Conservation Supply Data for Three Transportation Sectors
This appendix provides information on the calculation method used to determine cost-effectiveness for three transportation sectors: light-duty vehicles, heavy trucks, and aircraft.
Light-duty vehicle efficiencies were emphasized in Chapter 23 as the largest and most thoroughly studied transportation sector. Table E.1 shows the amount of fuel used by each type of vehicle for different modes of operation. As discussed in Chapter 23, the long trend to reduce operating costs via technology improvements while maintaining or improving other vehicle attributes is shown in Figure 23.1b. The fuel economy index (FEI), the product of vehicle mass and fuel economy in miles per gallon, controls for the fact that vehicle mass increased throughout the interval shown on the left-hand side of Figure 23.1b. This parameter, used to judge passenger cars for many decades, is a better indicator of powertrain efficiency than fuel economy alone. In the last decade the trend in the FEI, having the same units but measured at a differently specified test condition, is shown increasing at a similar rate. The recent trend was maintained, however, in a period of decreasing car mass and changing market demands for increased performance.
Research on the knocking properties of fuel in 1913 by Ricardo and later by Kettering provided the basis for many of the gains through 1970 (Amann, 1989). In recent times, applications of new computer technology to engine control, and applications of refined design techniques and new materials for weight reduction, have led to improved fuel economy (see Table E.2).
As the hedonic models of Atkinson and Halvorsen (1984, 1990) show,
TABLE E.1 Transportation Energy Use by Mode, 1987
(continued on page 729)
(Table E.1 continued from page 728)
consumers choose vehicles as a bundle of attributes that include style, comfort, performance, safety, fuel economy, and price. By definition, externalities associated with pollution and some safety issues are not a component of this bundle. Atkinson and Halvorsen calculate the demand elasticities for the attribute of personal safety, however, and their estimate of the revealed preference for this attribute provides a value of life ranging between $2.4
million and $6 million. Similarly, their results indicate that a significant number of consumers place a great value on performance.
Safety and performance are directly related to the mass and power of a vehicle, and fuel economy is inversely related to these variables. Although the literature on fuel economy (Bleviss, 1988) identifies vehicles having good performance (11 seconds, 0 to 60 mph) and exceptional fuel economy (81 highway miles or 63 city miles per gallon (mpg)), these prototypes are not subject to price or production constraints. Given a price constraint and the laws of physics, the consumer is forced to trade off desired attributes against one another.
TABLE E.2 Comparison of Vehicle Fuel Economy Technology Estimates
When forecasting cost-effective greenhouse gas reductions for future years, several uncertainties should be recognized. Figure E.1 makes clear that, given consumer preferences, the relative sizes of the automobile and lightduty truck markets are highly interactive and there has been a tendency to shift toward less efficient light-duty trucks. In addition, the number of person-miles traveled by each sector depends strongly on fuel price and other product attributes. On-road fuel economies are typically lower than those predicted by the EPA (Environmental Protection Agency) fuel economy test procedure as they depend on differences in highway speed, congestion, urban-rural travel mix, and average trip length. To provide a generous estimate of greenhouse gas reductions, a relatively low composite fuel economy baseline of 19.7 mpg will be used for the combined fleet of automobiles and light-duty trucks. Travel projections from the MOBILE3 Emission Model for the year 2000 were used (the MOBILE3 model is an emissions planning document that has been used by EPA to project vehicle miles traveled, emission levels, and grams per mile).
The CO2 emitted from the tailpipe of a vehicle must be adjusted for three additional global impacts. The first is the fact that other greenhouse gases such as CH4 and N2O often accompany CO2 emissions from the transportation sector. In addition, the processing and transportation of these fuels introduce greenhouse gases into the atmosphere.
In the case of gasoline, for every 311 g of CO2 emitted, approximately 78 g of CO2 equivalent is emitted as CH4 and 38 g as N2O. In addition,
CO2 emissions and venting during the processing and distribution of gasoline bring the total greenhouse gas emissions to between 455 and 507 g CO2 equivalent (see Tables 3 and 6 of Unnasch et al., 1989). The CO2 emission values in each table provided in this appendix are therefore multiplied by 1.55 to obtain CO2 equivalence for both the cost-effectiveness and the emission reduction axis in each figure.
Ten years ago, a fuel economy technology plan for automobiles and light-duty trucks was designed by the Energy Productivity Center of the Mellon Institute. The data from this study have been used in the calculations summarized in Table E.3 (Shackson and Leach, 1980) and plotted in Figure E.2 using the least-cost supply curve framework proposed by the Lawrence Berkeley Laboratory (Wright et al., 1981).
For each technology, an estimate of consumer purchase cost in 1990 dollars was used as the numerator in the cost-effectiveness ratio. Reductions in fuel consumption produced by each technology were used to estimate the corresponding 10-year benefit stream for the denominator (i.e., reduced greenhouse gas emissions in tons of CO2 equivalent).1 Because the benefit stream is directly proportional to vehicle usage, which falls rapidly with age, it is appropriate to discount the benefit terms for each interval between the time of purchase and the time benefits are realized. The effect of discounting is to depreciate future benefits and thereby raise the calculated cost-effectiveness values. For the cost-effectiveness calculations in this study, the CO2-equivalent emissions avoided were discounted at 3, 6, 10, and 30 percent, and were used in plotting the four curves in Figure E.2. The discount rates of 3, 6, and 10 percent represent a societal perspective, while 30 percent is closer to the discount rate chosen by a consumer when selecting a vehicle.
Because the technologies save fuel, the cost-effectiveness values were credited with $1.00/gal for discount rates of 3, 6, and 10 percent and $1.25/gal for the 30 percent ratehence the negative values on the conservation supply curves.
The horizontal axis in Figure E.2 represents an estimate of the cumulative annual reductions in greenhouse gas emissions as though each device were employed in the 1989 fleet. These quantities were not discounted (see Table E.3). The 24 technologies analyzed by Shackson and Leach are ordered by their cost-effectiveness in Table E.3hence the monotonic increase in the curve in Figure E.2.
As mentioned above, many of these most cost-efficient technologies were introduced, in part, during the 1975 to 1982 vehicle production era. The most expensive were not offered on the market by the industry. As the industry moved along the learning curve throughout the decade, old technologies became available at lower cost, and as the cost of fuel increased, new technologies became attractive to the consumer. Furthermore, these
TABLE E.3 Automobile and Light Truck Data (Shackson and Leach, 1980)
(continued on page 734)
(Table E.3 continued from 733)
(continued on page 735)
(Table E.3 continued from page 734)
(continued on page 736)
(Table E.3 continued from page 735)
NOTES: Cost in 1990 = Cost in 1979 × Consumer Price Index (1990)/CPI (1979) = 130/72.6 × D (cost to consumer in 1979).
Cumulative improvement = the sum of (penetrations × efficiency gains).
Cumulative cost = previous cost + new component cost × %penetration.
Fuel economy = (1 + %improvement) × 19.7.
Annual fuel savings = (1,889.28 billion miles traveled in 2000) × [(1/19.7) - (1/mpg)].
Tons of CO2 equivalent avoided = 0.00895 × 1.55 × annual fuel savings.
At 3%, marginal tons of CO2 equivalent for change in mpg = .00895 × 1.55 × ten years of vehicle miles traveled (discounted at 3%) × delta reciprocal mpg = .00895 × 1.55 × 92383 × ((1/mpg(i) - 1/mpg (i - 1)).
At 6%, marginal tons of CO2 equivalent for change in mpg = .00895 × 1.55 × ten years of vehicle miles traveled (discounted at 6%) × delta reciprocal mpg = .00895 × 1.55 × 81530 × ((1/mpg(i) - 1/mpg(i - 1)).
At 10%, marginal tons of CO2 equivalent for change in mpg = .00895 × 1.55 × ten years of vehicle miles traveled (discounted at 10%) × delta reciprocal mpg = .00895 × 1.55 × 69980 × ((1/mpg(i) - 1/mpg(i - 1)).
At 30%, marginal tons of CO2 equivalent for change in mpg = .00895 × 1.55 × ten years of vehicle miles traveled (discounted at 30%) × delta reciprocal mpg = .00895 × 1.55 × 20560 × ((1/mpg(i) - 1/mpg(i - 1)).
Corresponding $/ton of CO2 = marginal cost/marginal tons of CO2 equivalent - fuel credit @ $1.00 per gallon = (change in cumulative cost/marginal tons of CO2 equivalent) - 111.73/1.55.
aAccess. Ld. = accessory load reduction; Red. Roll. = reduced rolling resistance; Impr. Man. = improved manual transmission; Des. Param. = engine design parameters; TorqLoc=torque converter lock-up; Aero. Adds = aerodynamic add-on equipment; Eng. Des. = engine design parameters; 4Sp. Auto. = four-speed automatic with torque converter lock-up; DISC = diesel and direct injected stratified charge engines; and Oper. Par. = engine operating parameters.
technologies became attractive to larger segments of the market and their penetration was increased.
Such a portfolio of technologies, similar to that proposed by Shackson and Leach, has been proposed by DOE (Difiglio et al., 1990). The bulk of the fuel economy gains proposed by DOE are achieved by introducing known technologies to most models in the fleet.
While cautioning the reader on the firmness of the benefits and costs of the DOE portfolio of technologies, Ledbetter and Ross (1989) have also used a supply curve framework for the analysis of these data. The conservation supply curve data (utilizing the cost and fuel economy values for 17 technologies in Table 5 of Ledbetter and Ross) have been used to generate the curve in Figure E.3 for the four different perspectives of this study (see Table E.4).
Comparison of the cost-effectiveness values in Figures E.2 and E.3 shows significant differences. The differences do not result as much from differences in the technology portfolios as from differences in the estimates of the costs and benefits of individual technologies. The disagreement between those who design and build cars and those who have generated the DOE data lies in the estimation, measurement, and aggregation of the fuel economy gain made possible by the technologies themselves when their
interactions are taken into account (see Berger et al., 1990). Industry analyses suggest that the potential for fuel economy improvement solely through diffusion of existing technology is less than one-half that predicted by DOE (see estimates of technology gains listed in Table E.2).
Specific differences in estimates for engine efficiencies are driven primarily by two factors. The first has to do with performance parameters chosen by engineers doing the analysis. Although fuel economy at constant horsepower seems like a reasonable criterion for comparison, it turns out that domestic consumers are more interested in vehicles that deliver torque at low engine speed. The newer engines being proposed by DOE for introduction will deliver significant gains at constant horsepower; however, the gains are diminished when compared at constant torque.
The second factor involving engines has to do with how technologies are bundled and then labeled. While one technology at a time can produce a gain of a certain magnitude, systems compound in a way that is not always additive (Berger et al., 1990).
This phenomenon has been demonstrated for vehicles on the market in a statistical analysis of EPA test data from virtually all of the cars in the model year 1988 and 1989 fleets (Bussmann, 1989). The statistical study
TABLE E.4 Automobile and Light Truck Data, Department of Energy (Ledbetter and Ross, 1990)
(continued on page 740)
(Table E.4 continued from page 739)
(continued on page 741)
(Table E.4 continued from page 740)
predicts a 5.4 percent fuel economy gain with technologies for which the Office of Technology Assessment and DOE predict a 17.3 percent gain. The major difference between these studies is that the 17.3 percent estimate is made by summing the gains from 12 efficiency technologies considered individually and the 5.4 percent estimate is made by a model that considers the system of these same technologies.
Other uncertainties in many parameter estimates lead to significant differences in the cost-effectiveness values in Figures E.2 and E.3. In addition to differing estimates for fuel economy gains, if one allows for technology interactions, differing estimates of consumer preferences and differing methods of allocating costs could underlie the observed uncertainties.
Approximately 20 percent of the fuel expended on the U.S. highway system is utilized by heavy trucks and other heavy-duty vehicles, as indicated in Table E.1. The opportunities to conserve fuel and minimize emissions parallel those for light-duty vehicles and light-duty trucks. The method of analysis in this section parallels the methods used in the preceding section. Despite the fact that the data base for these vehicles is relatively meager, it will be possible to create conservation supply curves for three important categories of heavy vehicles.
In a study prepared for the Motor Vehicle Manufacturers Association by Energy and Environmental Analysis (1984), costs and efficiency gains for several technologies were presented for the categories ''light-heavy," "medium-heavy," and "heavy-heavy," corresponding to Classes 2B-5, Classes 6–8A, and Class 8 heavy trucks. In most cases the fuel savings and the equipment costs were tabulated directly by analysts from Energy and Environmental Analysis. In others, the level of market penetration corresponding to a break-even mileage for a 2-year payback on the investment cost was provided. Because histograms of the fraction of vehicles traveling each mileage interval were also given for each category, it was possible to calculate the equipment cost using the formula:
Fuel savings = [VMT × FP] × [1 - (1/(1 + f))] × [1/FE0]
where VMT = vehicle miles traveled, FP = fuel price, FE0 = base fuel economy before addition of a technology, and f = the fractional improvement in fuel economy. The results of this calculation and other data tabulated in the EEA analysis provide the basis for Tables E.5, E.6, and E.7.
The supply curves resulting from these calculations are presented in Figures E.4, E.5, and E.6. The panel has found no estimates to compare with this analysis for heavy trucks to demonstrate the uncertainties that surely underlie these results.
As is evident from these figures, the tons of CO2 saved and the cost-effectiveness values vary dramatically for the three categories of trucks.
Domestic Air Carriers
The third most significant component of transportation energy use is domestic airlines, as indicated in Table E.1. The most advanced new jet aircraft are far more efficient than older aircraft still in service. Carlsmith et al. (1990) estimate that on a 1000-mile trip, aircraft produced in the 1960s are capable of between 40 and 50 seat-mpg, while the new Boeing 757 and 767 now in service have a fuel efficiency of 70 seat-mpg. Improvements now being introduced arise from a combination of higher bypass ratio engines, increased compressor and turbine efficiencies, and more energy-efficient flight planning and operations. Like highway vehicles, aircraft can also benefit from weight reduction and better aerodynamics.
Although some estimate efficiencies approaching 130 to 150 seat-mpg from planned vehicles utilizing fanjets along with other new technologies, personal communication with a participant in that industry would attribute a 20 percent gain to the fanjet technology alone. This same source estimates that an approximately $20 billion investment will be necessary in research, development, training, tooling, and certification for retrofitting the existing fleet with new fanjet technology.
Component costs per vehicle are estimated at $8 million. If a return on investment of 15 percent for retrofitting half of the 4000 transport vehicles in the fleet with this technology, and a 20 percent fuel economy gain, are assumed, 0.016 Gt/yr per year of CO2 reduction is possible at a cost of $300/t CO2.
Again, the estimate is based on several unverified assumptions and extrapolations that call for great caution.
Uncertainties dominate the content of much of this appendix. The first uncertainty is due to disagreements among technologists as to the costs and benefits likely to accrue from a portfolio of devices at various stages of development or implementation. This is revealed in the wide range of values in the conservation supply curves derived above from the Shackson and Leach (1980), Difiglio et al. (1990), and Ledbetter and Ross (1989) data sets.
The second uncertainty relates to disagreements among social scientists as to the relative impact of fuel prices and mandated fuel economy levels on the supply of fuel-efficient vehicles. This is revealed when comparing the results from econometric analyses by Godek (1990), Greene (1989), and
TABLE E.5 Light-Heavy Truck Data (Energy and Environmental Analysis, Inc., 1984)
(continued on page 745)
(Table E.5 continued from page 744)
TABLE E.6 Medium-Heavy Truck Data (Energy and Environmental Analysis, Inc., 1984)
(continued on page 747)
(Table E.6 continued from page 746)
TABLE E.7 Heavy-Heavy Truck Data (Energy and Environmental Analysis, Inc., 1984)
(continued on page 749)
(Table E.7 continued from page 748)
Leone and Parkinson (1990), which differ on the relative importance of market and regulatory mechanisms.
Cross-sectional data sets for several nations are now emerging to provide additional insight into consumer preferences (see Schipper, 1991). Similarly, the wide range of values in the conservation supply curves based on the Shackson and Leach and the Ledbetter and Ross studies may well be reconciled with these emerging data. In light of these remaining uncertainties, it is interesting to examine the automobile and light truck conservation supply curves in the context of fuel prices and consumer preference for new car fuel economy.
To examine the interactions of consumer choice and technology, the panel has plotted conservation supply information in a format that allows a comparison with consumer decision making. Results from Tables E.3, E.4, and E.8 derived from Shackson and Leach (1980), Ledbetter and Ross (1989), and Difiglio et al. (1990), respectively, are plotted in Figure E.7. Values
TABLE E.8 A Supply Curve for Light-Duty Vehicles Fuel Economy Technologies (Difiglio et al., 1990)
(continued on page 753)
(Table E.8 continued from page 752)
for the average new-car fleet efficiencies, along with the average fuel prices for Japan, Sweden, the United Kingdom, the United States, and West Germany, are also plotted as squares on the same figure. The automobile fleet fuel economy values were reduced by a factor of 1.3 since increased urban congestion, higher highway speeds, and a larger fraction of total miles being driven in urban areas are projected to increase the difference between the EPA fuel economy test and actual on-road fuel economy from 15 percent in 1987 to 30 percent in 2010 (Ledbetter and Ross, 1990). Although Difiglio et al. (1990) estimated a 17 percent difference in the year 2000, their fuel economy values were also reduced by the factor 1.3 to be consistent with the other adjustments in Figure E.7. Since the Shackson and Leach (1980) supply curve was for a fleet that included light trucks, no adjustment was made to their base fuel economy of 19.7 mpg.
The perspective in Figure E.7 is that of a consumer expecting a 10-year benefit stream using a 30 percent discount rate on future benefits. If the consumers in these five nations had no preference between purchasing fuel economy technology and avoiding the cost of gasoline, they would choose a set of technologies on one of the three curves. If, on the other hand, consumers valued other attributes in their vehicles sacrificed by fuel economy
technologies, they would choose a level of fuel economy lower (i.e., to the left) of the curve in Figure E.7.
If a 30 percent discount rate and a 10-year lifetime are valid assumptions, consumers in the United States and West Germany choose a level of fuel economy appropriate for their fuel prices, provided the Shackson and Leach curve is an accurate indication of the cost of technology. The fact that three other nations lie a significant distance from the steepest supply curve indicates either that their vehicle use patterns are dramatically different or that the technology cost-effectiveness information is inappropriate.
To summarize the relative magnitude of the values along with their uncertainties, a sample of average values from two different regions of two different analyses is presented below.
The discontinuity in the slope of the Difiglio curve at approximately 31 mpg (on-road fuel economy) is the point at which sales shifts in the vehicle mix are required to gain higher average fuel economy levels. Difiglio has labeled this point the ''maximum technology" point for the year 2000.2 The panel has chosen the region beyond this point as the region in which life-style adjustments are incurred. Up to this point, the costs of attributes lost or compromised by fuel efficiency technologies are ignored even though consumers consider them substantial.
The average cost-effectiveness values, as distinct from marginal cost-effectiveness values, which increase significantly as one moves past 25 mpg, are summarized in Table E.9 as a function of the discount rate for the three cost curves (see calculations in Tables E.3 through E.8). The costs are constrained, however, by the fact that the panel has not considered cumulative costs exceeding $3850 (1990 dollars) from the Shackson and Leach study, and Ledbetter and Ross did not go beyond cumulative costs of $609 (1990 dollars).
To provide a visual impression of the relative magnitudes and their uncertainties, Figure E.8 illustrates values derived for the discount rate of 6 percent for light duty vehicles. One should keep in mind that data for the automobile and light truck calculations were from three sources, thereby providing the indication of uncertainty.
1. Throughout this report, tons (t) are metric; 1 Mt = 1 megaton = 1 million tons; 1 Gt = 1 gigaton = 1 billion tons.
2. Subsequent to the preparation of these results, K. G. Duleep (co-author with Difiglio and Green) has refined the estimate of the "maximum technology" point. Duleep estimates that in 1996 all available technologies could produce a CAFE mpg of 29.3 (22.5 on-road) and in 2001 all available technologies could produce a CAFE mpg of 36.0 (27.7 on-road) (Plotkin, 1991).
TABLE E.9 Implementation Cost of Vehicle Efficiency Improvements
Amann, C. A. 1989. The automotive spark-ignition engineAn historical perspective. In History of the Internal Combustion Engine, ICE, Volume 8, Book No. 100294-1989, E. F. C. Somerscales and A. A. Zagotta, eds. American Society of Mechanical Engineers.
Atkinson, S. E., and R. Halvorsen. 1984. A new hedonic technique for estimating attribute demand: An application to the demand for automobile fuel efficiency. Review of Economics and Statistics 66(3):416–426.
Atkinson, S. E., and R. Halvorsen. 1990. Valuation of risks to life: Evidence from the markets for automobiles. Review of Economics and Statistics 72(1):137–142.
Berger, J. O., M. H. Smith, and R. W. Andrews. 1990. A system for Estimating Fuel Economy Potential due to Technology Improvements. Ann Arbor, Mich: The University of Michigan, School of Business Administration.
Bleviss, D. 1988. The New Oil Crisis and Fuel Economy Technologies. New York: Quorum Books.
Bussmann, W. V. 1990. Potential Gains in Fuel Economy: A Statistical Analysis of Technologies Embodied in Model Year 1988 and 1989 Cars. Intra-Industry Analysis of Fuel Economy Efficiencies.
Carlsmith, R. S., W. U. Chandler, J. E. McMahon, and D. J. Santini. 1990. Energy Efficiency: How Far Can We Go? Report ORNL/TM-11441. Prepared for the Office of Policy, Planning and Analysis, U.S. Department of Energy. Oak Ridge, Tenn.: Oak Ridge National Laboratory.
Davis, S. C., D. B. Shonka, G. J. Anderson-Batiste, and P. S. Hu. 1989. Transportation Energy Data Book: Edition 10. Report ORNL-6565 (Edition 10 of ORNL-5198). Prepared for the U.S. Department of Energy. Oak Ridge, Tenn.: Oak Ridge National Laboratory.
Difiglio, C., K. G. Duleep, and D. L. Greene. 1990. Cost effectiveness of future fuel economy improvements. The Energy Journal 11(1):65–86.
Energy and Environmental Analysis (EEA). 1984. Documentation of market penetration forecasts. In Historical and Projected Emissions Conversion Factor and Fuel Economy for Heavy Duty Trucks, 1962–2002. Arlington, Va.: Energy and Environmental Analysis, Inc.
Godek, P. E. 1990. The Corporate Average Fuel Economy Standard 1978–1990. Working Paper, October 1990.
Greene, D. L. 1989. CAFE or Price?: An Analysis of the Effects of Federal Fuel Economy Regulations and Gasoline Price on New Car MPG, 1978–89. Prepared for the Office of Policy Integration, Office of Policy, Planning and Analysis, U.S. Department of Energy. November 1989. Washington, D.C.: U.S. Department of Energy.
Ledbetter, M., and M. Ross. 1989. Supply curves of conserved energy for automobiles. Draft paper prepared for Lawrence Berkeley Laboratory by the American Council for an Energy-Efficient Economy, Washington, D.C.
Leone, R. A., and Parkinson, T. W. 1990. Conserving energy: Is there a better way? Paper prepared for the Association of International Automobile Manufacturers.
Plotkin, S. E. 1991. Testimony before Senate Committee on Energy and Natural Resources. Washington, D.C.: Office of Technology Assessment. March 20, 1991.
Schipper, L. 1991. Energy saving in the U.S. and other wealthy countries: Can the momentum be maintained? Draft. International Energy Studies, Energy Analysis Program, Applied Science Division, Lawrence Berkeley Laboratory.
Shackson, R. H., and H. J. Leach. 1980. Using Fuel Economy and Synthetic Fuels to Compete with OPEC Oil. Pittsburgh, Pa.: Carnegie-Mellon University Press.
Unnasch, S., C. B. Moyer, D. D. Lowell, and M. D. Jackson. 1989. Comparing the Impact of Different Transportation Fuels on the Greenhouse Effect. Prepared by the Acurex Corporation for the California Energy Commission. April 1989. Sacramento: California Energy Commission.
Wright, J., A. Meier, M. Maulhardt, and A. H. Rosenfeld. 1981. Supplying Energy Through Greater Efficiency: The Potential for Conservation in California's Residential Sector. Report LBL-10738, EEB 80-2. January 1981. Berkeley, Calif.: Lawrence Berkeley Laboratory.