The fuel economy of cars and light trucks (light vehicles) rose dramatically after 1973, peaking in 1987-88. Oil prices plunged in 1986, squelching market interest in fuel economy. Corporate Average Fuel Economy (CAFE) standards helped drive the increase in fuel economy, but standards have not been meaningfully raised since the mid-1980s. By 1993 most older, less efficient vehicles had essentially been replaced, particularly in terms of annual usage. Because fuel economy improvement has ceased, light vehicle fuel use is again growing at the same rate as the amount of driving, which is expected to increase 50% over the next two decades. The United States has sent more than a trillion dollars overseas for oil imports, equal to 70% of the cumulative trade deficit over the past two decades. Light vehicle fuel use accounts for 21% of U.S. carbon dioxide emissions. A major portion of hydrocarbon emissions is also directly related to gasoline use. These problems will continue to grow unless new vehicle fuel economy is substantially improved.

An understanding of the opportunities for cost-effectively improving new car fuel economy underpins the development of balanced policies for controlling light vehicle fuel use. A number of recent studies address this question. Estimates of the potential fuel economy of the new automobile fleet for the year 2001 range from 28 mpg (essentially no improvement over recent levels) to 45 mpg. Disagreements can be traced to divergent assumptions about the benefits, costs, applicability, and marketability of the technologies considered. Published estimates for improvements over the near-term (roughly 10 years) are limited in that only existing technologies are considered. The primary assessments are the studies by Energy and Environmental Analysis, Inc. (EEA), sponsored by federal agencies, and studies by auto industry consultants such as SRI. The National Research Council (NRC) study of 1992 drew mainly on these sources. The technologies included in these data bases are now already five or more years old; newer technologies and further refinements of the existing ones are not fully included.

This analysis considers more widespread use of technologies already in production plus the introduction of emerging technologies. Our review is organized as a menu of options, grouped under major headings representing the engine, transmission, and tractive load aspects of vehicle design. While this discrete approach is convenient for analysis, in reality engineers take a much more integrated approach to design. In fact, the creativity of engineers and designers continually refines and expands the menu of options which can be used to increase vehicles' efficiency and improve them in other ways as well. To both capture the integrated nature of technology refinement and check our results, we also apply an engineering model to perform an integrated analysis of efficiency improvements to a typical vehicle.

We base our assessment on the technology status of the new car fleet in 1990, which is taken as the base year for the analysis. We consider technology improvements that will improve fuel economy while maintaining the same average vehicle size and performance as in 1990. Available cost information is reviewed and technologies are screened according to cost-effectiveness, considering the fuel savings to all consumers over an average vehicle lifetime. We examine contemporary auto industry product cycles, development times, and rates of technology change, obtaining an estimate that 8-11 years of lead time are needed to achieve full penetration of the efficiency improvements. Given the late 1993 timing of this report (model year 1994 has started), this implies that the industry can achieve the estimated degree of fuel economy improvement by model years 2002-2005. There have undoubtedly been increases in the use of some of the technologies since the 1990 base year assumed here. However, these technology improvements have not been directed toward fleetwide fuel economy improvement. Thus, achieving the vehicle efficiency increases estimated here could involve not only incorporation of new technology but also a redirection of existing technology applications. This suggests that the feasible improvements could actually happen more quickly or at lower cost than estimated here.

The resulting estimates of potential fuel economy improvement are presented in Box S1 (below). Reflecting the uncertainties surrounding new applications of technology, we present our results at three levels of technical certainty:

• Level 1 includes technologies already in use in at least one mass market vehicle worldwide and which have no technical risk in that they are fully demonstrated and available;
• Level 2 incorporates measures which are ready for commercialization and for which there are no engineering constraints (such as emissions control considerations) which inhibit their use in production vehicles but which entail risk in that some "debugging" may be needed because of limited production experience;
• Level 3 technologies are those in advanced stages of development but which may face some technical constraints before they can be used in production vehicles.

Back to index

Box S1.

                                                 Technology Certainty:
                                           Level 1      Level 2     Level 3

    Achievable, Cost-Effective MPG            40           46          51

 Average Added Cost per Car (1993$)          590          770         840

  Average Cost of Conserved Energy           0.55         0.53        0.51
                            ($/gal)

    Potential Savings in 2000 (Mbd)          0.4          0.5         0.6

    Potential Savings in 2010 (Mbd)          2.1          2.8         3.2

In order of increasing technical uncertainty, the resulting estimates of achievable new car fleet average fuel economy are 40 mpg, 46 mpg, and 51 mpg.These values correspond to improvements of 43%, 65%, and 85%, respectively,over the 1990 base year average of 28 mpg. We also performed sensitivityanalyses to investigate assumptions regarding fleet average accelerationperformance and technology penetration. Increasing performance to the 1993average lowers projected fuel economy by about 1 mpg; decreasing performance tothe 1987 average raises it by about 1.5 mpg. There is a smaller sensitivity tothe degree of technology penetration within the range considered. No change inaverage vehicle size is needed for the technology-based fuel economyimprovements analyzed here.

While much judgment is clearly involved in policy development, we believethat our Level 2 estimate--a new fleet fuel economy improvement of 65% by2002-2005--provides a reasonable target for public policies intended toincrease automotive fuel economy. More ambitious targets might be justifiedunder our Level 3 assumptions, since policy guidance can hasten the developmentand application of advanced technologies which have the potential forwidespread commercialization.

Two types of checks corroborate the fleet-average technology penetrationanalysis used to obtain our summary results: simulation analysis ofimprovements for a typical vehicle and comparison to fuel economy levelsactually achieved in a particular car. Applying an engineering model relating fuel consumption to vehicle tractive loads and engine performance for standard driving cycles enables us to simulate the effect of technologies on a specific vehicle. Taking a 1991 Ford Taurus as an example, we analyze a set of Level 2 technologies applied to reduce vehicleloads and decrease engine friction. Figure S1 illustrates energy losses in the current vehicle (lighter bars) and the reduced losses in the improved vehicle(darker bars). The result is a 43% cut in fuel consumption per mile, implyinga 75% improvement in fuel economy. Thus, applying technologies for tractiveload reduction, engine specific power enhancement, and optimized transmissioncontrol raises the base vehicle's fuel economy from 27 mpg to 47 mpg, anincrease just exceeding the fleet average we estimate at Level 2 certainty.Incorporating Level 3 technologies, such as lean-burn or two stroke engines anda greater degree of tractive load reduction, would permit an even greaterimprovement, to in excess of 50 mpg.

The 1992 Honda Civic VX provides a concrete example of fuel economy levelsin the ranges we estimate. The lean-burn version has a composite unadjustedfuel economy of 60 mpg, slightly higher than the 58 mpg obtained by applyingour Level 3 estimates to the 1990 subcompact fleet average of 31.5 mpg. TheCalifornia version of the Civic VX has a fuel economy of 55 mpg, also higherthan the 52 mpg average implied for subcompacts by our Level 2 technologyestimates (without lean-burn engines). Although the Honda Civic VX is not anaverage car (as in the modeled Taurus example), it demonstrates substantialfuel economy improvements over a comparable Civic hatchback without reducingsize or performance and using only some of the technologies reviewed here.Moreover, its improvements were already achieved, over one 4-year productcycle, while our projections allow 8-11 years of lead time for the fleet as awhole.

Accurately estimating the cost of improving fuel economy is difficultbecause of limitations in publicly available data and costing methodologies.Our technology cost estimates are derived largely from previously publishedinformation, such as studies by Energy and Environmental Analysis (EEA) andother sources. Costs for the Level 2 technologies are listed in Table S1.These estimates represent the incremental costs of improved vehicle technology,assuming the use of a mature technology averaged over a total production periodand without premature replacement of production facilities. The resultingaverage per-car incremental retail price estimates are $590, $770, and $840(1993$) for technology certainty Levels 1-3, respectively. The applicabilityof these cost estimates depends on assumptions regarding industry productcycles and other factors which affect the economics of motor vehicleproduction. In particular, costs are linked to lead time, since we estimatedlead time sufficient to validate the assumption of no premature replacement ofproduction facilities.

Given the above caveats, the estimated costs of fuel economy improvementare quite modest, in the range of 3%-5% of the average cost of a new car.These estimates are corroborated by the historical experience of pasttechnology-driven fuel economy improvements, of which retrospective analyseshave observed cost increases of roughly 5% of average new car price. While theestimates reported here are affected by industry economic factors which we arenot able to address, this larger uncertainty cuts both ways: while it possiblethat actual costs of making the fuel economy improvements identified here couldbe higher than estimated, it is also possible that the costs could be lower,particularly as experience is gained and opportunities arise for finding costsavings in the course of product development.

Annual fuel costs for an average new car are roughly $500 per year atcurrent fuel prices which, adjusted for inflation, are as low as they have everbeen. Thus, although market interest in fuel economy is low, improving fueleconomy is quite cost-effective to consumers, with an average payback time ofless than four years. In reporting that the efficiency-related technologyimprovements identified here are cost-effective, we are not saying that theywould necessarily be salable under today's market conditions. Much moreefficient vehicles could be sold under changed conditions which might bebrought about by various factors, such as national policies (fuel economyregulation, vehicle pricing incentives, or dramatically higher fuel taxes) orinternational events (wars, oil supply cartel decisions). Thus, policies toencourage or require efficiency improvement would change market conditions soas to lower the risk of applying technologies for efficiency improvement. Inthis regard, we distinguish the concerns of citizens from the concerns ofconsumers: citizens can collectively decide that higher fuel economy is neededto address problems of national concern and therefore support policy changes toraise fuel economy above the market level which they (and the auto industry)decide when acting as individual consumers.

The technology benefit, cost, and penetration estimates can be used toconstruct supply curves of potential fuel economy improvement and gasolinesavings, as given in Figure S2 and Table S1. Figure S2(a) plots potential newcar fleet average fuel economy against the Cost of Conserved Energy (CCE),expressed in 1993$ per gallon. The CCE is based on the ratio of incrementaltechnology cost to fuel savings discounted with a 5% real rate over a 12-yearvehicle life. It is an index of cost-effectiveness from the perspective ofconsumers in aggregate (all owners over the car lifetime rather than only thenew car buyer). Figure S2 gives costs under our Level 2 assumptions; similarcurves at other technology certainty levels are presented in Figure 7 of thereport. Each step in Figure S2 represents one of the technologies considered in ouranalysis, showing its incremental benefit for fuel economy improvement and itsmarginal cost expressed as an equivalent cost of avoided gasoline consumption.Steps are numbered by technology as listed in Table S1. For example, step 8 isvariable valve control (VVC), which offers an efficiency benefit of 12% andwould save 580 gallons of gasoline over an average vehicle lifetime. The costfor VVC is equivalent to having to pay only $0.46/gallon for this saved fuel,shown as the CCE level for step 8 in the figure. Technologies arecost-effective if their CCE is lower than the future price of gasoline expectedover the life of the improved vehicles, which we assume to be $1.65/gallon(1993$). The bottom part of Figure S2 shows the nationwide gasoline savings andgreenhouse gas emission reductions in 2010 for each increment of new vehiclefuel economy improvement achieved by 2005. This graph assumes proportionateefficiency improvements in light trucks and expresses the CCE as a crude oilprice equivalent, adjusting for the differences between oil prices and retailgasoline prices. Thus, savings of 2.8 million barrels per day (Mbd) can beobtained at a cost of just under $33 per barrel, roughly the oil priceprojected for 2010 by the U.S. Department of Energy. These savings wouldamount to a one-third cut in U.S. light vehicle fuel consumption, expected tootherwise reach 9 Mbd by 2010. The corresponding reduction in greenhouse gas emissions would be 27 millionmetric tons per year (MTC/yr) in 2000 and 140 MTC/yr in 2010 (full fuel cycleC-equivalent emissions expressed on a carbon mass basis). Achieving thislevel of new car fuel economy improvement would thus provide an 8% cut in U.S.C emissions otherwise expected for 2010, avoiding 38% of the projected growthin U.S. C emissions over 1990-2010. The cost of C emissions reduction iszero for fuel economy improvements having a CCE up to the avoided cost of fuelconsumption ($33/bbl in 1993$, equivalent to retail gasoline at $1.65/gallon).For modest levels of fuel economy improvement lower than the fullycost-effective level, greenhouse gas emissions reductions can be achieved atnet savings. Of the gasoline consumption and C emissions reductions estimated here,60% are from the improvements in passenger car fuel economy specificallyanalyzed in this report. The remainder are from proportionate improvements inlight truck fuel economy, which we believe are similarly feasible andcost-effective although a detailed analysis has not been done by ACEEE. The report also addresses the relationship between investments needed toimprove fuel economy and issues such as market risks and competitive factors inthe auto industry. Although not all firms are equally strong in all areas,competition induces ongoing enhancements of every firm's ability to respond to evolving market conditions. To meet changes in market conditions--be theyinduced by consumers, the world oil market, the government, or theircompetitors--a firm depends on its ability to develop quality products on atight schedule, to retool quickly, and to execute flexible, "lean," productionprocesses. An aspect of the advancing production efficiency includesrelationships among competitors in the industry, such as joint ventures,product sharing, and outsourcing of components to competitors as well as tospecialized suppliers. Thus, the issue of fuel economy improvement is largelyone of how the industry's substantial, competition-driven capabilities aredirected. In the absence of market signals or government policies to directadvances toward improving fuel economy, the industry's energies have recentlybeen directed toward greater performance, luxury, and product differentiation,some of these coming at the expense of fuel economy. We find no inherent reason why the industry's capabilities could nototherwise be channeled, with little change in risk or cost, given marketsignals or government policies pointing toward efficiency improvement. Givenadequate lead time and balanced policies that provide equitable treatment offirms in the U.S. market, the 43%-85% improvements in conventional vehicle fueleconomy identified here can be reached without added market risk and at modestper-vehicle cost, with overwhelming benefits in terms of fuel savings andavoided oil import and environmental costs over the life of the improvedvehicles. Our study shows that a number of technologies, implemented throughout thefleet to varying degrees, can yield a range of new car fuel economy levelsconsiderably higher than those of today. There is a rich array oftechnological approaches for improving fuel economy, so that automakers neednot count on the availability of only one circumscribed set of engineeringoptions for reaching modest or intermediate levels of new fleet average fueleconomy. The potential availability of less certain technologies, e.g., thoseidentified here as Level 3 technologies, reduces the risk for reaching low orintermediate levels of fleet wide fuel economy improvement. Thus, there aremultiple ways by which the new car fleet could evolve to reach, say, ourLevel 2 achievable potential of 46 mpg. Different approaches might, in fact,be taken by different manufacturers. It is important to emphasize the conservatism of the results presentedhere, which rely solely on incremental improvement of vehicles based ongasoline-burning internal combustion engines, without radical changes in eitherdesign or manufacturing technique. We do not consider the potentially dramaticimprovements in fuel efficiency that could be achieved through the use ofhybrid drivetrains for efficient power management, net-shaping of compositebody structures, along with advanced computer-aided design, manufacturing, andengine/transmission control technologies. The use of such approaches forautomotive design has already reached the prototype stage, and could well beused for commercial production within a decade. Policy impetus for achievingimprovements in new car fuel economy would do much to stimulate thecommercialization of these more advanced technologies. In summary, our review indicates that there is a wide array of availableand near-commercial technologies which can be applied to improve automotivefuel economy over the next decade. Improving new cars to the mid-range(Level 2) estimate of 46 mpg by 2005 and improving new light trucksproportionally would cut U.S. gasoline consumption by 2.8 million barrels perday and reduce oil imports by at least 2 million barrels per day in 2010.There would be corresponding annual cuts of 140 million tons of greenhouse gasemissions and nearly 500,000 tons of hydrocarbon emissions. This degree offuel economy improvement would add about $770 to the price of an average newvehicle. The overall annual cost increase in the new vehicle market would gradually rise to as much as $11 billion. Up-front investment costs by theauto industry will occur sooner but would be only a fraction of the overallretail cost increase. These costs are quite modest compared to annualexpenditures of over $200 billion in new light vehicle purchases. Viewed as anational investment, fuel economy improvement is very cost-effective, with thegasoline cost savings reaching $70 billion per year by 2010 and continuing torise thereafter. The enhanced economic growth from re-spending of thesegasoline cost savings would increase net U.S. employment by nearly 250,000 jobsby 2010, including nearly 50,000 new jobs in the auto industry. In short, thelarge benefits to the nation--direct consumer savings, lower oil imports,reduced hydrocarbon and C emissions, and job creation--indicate that fuel economy improvement is one of the best investments the country can make.

Back to index

Table S1. List of technologies, fuel economy benefits, and costs for mid-range (Level 2) estimates of technology certainty. *

Average Est. Fleet New Fleet Savings

Technology MPG unit avg MPG Car CCE ACE avg

benefit cost increase MPG $/gal $/gal cost in

2010

(Mbd)

1. Compression ratio increase 1.0% $ 0 1% 28.2 0.00 0.00 $ 0 0.08

2. Lubrication improvements 0.5% 2 2% 28.4 0.11 0.03 2 0.13

3. Lower tire rolling 4.8% 22 7% 29.8 0.12 0.10 24 0.46

resistance

4. Continuously variable trans. 6.0% 33 10% 30.6 0.15 0.12 37 0.63

5. Optimized manual 11.0% 66 12% 31.3 0.18 0.13 51 0.77

transmission

6. Optimized transmission 9.0% 66 19% 33.2 0.24 0.17 99 1.14

control

7. Accessory improvements 1.7% 14 21% 33.8 0.30 0.18 112 1.23

8. Variable valve control 12.0% 140 32% 36.8 0.46 0.26 232 1.71

9. Variable displacement 5.0% 70 34% 37.4 0.61 0.28 260 1.81

10. Overhead cam 3.0% 44 37% 38.0 0.64 0.29 284 1.90

11. Weight reduction 9.9% 160 47% 40.9 0.79 0.37 449 2.29

12. Friction reduction 6.0% 110 53% 42.4 0.97 0.42 536 2.47

13. Four valves per cylinder 6.6% 120 57% 43.9 1.03 0.46 621 2.64

14. Torque converter lockup 3.0% 60 58% 44.0 1.17 0.46 623 2.65

15. 5-speed automatic 5.0% 120 60% 44.6 1.42 0.48 667 2.72

transmission

16. Aerodynamic improvements 3.8% 100 65% 45.9 1.59 0.53 766 2.85

17. Multipoint fuel injection 3.0% 80 66% 46.2 1.73 0.53 784 2.88

18. Super-/turbo- charging 5.0% 180 70% 47.3 2.27 0.59 903 3.00

19. Idle off 6.0% 290 74% 48.3 3.19 0.67 1046 3.09

Average MPG benefit is for the technology applied to an individual car withLevel 2 assumptions, as given in Table 1 of the report. Estimated unit cost and fleet average cost increase are based on Table 4 but given in 1993$ (using a GDP inflator of 1.10 to update from 1990$ to 1993$).Fleet average MPG increase is cumulative, based on an average of the High andFull penetration assumptions given in Table 2(b), and reflects an interpolatedoptimization factor to account for the multiplicative interaction of loadreduction and drivetrain measures (based on Table 3).Marginal (CCE) and average (ACE) cost of conserved energy are based on 5% realdiscount rate and 12-year, 10,000 mi/yr lifetime; CCE and ACE values would be30% higher using a 10% discount rate.Nationwide gasoline savings in million barrels per day (Mbd) in 2010 assume thegiven percentage MPG increase is achieved in new cars and light trucks by 2005and are calculated relative to new fleet fuel economy frozen at the 1990 levelof 25.2 mpg, using a fuel economy shortfall of 20%, a cost of driving("rebound") elasticity of 10%, and total light duty Vehicle Miles of Travel(VMT) of 2.748 x 1012 miles/year in 2010.

99 pp., 1993, $17.00/T932

Back to: Transportation