers to improve the mileage of
their passenger cars sold in California
by 30 percent, to 36 mpg, by 2015.
By comparing our scenarios with
present and past proposals, we do not
mean to imply that the changes will be
painless. The original CAFE enhance-
ments were obtained by relatively easy
weight reductions and by plucking
other low-hanging fruit. By contrast,
we are now talking about improving
engines and drivetrains that have al-
ready undergone 30 years of optimiza-
tion. Advances will be hard-earned.
And both scenarios will require man-
ufacturers to produce vehicles that
they would not otherwise choose to
make. Consumers will have to accept
these same vehicles, and pay more for
them up front (the savings come later
as drivers pay less for fuel). Even so,
Figure 3. In 2002, the National Research Council (NRC) identified many technologies that could be used to further improve fuel efficiency. These
technologies were divided into two categories: “production-intent” (already suitable for use in mass-produced vehicles) and “emerging technologies”
(expected to be ready for mass production within 10 years). The authors’ economic model uses these estimates as inputs for two different scenarios,
one with moderate increases in the CAFE standards and one with larger increases that would require deployment of some emerging technologies.
136 American Scientist, Volume 93 © 2005 Sigma Xi, The Scientific Research Society. Reproduction
with permission only. Contact perms@amsci.org.
the standards will require a decade or
more to curtail gasoline consumption
significantly, because they apply only
to new cars—they do nothing about
existing cars with poorer mileage.
The cost estimates that we quoted
above are consistent with those of
CARB, which figured that compliance
with its regulations would add $1,050
to the price of a typical vehicle. They
are not consistent with estimates by the
Alliance of Automobile Manufacturers,
which claims that the CARB mandate
would cost about $3,000 per vehicle—a
higher cost for a good deal less im-
provement than our advanced scenar-
io. We believe this estimate is inflated.
Indeed, the auto industry has a long
history of greatly overestimating the
costs of vehicle environmental, safety
and fuel-efficiency improvements. We
chose to use estimates based on the
work of the National Research Coun-
cil, a body we believe to be impartial.
Of course, the results of our economic
model depend critically on these cost
estimates, and those who think that
our figures are too low may disagree
with our conclusions.
Input-Output Models
Before we turn to the results of our
forecast, we would like to explain how
our model works. The main engine,
input-output (or “I-O”) analysis, was
first developed by economists in the So-
viet Union in the 1920s. Input-output
analysis might have ended up in the
dustbin of history, had not Wassily Le-
ontief (1906–1999), a student of this
science, emigrated to the United States
and joined the faculty of Harvard Uni-
versity in 1931. While there, Leontief
adapted input-output analysis to the
U.S. economy. However, in the post-
World War II frenzy of anti-Communist
paranoia, Leontief’s funding was ter-
minated, and such models fell out of
favor. Only in the 1960s did economists
finally recognize I-O analysis as useful
for other purposes than the centralized
“planning” of a socialist economy.
Fortunately, Leontief lived to see
his work vindicated: He received the
1973 Nobel Prize in Economic Sciences
for his pioneering efforts in develop-
ing this method. Today, I-O analysis
is used regularly as a national and
regional economic impact and fore-
casting tool. Probably its most visible
and publicized use includes projecting
the economic impacts of sports facili-
ties. Another prominent use is pro-
jecting the negative impact brought
to communities when military bases
are closed. And anytime you hear es-
timates of what the tourism industry
has brought into the state or local-area
economy and the number of jobs gen-
erated, the projections were almost
surely made by I-O analysis.
An input-output model divides
the national or regional economy into
various industrial sectors and tracks
how much each industry must pur-
chase from every other industry to
produce one unit of output. In a social-
ized economy, a centralized planning
bureau would determine the targeted
level of output for each industry; in a
capitalist economy, market forces in-
stead make that determination. The
trickiest point to understand is that
the model contains feedback loops that
force most industries to produce more
than the “direct output requirements”
would seem to imply. For example, a
demand for x percent more automo-
biles than last year requires y percent
more steel. But steel mills require elec-
tricity to run. And an electric utility
requires turbines from a factory to pro-
duce electricity. That factory in turn
needs steel from steel mills to produce
turbines, and the steel mill requires
more electricity, and so on.
Leontief discovered an ingenious
mathematical method through matrix
inversion that collapses all of these feed-
back loops into one step and calculates
the extra (“indirect”) output require-
ments they create. The ratio of the total
requirements to the direct requirements
is called the input-output multiplier.
To apply our model, we first trans-
lated the increased expenditures for re-
configured motor vehicles (those meet-
ing the revised CAFE standards) into
per-unit output requirements for ev-
ery industry in the economy. We used
these demands to derive the “direct”
output requirements for each industry
and then applied Leontief’s mathemat-
ical formulae to compute the indirect
production needed. Next, we used the
total output requirements to compute
sales volumes, profits and value added
belt
“variator”
solenoid
valve
magnet
coil
Figure 4. Automobile manufacturers are exploring and in some cases have already begun introducing technologies identified in the NRC report.
Continuously variable transmission (left) in the Audi A6 features a “variator” with a wide interlinked chain passing over it. As the plates of the
variator move together or apart, the radius of the chain on both ends—and therefore the gear ratio—changes in a smooth, continuous fashion. An
emerging engine technology is camless valve actuation (center), shown here in a prototype configuration that uses solenoids to open and close
the intake and exhaust valves. (In conventional engines, valves are controlled by a rotating cam shaft, which cannot adapt to changing engine
conditions.) The Honda Accord Hybrid includes an integrated starter-generator (right), which allows the gasoline engine to shut down at stop-
lights instead of wasting gas while idling. (Image at left courtesy of Audi AG; one at right courtesy of Honda Motor Co., Ltd.)
2005 March–April 137
www.americanscientist.org © 2005 Sigma Xi, The Scientific Research Society. Reproduction
with permission only. Contact perms@amsci.org.
for each industry. Finally, using gov-
ernmental data on man-hours, labor
requirements and productivity, we
estimated the number of jobs created
within each industry. (Jobs were the
main focus of our project, but we could
just as easily have estimated the effects
of CAFE on personal income, corporate
profits or government tax revenues.)
For this study, we went into even
greater detail, breaking the effects of
the CAFE standards down by occupa-
tion and geography. The geographic
analysis uses a region-by-region ver-
sion of our national model, which
we can take all the way down to the
county level of detail if needed. (For
this project, though, we went down
only to the state level.) Because of the
comprehensive nature of the modeling
system, the regional analysis uses the
same data and gives results consistent
with the national analysis.
In all, our national model currently
includes 495 industries and 699 occu-
pations. At the state level we used a
more coarse-grained model with 85 in-
dustries, but kept the full occupational
detail. We mention these numbers
solely to give an idea of the scale of the
undertaking. It is very important to
realize that size alone is no guarantee
of a model’s accuracy. Just as impor-
tant, in our opinion, is the quality of
the data that go into the model. We use
data that come directly from U.S. gov-
ernment statistical agencies and very
rarely from other sources. These data
are unbiased, are respected worldwide
and, most important, are gathered in a
comparable fashion across states and
across industries.
Running the Numbers
Our first job was to estimate the over-
all effects of the “moderate” and “ad-
vanced” CAFE scenarios on gasoline
consumption. Do they solve the prob-
lem they are intended to solve? In both
cases, the country uses less fuel than it
does in the base scenario, which is of
course as one would expect. In 2000,
the United States consumed about 125
billion gallons of gasoline. Under the
base scenario, this amount would in-
crease to 150 billion gallons by 2015
and 190 billion gallons by 2030. Under
the moderate scenario, the increases
are smaller: 140 billion gallons by 2015
and nearly 170 billion gallons by 2030.
Clearly, the moderate plan reins in
America’s galloping appetite for oil but
fails to end it. The advanced solution
has more dramatic effects: After rising
through 2010, gasoline consumption
actually begins to drop as the vehicle
fleet is gradually transformed—older
vehicles are scrapped and replaced with
new, more fuel-efficient ones. Eventu-
ally, though, because of the continually
increasing number of vehicles on the
road, the trend turns around. By 2030,
the nation is back to consuming about
130 billion gallons per year, about the
same amount as in 2005. In effect, the
advanced scenario “buys” the country a
25-year delay in its tendency to consume
more gasoline with each passing year.
The reductions in fuel consump-
tion translate into financial savings for
American consumers. The extent of the
windfall depends, of course, on the cost
of gasoline. Because we cannot forecast
gas prices with any accuracy, we used a
range of hypothetical values from $1.25
to $1.75 per gallon (in 2002 dollars). In
the moderate scenario, by 2030, con-
sumers spend $35 billion to $50 billion
less on gasoline than they do in the base
scenario. But the savings come at a cost:
the higher price of the more fuel-effi-
cient vehicles. By 2030, the increased
vehicle cost will be $16 billion per year.
Thus the stricter CAFE standards prove
to be a good bargain, not only in 2030
but throughout the lifetime of the mod-
el. And, if gasoline prices are higher
than we assumed (as they currently
are), consumers will save even more.
The story for the advanced scenario is
a little different. There would be a little
more pain at first for American con-
sumers, because the CAFE standards
do not start paying for themselves un-
til roughly the year 2020. However, by
2030, consumers are comfortably in the
black: They would save $75 billion to
Figure 5. Input-output models track numerous connections between different sectors of the
economy. In this schematic representation, only three out of the 85 sectors actually mod-
eled are shown. Here an increased demand for automobiles (center) will generate increased
production requirements for steel (top). To fill this demand, steel plants will require more
electricity (bottom). To meet this capacity, the electric-power sector will require more steel to
build turbines and power lines, thus setting up a feedback loop. Input-output analysis allows
economists to calculate the effect of such complicated feedback loops.
138 American Scientist, Volume 93 © 2005 Sigma Xi, The Scientific Research Society. Reproduction
with permission only. Contact perms@amsci.org.
$100 billion per year on gasoline, while
spending about $55 billion more per
year on vehicles.
Besides tracking the effect of the
CAFE standard on gasoline consump-
tion, we also wanted to estimate its
effect on jobs. Clearly, substantial job
shifting and displacement would oc-
cur. For example, in 2020, under the
moderate scenario, 101,000 jobs are
created but 72,000 jobs are displaced,
creating a net employment increase of
29,000. Under the advanced scenario,
433,000 new jobs are created by that
year and 86,000 are displaced, result-
ing in a net gain of 347,000. While sig-
nificant, these gains must be put into
perspective: In 2010, U.S. employment
will total 142 million, and in 2030 it
will total 166 million. So the increases
amount to only a fraction of a percent
of the national workforce.
Will It Play in Peoria?
We felt that the overall numbers for
the country did not convey in suffi-
cient detail what would happen to the
economy as a result of the proposed
CAFE changes. The senators who were
debating the changes in 2002 needed
to know what would happen in their
own states, though of course they also
needed to be mindful of the national
picture. And the arguments of auto
manufacturers and organized labor
could best be addressed if we knew
what happened to employment on an
industry-by-industry basis.
We found that the motor-vehicle
and related industries would be major
winners (quite the opposite of the dire
predictions of auto manufacturers),
while employment in the petroleum
industry would suffer. For example, in
the advanced scenario in 2020 (com-
pared, as always, with the business-as-
usual case), jobs in the “motor vehicle
and equipment” sector would increase
by 155,000; jobs in the “rubber” sec-
tor would increase by 22,000; and jobs
in the “electronic component” sector
would increase by 9,500. But jobs in
“crude petroleum and natural gas”
would decrease by 32,000 and “petro-
leum refining” would lose 17,000 jobs.
Our model also allows us to break
down the employment trends by occu-
pation as well as industry. In 2020, under
the advanced scenario, jobs would be
created for 700 computer programmers,
900 mechanical engineers, 1,500 com-
puter-controlled machine tool operators
and 2,700 machinists. On the downside,
petroleum engineering jobs would de-
crease by 700 and petroleum-pump op-
erators would decrease by 6,100. All of
these are net figures; for example, the
net gain of 700 computer programming
jobs results from 1,100 new positions
created and 400 displaced.
Consistent with the above industry-
by-industry results, the places that ben-
efit the most from the enhanced CAFE
regulations are the auto-producing
states of the upper Midwest. Michigan
gains 54,500 jobs, Ohio adds 29,300,
and Indiana receives 22,300. California
also scores an impressive increase of
28,400 jobs, albeit on a much larger
base. Only four states—Louisiana,
Wyoming, Alaska and New Mexico—
suffer net job losses, with Louisiana
losing the most (1,100). Texas, a state
one might expect to be hard-hit by the
loss of petroleum-refining jobs, actu-
ally experiences a modest gain of 2,500
jobs. Its losses in sectors such as “crude
petroleum and natural gas” and “con-
struction” are more than offset by jobs
created in “motor vehicles and equip-
ment,” “fabricated metal products,”
“services” and other industries.
Our results are consistent with simi-
lar studies that have been conducted
over the past three decades. For exam-
ple, in 1980, Douglas Dacy, Robert Kue-
nne and Paul McCoy of the University
of Texas and Princeton estimated the
impact of the original CAFE standards
and projected a net increase in employ-
ment of 140,000 jobs by 1985, with the
jobs projected in various service indus-
tries, plastics, metal stampings and oth-
Figure 7. Authors’ model generates sector-by-
sector predictions of total employment. In the
advanced scenario, most sectors experience a
small amount of growth by 2020. The “motor
vehicles and equipment” sector grows the
most rapidly. Job losses in “petroleum refin-
ing” and “crude petroleum and natural gas”
are smaller in absolute numbers but larger in
percentage terms.
Figure 6. Authors’ input-output model predicts the effect of enhanced CAFE standards on the
economy. As shown in the top graph, annual gasoline consumption continues to increase un-
der the moderate scenario, though not as fast as in the base scenario (no change to CAFE regu-
lations). Only the advanced scenario brings the annual increases to a halt. Both the moderate
and advanced scenarios lead to a net creation of new jobs (bottom), with the effect peaking at
more than 350,000 new jobs in the advanced scenario in 2015.
2005 March–April 139
www.americanscientist.org © 2005 Sigma Xi, The Scientific Research Society. Reproduction
with permission only. Contact perms@amsci.org.
er sectors outweighing projected losses
in steel, petroleum and gas, and whole-
sale and retail trade. In 1989, Arvind
Teotia and his associates at Argonne
National Laboratory estimated the im-
pacts of the use of clean diesel engine
technology in light trucks to comply
with CAFE standards and found that
between 70,000 and 110,000 jobs would
be created. In 1992, the American Coun-
cil for an Energy-Efficient Economy es-
timated that by increasing the fuel effi-
ciency of passenger cars from 28 mpg in
1990 to 40 mpg in 2000 and 50 mpg in
2010, 244,000 additional jobs would be
created by 2010. A 2001 Union of Con-
cerned Scientists study analyzed the
economic effects of increasing CAFE
standards to 40 mpg by 2012 and to 55
mpg by 2020, and projected an increase
of 104,000 jobs by 2020.
Our analysis shows that enhanced
CAFE standards would increase overall
employment in the United States, but it
is still conceivable that there could be
a net decrease in union jobs. Indeed,
labor unions such as the United Auto
Workers are concerned that the new
standards would exacerbate the shift
in production toward imports and ve-
hicles produced in U.S. factories owned
by foreign companies (“transplant”
facilities, which are rarely unionized).
They might be right. Nevertheless, there
are reasons to question the presumed
inverse correlation between union jobs
and CAFE standards.
First, union representation in the
motor-vehicle industry has been de-
clining for two decades, even during
the 1990s, which was one of the most
robust and profitable decades in his-
tory for domestic vehicle manufactur-
ers, whereas CAFE standards have not
changed since 1985. Organized labor’s
problems in the industry thus appear to
be deep-seated and cannot be attributed
to CAFE. Second, foreign manufactures
are rapidly improving their technol-
ogy, largely because their main markets
are in countries with high fuel prices
or high fuel-economy standards. To
the extent that stricter standards force
domestic manufacturers to adopt new
technology, enhanced CAFE could actu-
ally improve their competitiveness—and
thus preserve and expand union jobs.
Although we are not in a position
to offer any firm answers about this
particular issue, our I-O modeling rec-
ognizes that a substantial portion of
U.S. vehicle expenditures are made on
imports and create sales and jobs in-
ternationally. Imports will continue to
be sold in the United States, and those
countries that cater to our markets
can expect positive economic gains,
including additional jobs. Of course,
U.S. vehicles and parts sold in Europe
and Asia also create jobs—both union
and nonunion—here at home.
In summary, our modeling leads us
to an unambiguous conclusion: An in-
crease in CAFE standards would save
consumers money in the long run,
would not harm the U.S. economy and
would in fact lead to a significant net
increase in jobs, mostly within the first
five or ten years. A moderate improve-
ment in vehicle mileage will not re-
duce America’s consumption of oil,
although it will slow down the rate of
increase. To actually reduce consump-
tion, the government would have to
institute considerably stricter—though
still feasible—standards, amounting to
a 50-percent improvement in mileage
by 2015. Finally, whereas the overall
economic impact on the national scale
is positive, the benefits would not be
uniformly distributed. A few indus-
tries and states would see net losses
in employment, and job shifts would
occur even within industries, occupa-
tions and states that added jobs.
It’s time we put CAFE reform back
on the national agenda. We believe
that the results of models like ours
can influence the country’s decisions
about fuel efficiency, jobs and the econ-
omy, thus moving the debate beyond
rhetoric and political posturing. The
nation’s transportation system, petro-
leum needs and the jobs of hundreds
of thousands of Americans are at stake.
Reliable information and objective
analysis are required. One would hope
that the work we’ve summarized here
will improve the rigor and quality of
this critically important debate.
Acknowledgments
The work summarized here was supported, in
part, by the Energy Foundation.
Bibliography
Bezdek, R. H, and R. M. Wendling. 2005. Po-
tential long-term impacts of changes in U.S.
vehicle fuel efficiency standards. Energy
Policy 33:407–419.
Dacy, D. C., R. E. Kuenne and P. McCoy. 1980.
Employment impacts of achieving auto-
mobile efficiency standards in the United
States. Applied Economics 12:295–312.
Friedman, D., J. Mark, P. Monahan, C. Nash
and C. Ditlow. 2001. Drilling in Detroit: Tap-
ping Automaker Ingenuity to Build Safe and Ef-
ficient Automobiles. Cambridge, Mass: Union
of Concerned Scientists Publications.
National Research Council. 2002. Effectiveness
and Impact of Corporate Average Fuel Economy
(CAFE) Standards. Washington, DC: Nation-
al Academy Press.
Figure 8. In a state-by-state breakdown, job increases in the advanced scenario are greatest in
the upper Midwest and California (light green). Only four states (tan) experience net decreases
in employment, and the decreases are much smaller than the increases elsewhere.
For relevant Web links, consult this
issue of American Scientist Online:
http://www.americanscientist.org/
IssueTOC/issue/701
