This is the second of three files comprising the document:
Impacts of the Kyoto Protocol
on U.S. Energy Markets
and Economic Activity
October 1998
Energy Information Administration
Office of Integrated Analysis and Forecasting
U.S. Department of Energy
Washington, DC 20585
197019801996Reference1990+24%
1990+9%
1990-3%
0246810121416QuadrillionBtuPetroleumNaturalGasSteamCoalElectricityLossesHistoryProjectionsfor2010Figure 47. Industrial Sector Energy Consumption
by Fuel, 1970, 1980, 1996, and 2010
Sources: History: Energy Information Administration, State Energy Data
Report 1995, DOE/EIA-0214(96) (Washington, DC, December 1997).
Projections: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.
D080398B, and FD03BLW.D080398B.
1995200020052010201520200.70.80.91.0Index,1990=1.0Reference1990+24%
1990+14%
1990+9%
19901990-3%
1990-7%
1970197519801985199019950.50.81.01.31.5Index,1990=1.0IndustrialDeliveredEnergyIntensity,
HistoryFigure 48. Projected Energy Intensity in the
Industrial Sector, 1995-2020
Sources: History: Consumption: Energy Information Administration, State
Energy Data Report 1995, DOE/EIA-0214(96) (Washington, DC, December
1997). Output: Constructed by Standard & PoorÕs DRI from U.S. Department of
Commerce, ÒBenchmark Input-Output Accounts for the U.S. Economy, 1992:
Make, Use, and Supplementary Tables,Ó Survey of Current Business, November
1997, and predecessor benchmark tables. Projections: Office of Integrated
Analysis and Forecasting, National Energy Modeling System runs KYBASE.
D080398A, FD24ABV.D080398B, FD1998.D080398B, FD09ABV.D080398B,
FD1990.D080398B, FD03BLW.D080398B, and FD07BLW.D080398B.
reduced, resulting in smaller incentives for the addition
of new, less energy-intensive capital equipment. The
changes in energy intensity for the industrial subsectors
(Figure 49) indicate that slower growth in output can
lead to less pronounced declines in energy intensity in
the more stringent carbon reduction cases.
The change in aggregate industrial energy intensity can
be decomposed into two effects. One is the change in
energy intensity that results from a change in the
composition of industrial output. For example, if the
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
FoodPaperChemicalsGlassCementSteelAluminumAgricultureMiningConstructionMetal-BasedDurablesOtherMfg.
Total01020-10-20-30TotalPercentChangeReference1990+24%1990+9%1990-3%
Figure 49. Projected Change in Industrial Sector
Energy Intensity, 1996-2010
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.
D080398B, and FD03BLW.D080398B.
output of the most energy-intensive industries grows
more slowly than other parts of the industrial sector,
aggregate energy intensity will fall even though no
individual industryÕs energy intensity has changed. This
is the ÒstructuralÓ effect. The other is increased energy
efficiency and shifts toward less energy-intensive
products in individual industries (the Òefficiency/
otherÓ effect). The relative contributions of these two
effects to the reduction in aggregate industrial intensity
have varied substantially over time (Figure 50).47 For
example, between 1980 and 1985, when aggregate
industrial intensity fell by 3.6 percent annually, the
structural and efficiency/other effects made equal
contributions to the decline. Over a longer period, from
1980 to 1996, the structural effects dominated the
reduction in aggregate industrial energy intensity.
Similarly, in the projections, the structural and
efficiency/other effects can be decomposed. About two-
thirds of the projected reduction in aggregate industrial
intensity is attributable to the structural effect, which is
slightly larger in the carbon reduction cases than in the
reference case.
Total expenditures for energy purchases in the industrial sector are projected to be $121 billion in 2010 in the
reference case. In the carbon reduction cases, the effects
of higher energy prices are reduced by fuel switching
and reduced consumption. Nevertheless, energy expenditures in 2010 are projected to be $24 billion (20 percent)
higher in the 1990+24% case and $60 billion (50 percent)
higher in the 1990+9% case than in the reference case,
and in the 1990-3% case they are projected to be even
higherÑ$101 billion (83 percent) higher than in the
reference case at $222 billion (Figure 51).
1980-
19851980-
1996Reference1990+24%
1990+9%
1990-3%
0.00.5-0.5-1.0-1.5-2.0AnnualPercentChangeEfficiency/OtherStructural1996-2010Figure 50. Structural and Efficiency/Other Effects
on Industrial Energy Intensity,
1980-1985, 1980-1996, and 1996-2010
Sources: History: Consumption: U.S. Department of Commerce, National
Technical Information Service, National Energy Accounts, PB89-187918
(Springfield, VA, February 1989). Output: Constructed by Standard & PoorÕs DRI
from U.S. Department of Commerce, ÒBenchmark Input-Output Accounts for the
U.S. Economy, 1992: Make, Use, and Supplementary Tables,Ó Survey of Current
Business, November 1997, and predecessor benchmark tables. Projections:
Office of Integrated Analysis and Forecasting, National Energy Modeling System
runs KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.D080398B, and
FD03BLW.D080398B.
1990+24%1990+9%1990+9%
LowTechnology1990+9%
HighTechnology1990-3%
020406080100PercentFigure 51. Change From Projected Reference Case
Energy Expenditures in the Industrial
Sector for Alternative Carbon Reduction
Cases, 2010
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.
D080398B, FREEZE09.D080798A, HITECH09.D080698A, and FD03BLW.
D080398B.
47The decomposition is done with the divisia index. For an explanation of the calculation of the index, see Boyd et al., ÒSeparating the
Changing Composition of U.S. Manufacturing Production from Energy Efficiency Improvements: A Divisia Index Approach,Ó The Energy
Journal, Vol. 8, No. 2 (1987). Alternative decomposition methods are discussed in Greening et al., ÒComparison of Six Decomposition Methods: Application of Aggregate Energy Intensity for Manufacturing in Ten OECD Countries,Ó Energy Economics, Vol. 19 (1997). Note that using different time periods or subsector aggregations may also yield different results.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Sensitivity Cases assumes that no additional technology changes (as
reflected in energy intensity) will occur after 1998. Nor-
The projections of industrial sector energy expenditures mal turnover of capital, however, would result in some
in the carbon reduction cases are based on the reference decline in energy intensity as old equipment is replaced
case assumptions about technology improvements and with currently available equipment with lower energy
likely industrial response. Expenditures would be much intensity. The high technology case assumes an aggreshigher if technology improvements occurred at a slower sive private and Federal commitment to energy-related
rate than in the reference case. On the other hand, a more research and development, which results in successful
optimistic technology outlook would reduce energy commercialization of energy-saving technologies.48
expenditures.
As noted earlier, the analysis uses technology bundles
To span the technology alternatives, low and high tech-to characterize technological change in the energynology sensitivity cases, based on the 1990+9% carbon intensive industries. This approach is illustrated
reduction case, were analyzed. The low technology case in Table 9. For example, the energy intensity of the
Table 9. Projected Energy Intensities for Industrial Process Steps and End Uses
Industry/Process Step or End Use 1990+9% Low Technology 1990+9% 1990+9% High Technology
Food ................................ 1.00 0.89 0.79
Direct Fuel ........................... 1.00 0.88 0.79
Hot Water/Steam ...................... 1.00 0.89 0.79
Refrigeration ......................... 1.00 0.90 0.79
Other Electric ......................... 1.00 0.90 0.79
Pulp and Paper........................ 1.00 0.78 0.64
Paper Making......................... 1.00 0.77 0.62
Bleaching ............................ 1.00 0.86 0.78
Waste Fiber Pulping.................... 1.00 0.94 0.87
Mechanical Pulping .................... 1.00 0.92 0.96
Semi-Chemical........................ 1.00 0.86 0.91
Kraft, Sulfite, misc. ..................... 1.00 0.78 0.61
Wood Preparation ..................... 1.00 0.95 0.92
Bulk Chemicals ....................... 1.00 0.95 0.85
Electrolytic ........................... 1.00 0.91 0.83
Other Electric ......................... 1.00 0.90 0.83
Direct Fuel ........................... 1.00 0.88 0.83
Steam/Hot Water ...................... 1.00 0.89 0.83
Feedstocks........................... 1.00 0.99 0.87
Glass ................................ 1.00 0.73 0.59
Post-Forming ......................... 1.00 0.91 0.94
Forming ............................. 1.00 0.89 0.88
Melting/Refining ....................... 1.00 0.63 0.41
Batch Preparation ..................... 1.00 0.96 0.99
Cement .............................. 1.00 0.85 0.77
Finish Grinding........................ 1.00 0.82 0.72
Dry Process .......................... 1.00 0.83 0.66
Wet Process.......................... 1.00 0.93 0.97
Steel ................................ 1.00 0.81 0.50
Cold Rolling .......................... 1.00 0.56 0.33
Hot Rolling ........................... 1.00 0.65 0.37
Ingot Casting/Primary Rolling ............ 1.00 1.00 1.00
Continuous Casting .................... 1.00 1.08 1.06
Blast Furnace/Basic Oxygen Furnace ...... 1.00 1.10 0.50
Electric Arc Furnace.................... 1.00 1.00 0.62
Coke Oven ........................... 1.00 1.00 0.98
Primary Aluminum ..................... 1.00 0.87 0.71
Notes: The energy intensity for the low technology case is defined as 1.0. The 1990+9% case and high technology case energy intensities are
indexed against the energy intensity for the low technology case. The intensities are not additive within an industry.
Source: The high technology sensitivity case is based in part on an analysis prepared by Arthur D. Little, Inc., Aggressive Technology Strategy for the NEMS Model
(1998).
48The high technology sensitivity case is based in part on an analysis prepared by Arthur D. Little, Inc., Aggressive Technology Strategy for
the NEMS Model (1998).
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Cogeneration Systems
In every carbon reduction case considered in this report, significant hurdles must be overcome if the potential is to
neither photovoltaics nor fuel cells are projected to gain be realized. Siting one or more power and steam genera-
significant market penetration, because of their high tors in an area already dense with buildings could prove
costs. With payback periods of more than 20 years, the to be a challenge, as could the installation, maintenance,
success of these technologies seems largely dependent on and repair of lines to carry steam and hot or chilled water
reducing production costs and increasing efficiency supplies in cities with under-street congestion of existing
(which would result in further cost reductions for the con-gas, water, sewage, and electricity lines. Also, construcsumer). Federal financial assistance would also play a role tion costs for district energy systems are about one-third
in their success. higher than those for conventional generating technolo-
A key issue facing power producers and their customers gies.
is whether the types of cogeneration systems currently Although it is possible that fuel cost savings over the life
used in the United States will be extended to include dis-of a district energy plant could offset its higher initial contrict energy systems and advanced turbine systems struction cost, electricity producers might be reluctant to
(ATS). Cogeneration systems, also called combined heat invest significant capital during a period of regulatory
and power systems, simultaneously produce heat in the reform. Even after the current restructuring process in
form of hot air or steam and power in the form of electric-U.S. electricity markets is completed, the risk of nonreity by a single thermodynamic process, usually steam covery of capital for capital-intensive technologies in a
boilers or gas turbines, reducing the energy losses that competitive environment will make finding investors in
occur when process steam and electricity are produced such projects a challenge. Moreover, the development of a
independently. Thus, cogeneration systems could play a district energy system involves the coordinated effort of
significant role in reducing U.S. greenhouse gas emis-local and State governments, investors, and the commusions. nity as a whole, together with the subsequent legal, finan
cial, and environmental issues that arise with the
In 1996, electric utilities used more than 21 quadrillion inclusion of many and diverse stakeholders.
Btu of energy from the combustion of coal, natural gas,
and oil to produce the equivalent of only 7 quadrillion Btu Another technology that some energy analysts believe
of electricity available at the plant gate, representing a could significantly reduce greenhouse gas emissions is
conversion loss of 67 percent.a Consequently, unused the next-generation, very-high-efficiency ATS. These tur
waste heat at utility plants accounted for 346 million met-bines are expected to operate, at minimum, 5 to 10 percent
ric tons or nearly 24 percent of U.S. carbon emissions in more efficiently than steam boilers and to cost less than
1996. Additional losses on the order of 7 percent are $350 per kilowatthour when used as a simple-cycle tur
incurred during transmission and distribution of electric-bine.b Their small size (5 megawatts) and short construc
ity to customers.b Because cogeneration systems capture tion and delivery schedule (18 months) result in relatively
and use a significant portion of the waste heat energy, smaller capital outlays and faster capital recovery, which
they are nearly twice as efficient as conventional power are expected to give them an economic advantage over
plants in extracting usable energy. About 6 percent of large central-station turbines.
total U.S. generating capacity includes some type of Commercialization of ATS turbines is not expected until
cogeneration system, in such diverse industries as manu-2001, and penetration is expected to occur first where
facturing, mining, and refining.c
there is a need to satisfy internal power and steam
Some energy analysts believe that there is even greater requirements at industrial and large commercial estab
potential to increase the penetration of cogeneration sys-lishments. But large-scale penetration of the ATS technol
tems and reduce carbon emissions by wide-scale con-ogy as envisioned by its advocates depends on the
struction of district energy systems.b District energy development of a significant niche market for this
systems distribute chilled water, steam, or hot water to cogeneration systemÑa market characterized as having a
buildings to provide air conditioning, space heating, small, but not constant, demand for steam. ATS in
domestic hot water, and industrial process energy. About electric-only mode may not be competitive with other pri
5,800 district energy systems are installed in the United mary power technologies, and a constant demand for
States, serving more than 8 percent of commercial floor-steam could be satisfied more economically by conven
tional gas and combined-cycle steam boilers.b Conse
spaceÑprimarily military bases, universities, hospitals,
downtown areas, and other group buildings.d quently, the competitiveness of ATS with other
generating technologies depends on locating markets
The greatest growth potential for district energy systems with an optimal demand for steam during part of the day
is in the area of utility-financed cooling systems for down-and maximum demand for electricity for the remainder of
town areas where there is a large amount of commercial the day, even during off-peak periods. Few, if any, power
floorspace located in a relatively small area; however, markets would meet such stringent criteria.
aEnergy Information Administration, Annual Energy Review 1996, DOE/EIA-0384(96) (Washington, DC, July 1997).
bInterlaboratory Working Group on Energy-Efficient and Low-Carbon Technologies, ÒScenarios of U.S. Carbon Reductions,Ó
LBNL-40533, ORNL/CON-444 (September 1997).
cEnergy Information Administration, Annual Energy Outlook 1998, DOE/EIA-0383(98) (Washington, DC, December 1997).
dSee web site www.energy.rochester.edu/us/climate/abstract.htm, ÒDistrict Energy in U.S. Climate Change Strategy.Ó
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
paper-making process step in the pulp and paper
industry is 19 percent lower in the 1990+9% case than in
the low technology sensitivity case. For the same process
step, energy intensity is 36 percent lower in the high
technology case than in the low technology case. For
some process steps where the change in intensity is very
small, the higher energy prices in the 1990+9% case lead
to a slightly lower intensity than in the high technology
case, where energy prices are lower. (The technology
cases were modeled across all sectors simultaneously.
The resulting lower consumption in the high technology
case also resulted in lower prices.)
In the 1990+9% low technology case, industrial energy
expenditures in 2010 are projected to be nearly double
those in the 1990+9% carbon reduction case and $110
billion higher than those in the reference case. In the
high technology sensitivity case, energy expenditures
are projected to be only $23 billion higher than in the
reference case, which has no carbon reductions, in 2010.
The high technology case reduces, but does not
eliminate, the impact of higher energy prices, producing
$37 billion in savings attributable to the assumed
technology advances (Figure 51).
Another sensitivity case for the 1990+9% carbon reduction case was implemented to examine the impacts of
alternative assumptions about the use of cogeneration
and biomass for electricity generation. These assumptions reflect the possibility that natural gas cogeneration
and biomass could be used more extensively than projected in the other cases. Natural-gas-fired cogeneration
is posited to be a function of two economic factors. One
is demand for process steam, with higher demand leading to more cogeneration. (In the carbon reduction cases,
industrial steam demand is reduced because the
requirements for process steam fall when industrial output falls.) The other is the spread between electricity and
natural gas prices, with a higher price difference leading
to more gas-fired cogeneration. The assumption used
here is that natural-gas-fired cogeneration is more
responsive to increasing prices.
Industrial biomass consumption is dominated by activities in the pulp and paper industry, where biomass residue and pulping liquor are used to supply more than
half the industryÕs energy requirements. Consumption
of biomass residue and pulping liquor is a function of
the industryÕs output. Consequently, biomass consumption tends to fall in the carbon reduction cases, because
industrial output is projected to be lower. The 1990+9%
aggressive cogeneration/biomass sensitivity case
assumes that the reduction in biomass consumption will
be attenuated by additional biomass recovery and utilization. Additional biomass recovery also leads to an
increase in cogeneration from biomass, which further
reduces the requirements for other fossil fuels.
The aggressive cogeneration/biomass case results in a 9percent increase (20 billion kilowatthours) in the level of
gas-fired cogeneration in 2010 relative to the reference
case (Figure 52). This is smaller than the change seen in
the high technology sensitivity case, because industrial
output is lower in the aggressive cogeneration/biomass
sensitivity than in the high technology case. (Industrial
output is lower in the aggressive cogeneration case than
in the high technology case, because the projected
energy prices are higher in the aggressive cogeneration
case.) Biomass consumption in 2010 is projected to be 1.2
percent (27 trillion Btu) higher in the aggressive
cogeneration/biomass sensitivity case than in the reference case (Figure 52). As with cogeneration, this increase
is slightly less than the change seen in the high technology sensitivity case, again because of the lower industrial output projected in the aggressive cogeneration/
biomass case. Projected energy expenditures in the
industrial sector in 2010 in this sensitivity case are $15
billion less than in the 1990+9% case. It should be noted
that neither the cost nor the likelihood of achieving the
assumed changes in the high technology or aggressive
cogeneration/biomass sensitivity case has been evaluated. Instead, the experiments were an attempt to span
the range of possible outcomes.
1990+9%
1990+9%
LowTechnology1990+9%
HighTechnology1990+9%
AggressiveCogeneration1990-3%
0.02.04.06.08.010.012.0-2.0-4.0PercentChangefromReferenceCaseBiomassCogenerationFigure 52. Natural-Gas-Fired Cogeneration and
Biomass Consumption in the Industrial
Sector in Alternative Carbon Reduction
Cases, 2010
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FD09ABV.D080398B, FREEZE09.
D080798A, HITECH09.D080698A, BEHAVE09.D080498A, and FD03BLW.
D080398B.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Industrial Composition
Because non-Annex I countries are not required to reduce industries or markets. The industries affected and the per-
emissions under the Kyoto Protocol, their energy prices centage reductions in projected industrial output in the
are likely to be lower than those in the Annex I countries, reference case were as follows: bulk chemicals, 28.5; aluincluding the United States. As a result, more energy-minum, 13.7; pulp and paper, 10.2; steel, 30.5; and cement,
intensive industries could migrate from areas with high 38.2.
energy costs, and those that remain could lose markets to A second study was conducted at EIAÕs request by Char-
lower-cost foreign competition. Energy-intensive indus-les River Associates (CRA),b using a more general
tries also may face reduced demand as consumers shift approach. Explicit linkages to international trade were a
their consumption patterns to less energy-intensive fundamental part of the modeling framework for the
goods and services. There are several counter arguments study, which was conducted under assumptions similar
to this hypothesis: the relatively small share of energy to those of the 1990+14% carbon reduction case in this
expenditures in annual manufacturing expenditures analysis. The industries affected and the percentage
makes the impact of differential energy prices relatively reductions from reference case output projections were as
unimportant; energy prices are not important determi-follows: total chemicals, 3.9; nonferrous metals, 1.5; pulp
nants of international trade or capital flows, which and paper plus printing, 0.7; steel, 1.4; and nonmetallic
implies that U.S. energy-intensive industries are not minerals, 1.4. The percentage output reductions from the
likely to be seriously affected by an energy price disad-comparable NEMS case (1990+14%) are about double the
vantage; and a large number of business opportunities CRA values: nonferrous metals, 4.4; pulp and paper plus
related to climate change mitigation will become avail-printing, 2.0; steel, 3.1; and nonmetallic minerals, 3.5. The
able both domestically and in non-Annex I countries. exception is total chemicals for which the NEMS results
Needless to say, there are widely divergent points of view project a slightly smaller reduction of 3.5 percent. The
about the likelihood of significant industrial migration projections from NEMS, which estimates only domestic
and the extent of adverse impacts on U.S. industry.a An output reductions, and from CRA, which treated both
analysis of the change in industrial composition, which international capital flows and domestic output reducwould require an analysis of all the relative costs of manu-tions, are significantly lower than those from the Argonne
facturing inputs, of which energy costs are only one, National Laboratory study.
monetary issues, and international trade issues, is beyond
the scope of this report. In view of the above results, it is difficult to distinguish
the effects of reduced output from those that could result
One published study has attempted to evaluate the poten-from industrial migration abroad in response to differtial effects of differential changes in international energy ences in international energy prices. There are many ana-
prices on the U.S. industrial sector. The study was con-lytical complexities in the assessment of potential effects
ducted by Argonne National Laboratory in a workshop of carbon reductions on industrial output. A complete
format (see Argonne National Laboratory, The Impact of analysis of the issue would require consideration of all
High Energy Price Cases on Energy-Intensive Sectors: Per-input costs, including infrastructure and locational
spectives from Industry Workshops (July 1997)). Industry-advantages, monetary issues, and trade issues. Signifispecific discussion papers circulated to workshop partici-cant additional research would be required to examine
pants contained analyses that examined impacts for each the differential impacts of climate change policies on the
individual industry, assuming no price changes for other United States and other countries.
aThe following authors provide a sample of the breadth of disagreement in this area: American Petroleum Institute, Impacts of
Market-Based Greenhouse Gas Emission Reduction Policies on U.S. Manufacturing Competitiveness, January 1998; American Automobile
Manufacturers Association, Economic Implications of the Adoption of Limits on Carbon Emissions from Industrialized Countries, November
1997; Argonne National Laboratory, The Impact of High Energy Price Cases on Energy-Intensive Sectors: Perspectives from Industry Workshops, July 1997; Matthewson, et al., The Economic Implications for Canada and the United States of International Climate Change Policies,
1997 Canadian Energy Research Institute Environment-Energy Modeling Forum, October 1997; Repetto, et al., U.S. Competitiveness is
Not at Risk in the Climate Negotiations, (World Resources Institute, October 1997); and WEFA, Inc., Global Warming: The High Cost of the
Kyoto Protocol, National and State Impacts, 1998.
bCharles River Associates, Report to the Energy Information Administration (August 1998).
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Transportation Demand
Background
In terms of primary energy use in 1996, transportation
sector carbon emissions, which almost equaled industrial carbon emission levels, were the second highest
among the end-use demand sectors. Nearly 33 percent
of all carbon emissions and 78 percent of carbon emissions from petroleum consumption originate from the
transportation sector. In the reference case, carbon emissions from transportation are projected to grow at an
average annual rate of 1.9 percent to 2010, compared
with 1.4 percent for the commercial sector and 1.2 percent for both the residential and industrial sectors. In
addition, transportation is the only sector with increasing carbon emissions projected for the period from 2010
to 2020 in the carbon reduction cases. Therefore, if there
are no specific initiatives to reduce carbon emissions in
the transportation sector, especially beyond 2010,
increasing pressure may have to be exerted in the other
sectors in order to reach and then maintain 2010 carbon
emissions targets beyond 2010.
Consumers select light-duty vehicles (cars, vans, pickup trucks, and sport utility vehicles) based on a number
of attributes: size, horsepower, price, and cost of driving;
weighting these attributes by their personal preferences.
This analysis uses past experience to determine the
weights that each of these attributes have in terms of
consumer preferences for conventional vehicles. Technologies are represented by component (e.g., front
wheel drive, electronic transmission type) with each
technology component defined by a date of introduction, a cost, and a weight that indicates its impact on efficiency and horsepower. The vehicles are categorized by
the 12 size classes for cars and light trucks defined by the
Environmental Protection Agency and includes 2 conventional engine technologies, and 14 alternative fuel
vehicle engine technologies. Technologies penetrate
based on both their cost-effectiveness and by consumer
preference based on past experience with similar technologies in the automotive industry. Consumers are
assumed to consider only current energy prices when
evaluating technologies. However, it is assumed that the
automobile industry requires 3 years for minor technology makeovers and 5 years for major redesigns, estimating future fuel prices based on their rate of growth in the
past 3 to 5 years. Therefore, manufacturers consider
whether future fuel prices will enable their technologies
to be cost-effective from a consumer standpoint.
Penetration of alternative-fuel vehicles is based on four
consumer criteriaÑvehicle price, cost of driving per
mile, vehicle range, and availability of refueling stations.
Each of these attributes is weighted according to consumer surveys and expected changes over the forecast
period as a result of technological improvements, larger
scales of production, the availability and cost of fuel-
saving technologies, and the availability of alternative-
fuel refueling stations as more alternative-fuel vehicles
penetrate the market. Production levels for alternative-
fuel vehicles are constrained by the lead time to switch
production to a particular technology and the availability of technologies in each size class.
Depending on per capita income, fuel prices, and fuel
economy, consumers may switch to either smaller size
classes or smaller vehicles with lower horsepower
requirements within a size class. The trend in vehicle
sales toward or away from light trucks (vans, sport utility vehicles, and pickups) is determined by fuel prices.
Vehicle travel is determined by the cost of driving per
mile and per capita income. For flex-fuel or bi-fuel
alternative-fuel vehicles, the percentage use of each fuel
is based on the price differential between gasoline and
the alternative fuel.
Responding to changes in fuel prices, gasoline has a 2year demand elasticity of -0.25 and a 20-year elasticity of
-0.45. In the long term, consumers are expected to alter
their purchasing patterns and manufacturers to incorporate more fuel-saving technologies. Because fuel use for
freight trucks and trains depends primarily on requirements for freight movement as a result of economic
activity and the slow turnover of the stock, distillate fuel
has lower 2-year and 20-year price elasticities, at -0.09
and -0.13, respectively. In addition to fuel prices, business and personal air travel also depend on gross
domestic product (GDP) and per capita income, respectively, and have very slow rates of stock turnover. Jet
fuel has 2-year and 20-year elasticities of -0.12 and -0.15.
Energy intensity in the transportation sector is defined
as energy use (in terms of gallons of gasoline) per vehicle
per year. In the reference case, transportation energy
intensity in 2010 is projected to be about 635 gallons of
gasoline per vehicle, or about 53 gallons per month
(Figure 53). Energy intensity in the 1990+24% case is
lower than in the reference case but only by 12 gallons of
gasoline per car per month. In the 1990+9% case, the
projected energy intensity in 2010 is almost 53 gallons
lowerÑequivalent to 1 monthÕs use of gasoline. In
the 1990-3% and 1990-7% cases, the corresponding
reductions in gasoline consumption in 2010 are
equivalent to nearly 1.5 and 2 months of gasoline use,
respectively.
In the absence of fuel price changes, transportation
energy intensity will change in response to stock
turnover, technology availability, and income effects
(Table 10). Because 1998 prices are lower than those
projected for 2010 in the reference case, vehicle-miles
traveled would be higher and fuel efficiency lower than
in the reference case if the 1998 price level continued.
Constant 1998 fuel prices would slightly increase air
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
1996Reference1990+24%
1990+14%
1990+9%
19901990-3%
1990-7%
0100200300400500600700GallonsofGasolineperVehicleperYearProjectionsfor2010Figure 53. Light-Duty Vehicle Energy Intensity,
1996 and 2010
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD1998.
D080398B, FD09ABV.D080398B, FD1990.D080398B, FD03BLW.D080398B,
and FD07BLW.D080398B.
Table 10. Projected Average Transportation
Energy Intensities by Mode of Travel,
2010
(Million Btu per Vehicle per Year)
Travel Mode
1998
Average Reference
Constant
1998 Prices
Light-Duty Vehicles . . 78.4 79.5 80.9
Freight Truck ....... 699.5 787.7 787.7
Aircraft ............ 486,100 517,100 521,900
Source: Office of Integrated Analysis and Forecasting, National Energy Mod-
eling System run KYBASE.D080398A.
travel, but aircraft efficiency levels would not decline
relative to those in the reference case. More air travel
would necessitate higher aircraft stock levels in 2010, but
the increase would be more than offset by higher levels
of travel per plane. Freight truck fuel intensity would
not change with constant prices, because freight travel is
determined primarily by economic activity rather than
fuel prices. The slightly lower fuel prices in the constant
price case would not be enough to lower the fuel
economy of freight trucks relative to their projected fuel
economy in the reference case.
Carbon Reduction Cases
The transportation sector is the only sector that does not
reach 1990 carbon emissions levels by 2010 in any of the
carbon reduction cases (Figure 54). In the reference case,
energy demand in the transportation sector is projected
to exceed 1990 levels by approximately 10.7 quadrillion
Btu in 2010, a 49-percent increase (Figure 55). The corresponding increases are 9.4 quadrillion Btu in the
1990+24% case, 8.6 quadrillion Btu in the 1990+9% case,
and 6.6 quadrillion Btu in the 1990-3% case.
19901996Reference1990+24%
1990+14%
1990+9%
19901990-3%
1990-7%
0100200300400500600700MillionMetricTonsHistoryProjectionsfor2010Figure 54. Carbon Emissions in the Transportation
Sector, 1990, 1996, and 2010
Sources: History: Energy Information Administration, Emissions of
Greenhouse Gases in the United States 1996, DOE/EIA-0573(96) (Washington,
DC, October 1997). Projections: Office of Integrated Analysis and Forecasting,
National Energy Modeling System runs KYBASE.D080398A, FD24ABV.
D080398B, FD1998.D080398B, FD09ABV.D080398B, FD1990.D080398B,
FD03BLW.D080398B, and FD07BLW.D080398B.
197019751980198519901995200020052010201520200510152025303540QuadrillionBtuReference1990+24%
1990+14%
1990+9%
19901990-3%
1990-7%
HistoryProjectionsFigure 55. Fuel Consumption in the Transportation
Sector, 1970-2020
Sources: History: Energy Information Administration, State Energy Data
Report 1995, DOE/EIA-0214(96) (Washington, DC, December 1997).
Projections: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD1998.
D080398B, FD09ABV.D080398B, FD1990.D080398B, FD03BLW.D080398B,
and FD07BLW.D080398B.
Relative to the reference case, only 14 percent of the
projected reduction in total energy demand for all sectors in 2010 occurs in the transportation sector in the
1990+24% case, 19 percent in the 1990+9% case, and 24
percent in the 1990-3% case. In the 1990-3% case, the
reduction in carbon emissions from all sectors in 2010 is
approximately 492 million metric tons, of which 18
percent comes from the transportation sector.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Light-Duty Vehicles
Travel Demand. Light-duty vehicle travel (cars, pickup
trucks, vans, and sport utility vehicles) in 2010 is
projected to be 1.3 percent lower than in the reference
case in the 1990+24% case, 5.2 percent lower in the
1990+9% case, and 11.2 percent lower in the 1990-3%
case (Figure 56). Declines in light-duty vehicle travel
have been seen historically in 1973-1974 (2.7 percent)
and 1979-1980 (1.6 percent). In the 1990+24% and
1990+9% cases, the levels of light-duty vehicle travel rise
between 2005 and 2008, they are projected to decline by
an average of 1.2 percent per year over the same period
in the 1990-3% case (comparable to the rate of decline
from 1979 to 1980). In 1973-1974 and 1979-1980,
disposable per capita income was declining, at 0.7percent and 0.3-percent annual rates, respectively.
Those historical declines in income per capita, combined
with rising fuel prices, further reduced vehicle travel. In
contrast, from 2005 to 2008 income per capita is
projected to rise at an average annual rate of 0.8 percent,
more than twice the projected rate in the reference case,
partially offsetting the reductions in travel that are
expected to accompany higher fuel prices.
Slowing growth in vehicle-miles traveled is projected
even in the reference case, for several reasons. First, as
the Òbaby boomersÓ age, they are expected to drive less
(although they probably will drive more than previous
generations of the same age group).49 Second, as more
women have entered the workforce over the past three
19701980199020002010202005001,0001,5002,0002,5003,0003,500BillionVehicleMilesTraveledReference1990+24%
1990+9%
1990-3%
HistoryProjectionsFigure 56. Light-Duty Vehicle Travel, 1970-2020
Sources: History: U.S. Department of Transportation, Federal Highway
Administration, Highway Statistics, various years, (Washington, DC).
Projections: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.
D080398B, and FD03BLW.D080398B.
decades, resulting in more two-income households,
female drivers have logged more vehicle-miles of travel;
however, that growth will eventually slow as the
vehicle-miles traveled by women approaches that of
men. Finally, consumers have been keeping their
vehicles longer than in past decades, and older cars tend
to be driven less than newer cars. A countervailing trend
is the recent growth in purchases of light trucks, which
are driven 4.7 percent more per year than cars. In the
carbon reduction cases, a reversal of this trend back to
car sales as a result of higher fuel prices is expected,
leading to slower growth in vehicle-miles traveled.
After 2010, vehicle-miles of travel, total fuel use, and
total carbon emissions for light-duty vehicles are projected to begin rising again in the 1990-3% case and to
continue on an upward path through 2020, paralleling
the trends in the reference, 1990+24%, and 1990+9%
cases for the later years of the forecast. There are three
reasons for the continued growth in vehicle-miles traveled after 2012. First, carbon prices are projected to
decline in most cases after 2010. Second, lower demand
for gasoline is projected to result in lower refining costs,
lower world oil prices, and lower gasoline prices.
Finally, increases in disposable income after 2012Ñparticularly after 2015, when the U.S. average disposable
income in the 1990+24%, 1990+9%, and 1990-3% cases is
expected to exceed that projected in the reference case as
the economy rebounds from the initial response to carbon reduction effortsÑlead to more rapid increases in
light-duty vehicle travel from 2012 through 2020.
Increased telecommuting, which is assumed to reduce
vehicle-miles traveled by 0.13 percent in 2000 according
to the Climate Change Action Plan,50 is also assumed in
all the cases for this analysis, resulting in fuel savings of
21.6 trillion Btu in 2000. The 0.13-percent reduction is
assumed to continue throughout the projections, so that
as vehicle-miles traveled increase over time, the savings
from telecommuting increase proportionately.
Fuel Efficiency. In the carbon reduction cases, the fuel
economy of newly purchased light-duty vehicles in 2010
is expected to be higher than projected in the reference
case. Higher fuel prices are expected to encourage the
development of advanced fuel-saving technologies, as
well as changes in consumer purchasing patterns. For
example, average fuel efficiency for all new light-duty
vehicles in 2010, projected to be just under 25 miles per
gallon in the reference case, surpasses 27 miles per
gallon in the 1990+9% case (Figure 57), and even higher
levels might be achieved with more rapid advances in
technology, as described in the discussion of sensitivity
49Federal Highway Administration, National Personal Travel Survey: 1990 NPTS Databook, Vol. I (Washington, DC, November 1993), p. 3
18.
50U.S. Department of Energy, Office of Policy, Planning, and Program Evaluation, The Climate Change Action Plan: Technical Supplement
(Washington, DC, March 1994).
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
PassengerCarsLightTrucksAllLight-DutyVehicles010203040MilesperGallonReference1990+9%LowTechnology1990+9%1990+9%HighTechnologyFigure 57. Projected New Car and Light Truck Fuel
Economy, 2010
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FREEZE09.D080798A, FD09ABV.
D080398B, and HITECH09.D080698A.
cases below. The projections of new vehicle fuel
efficiency in the reference, 1990+24%, 1990+9%, and
1990-3% cases in 2010 are as follows: for cars, 30.6, 32.0,
33.6, and 35.6 miles per gallon; and for light trucks, 20.4,
21.2, 22.1, and 23.3 miles per gallon.
In the past, a 4.3-percent average annual increase in new
car fuel efficiency was achieved by automobile manufacturers from 1976 to 1988. Thus, the projected increases of 0.9 percent per year from 1996 to 2010 in the
1990+24% case, 1.3 percent per year in the 1990+9% case,
and 1.7 percent per year in the 1990-3% case appear to be
possible. On the other hand, those historical improvement rates resulted from the introduction of fuel-saving
technologies that involved radical changes in structural
design and were relatively inexpensive to implement.
For example, space and size reductions resulting from
downsizing to front wheel drive designs actually
reduced costs while also permitting the spatial redesign
of engine compartments, but further downsizing and
weight reductions may be difficult to achieve, because
they could eliminate larger vehicles from the marketplace and, possibly, increase the safety concerns associated with smaller light-weight vehicles. Diminishing
returns to scale have limited the potential for future fuel
savings, because many of the least expensive options
have already been implemented.
Light trucks have not achieved fuel efficiency improvements equivalent to those for automobiles, because consumers have sought higher horsepower for personal use
(particularly in sport utility vehicles), hauling (pickup
trucks), and commercial applications (standard vans).
Historically, the highest average annual growth rate in
fuel efficiency for new light trucks was 2.9 percent per
year from 1976 to 1986. In contrast, light truck fuel
economy is projected to grow by only 0.1 percent annually in the 1990+24% case, 0.4 percent annually in the
1990+9% case and 0.8 percent annually in the 1990-3%
case between 2000 and 2010. Lower growth rates occur
for light trucks in the carbon reduction cases than
historically because of the difference described above
regarding inexpensive and one time technological
improvements.
Among the 55 fuel-saving technologies that are assumed
to be available to manufacturers of light-duty vehicles in
the reference and carbon reduction cases, the most
significant market penetration is expected for drag reduction, continuously variable transmissions, electronic
transmission controls, cylinder friction reduction
technologies, advances in low-rolling-resistance tires,
variable valve timing, and accessory control units (Table
11). Aerodynamic improvements (drag reduction) have
already been implemented on many vehicles, but
further market penetration may be possible, especially
in the larger size classes. Continuously variable
transmissions match the gear ratio in a continuous
manner over the wide spectrum of gear ratios
demanded by the engine, rather than having a discrete
number of gears. Electronic transmission controls assist
the transmission by matching more precisely the gear to
be used with a given engine load. Cylinder friction
reduction technologies, such as low-friction pistons and
rings, lower the thermal and mechanical losses of the
engine. Low-rolling-resistance tires limit energy losses
from friction between tires and road surfaces. Variable
valve timing improves the thermal efficiency of an
engine by precisely timing when the ignition sparks
within the cylinder. Electronic controls and electric
motors for accessory drives on vehicles (cooling fan,
water pump, alternator, power steering and windows)
could improve fuel economy by reducing engine loads.
Changes in consumer purchasing patterns also are
expected to contribute to the fuel economy improvements for light-duty vehicles in the carbon reduction
cases. For that to happen, however, trends in consumer
choices over the past decade would have to be reversed.
With low fuel prices and high disposable income per
capita, average fuel economy has been flat from 1990 to
1996. Consumer purchases have tended toward larger
cars and light trucks, especially sport utility vehicles,
and there has been a growing preference for light trucks
over cars. Similarly, within each size class, consumers
have tended to purchase cars and light trucks that are
larger and have more horsepower.
In 1996, compact cars accounted for 45 percent of new
automobile sales, an increase from 34 percent in 1990;
however, the subcompact share of new car sales fell
from a high of 26 percent in 1991 to 19 percent in 1996.
Small pickup trucks, which captured 25 percent of the
market for new light trucks in 1990, reached a low of
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Table 11. Projected Penetration of Selected Technologies for Domestic Compact Cars, 2010
(Percent of New Sales)
Technology Reference 1990+9% 1990+9% High Technology
Drag Reduction (I) ....................... 52 73 63
Drag Reduction (II) ....................... 14 19 17
Continuously Variable Transmission ......... 48 54 49
Electronic Transmission Controls (I).......... 21 26 23
Electronic Transmission Controls (II) ......... 22 28 24
Cylinder Friction Reduction (I) .............. 46 65 56
Cylinder Friction Reduction (II) .............. 7 9 8
Low-Rolling-Resistance Tires (I) ............ 46 67 57
Low-Rolling-Resistance Tires (II) ............ 22 30 26
Variable Valve Timing..................... 79 82 52
Accessory Control Units (I)................. 24 33 28
Accessory Control Units (II) ................ 21 27 24
Source: Office of Integrated Analysis and Forecasting, National Energy Modeling System runs KYBASE.D080398A, FD09ABV.D080398B, and HITECH09.D080698A.
19 percent in 1996. Concurrently, standard and compact
sport utility vehicles, which had only a 20-percent share
of the light truck market in 1990, had a 45-percent share
in 1996. The average fuel economy of small pickup
trucks is 26.3 miles per gallon, as compared with 21.3
miles per gallon for small utility trucks and 18.1 miles
per gallon for large sport utility vehicles, which are now
growing in share at a much faster pace than even small
utility trucks. Sales of large sport utility vehicles
increased from 3.3 percent of all new light truck sales in
1991 to a high of 10.3 percent in 1996. In addition, sales of
small vans, which currently have an average fuel economy rating of about 22.7 miles per gallon, are being displaced by sales of small and large sport utility vehicles.
With a large supply of sport utility vehicles available to
consumers and a lack of station wagons designed from
sedan autos, which have a much higher fuel efficiency
rating, the fuel economy options for new vehicle buyers
are becoming limited.
With higher fuel prices in the carbon reduction cases in
2010 than in the reference case, it is projected that size
class shares will return to near 1976 levels. The
subcompact share of new car sales in 2010 is projected to
be 15 percent in the 1990+24% case, 19 percent in the
1990+9% case, and 24 percent in the 1990-3% case,
compared with 12 percent in the reference case (Figure
58). Similar trends are projected for all size classes in the
carbon reduction cases, as consumers move their vehicle
purchases down to lower size classes and sales of
compact, mid-size, and large cars are reduced. Although
shifting vehicle lines back to production of smaller cars
would require major changes in production facilities,
the lead time associated with those changes has
narrowed from about 4 years to 2 years.
Since 1990, the growth in light trucks sales at the expense
of car sales, and the growth in sales of standard and compact sport utility vehicles and minivans at the expense of
station wagons has slowed the rate of improvement in
efficiency for new light-duty vehicles. Light truck sales
shares have grown from about 37 percent of all light-
duty vehicle sales in 1990 to 43 percent in 1997, with a net
loss on average of more than 8 miles per gallon between
new cars and light trucks. In 2010, light trucks sales are
projected to be 46.1 percent of light-duty vehicle sales in
the 1990+24% case, 44.4 percent in the 1990+9% case,
and 42.5 percent in the 1990-3% case, compared with 47
percent in the reference case. Reversing the trend back
toward cars and away from truck purchases will not be
costless, however. Vehicle manufacturers reap much
higher profits from sales of light trucks than from car
sales. In addition, consumers may have difficulty finding fuel-efficient vehicles suitable for larger families
with the disappearance of many station wagons from
the new car market.
Reference1990+24%1990+9%1990-3%
01020304050PercentSubcompactCompactMidsizeLargeFigure 58. Projected Shares of Automobile Sales
by Size Class, 2010
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.
D080398B, and FD03BLW.D080398B.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Horsepower. Growth rates in new vehicle horsepower
in the light-duty vehicle market are currently at their
highest historical levels. From 1990 to 1997, new vehicle
horsepower increased at annual rates of 3.2 percent for
cars and 4.3 percent for light trucks. Between 1996 and
2010, horsepower for both cars and light trucks is
projected to increase at an annual rate of 2.4 percent in
the reference case, as a result of high per capita incomes
and low fuel prices. The higher fuel prices in the carbon
reduction cases are projected to lower the growth rate of
horsepower for cars to 1.9 percent between 1996 to 2010
in the 1990+24% case, 1.2 percent in the 1990+9% case,
and 0.3 percent in the 1990-3% case (Figure 59).
1990+24%
1990+9%
1990-3%
1990+24%
1990+9%
1990-3%
0102030405060HorsepowerCarsLightTrucks20102020Figure 59. Projected Reductions From Reference
Case Projections of Car and Light Truck
Horsepower in the Carbon Reduction
Cases, 2010 and 2020
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.
D080398B, and FD03BLW.D080398B.
Fuel Consumption. Reductions in fuel use by light-duty
vehicles (cars, pickup trucks, vans, and sport utility
vehicles) are projected to account for more than two-
thirds of the reduction in transportation energy consumption in 2010 in the carbon reduction cases relative
to the reference case projections. In the reference case,
light-duty vehicles are responsible for 57 percent of all
transportation use in 2010 (Figure 60). The difference in
gasoline consumption by light-duty vehicles (Figure 61)
results from both a decline in vehicle-miles traveled and
an increase in new car and light truck efficiency in
response to higher gasoline prices and lower levels of
disposable income. As fuel-saving technologies penetrate the light-duty vehicle market, higher fuel efficiencies lower the cost of driving per mile, which increases
vehicle travel, offsetting some of the fuel savings.51 The
increase in fuel efficiency also reduced the demand for
gasoline, leading to lower gasoline prices than would
otherwise have occurred. Gasoline prices in real 1996
dollars in 2010 are projected to be 14 cents per gallon
higher in the 1990+24% case than in the reference case,
30 cents per gallon higher in the 1990+9% case, and 55
cents per gallon higher in the 1990-3% case. Comparable
increases in gasoline prices were last seen during the oil
crisis of 1973-1974 (33 cents a gallon in 1996 dollars) and
during the oil embargo of 1979-1980 (47 cents a gallon).
Light-DutyVehicles56.7%
FreightTrucks19.0%Aircraft16.1%
Rail2.1%
Marine6.1%
Figure 60. Projected Fuel Consumption in the
Transportation Sector by Mode in the
Reference Case, 2010
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System run KYBASE.D080398A.
GasolineDistillateJetFuel05101520QuadrillionBtuReference1990+24%1990+9%1990-3%
Figure 61. Projected Fuel Consumption in the
Transportation Sector by Fuel Type,
2010
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.
D080398B, and FD03BLW.D080398B.
51This secondary effect has been estimated at about 10 to 12 percent. See L.A. Greening and D.L. Greene ÒEnergy Use, Technical Efficiency, and the Rebound Effect: A Review of the Literature,Ó draft report prepared for the Office of Policy Analysis and International Affairs,
U.S. Department of Energy (Washington, DC, November 6, 1997).
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Air Travel
Personal, business, and international air travel are
expected to decline in response to higher jet fuel prices
and higher ticket prices in the carbon reduction cases, as
compared with the reference case, from 2005 through
2015. The projected levels of air travel in 2010 are 1.4 percent lower in the 1990+24% case than in the reference
case, 7.4 percent lower in the 1990+9% case, and 16.0 percent lower in the 1990-3% case. Higher fuel prices in 2010
are projected to increase ticket prices by 5 percent, 13
percent, and 23 percent in the 1990+24% case, 1990+9%
and 1990-3% cases, respectively, over the reference case
prices. Lower merchandise exports (0.9 percent lower in
the 1990+24% case, 2.5 percent in the 1990+9% case, and
4.9 percent in the 1990-3% case than in the reference
case) have comparable effects on dedicated air freight
travel.
Between 2005 and 2008, air travel is projected to decline
by 1.2 percent annually in the 1990-3% case as a result of
a 19-percent average annual increase in jet fuel prices. In
comparison, air travel declined by 2.2 percent from 1980
to 1981, when jet fuel prices increase by 49 percent. Similar to light-duty vehicles, differences in the responses to
higher fuel prices between history and the carbon reduction cases can be explained by comparing growth rates
in income levels. Income during 2005 to 2008 is expected
to increase by 0.8 percent annually in the carbon reduction cases, however from 1980 to 1981 income was rising
even faster at 2.3 percent per year, which mitigated the
decline in air travel.
In 2010, the projected use of jet fuel is lower by 1.4 percent in the 1990+24% case than in the reference case, by
6.6 percent in the 1990+9% case, and by 14.2 percent in
the 1990-3% case (Figure 61). Jet fuel prices are projected
to be 15 cents per gallon higher than in the reference case
in 2010 in the 1990+24% case, 34 cents per gallon higher
in the 1990+9% case, and 63 cents per gallon higher in
the 1990-3% case.
Only relatively minor changes in the average fuel efficiency of new aircraft are expected to result from the
imposition of carbon reduction targets. For example, in
the 1990+9% case, new aircraft fuel efficiency is projected to improve at an annual rate of just 0.9 percent
between 1996 and 2010, compared with the 0.7-percent
rate projected in the reference case. As a result, the average efficiencies projected for the entire U.S. stock of aircraft are nearly the same in the two cases (Figure 62).
Less air travel is expected in the carbon reduction cases
than in the reference case, leading to slower rates of aircraft stock turnover, which in turn limit the penetration
of new aircraft into the aircraft stock. Higher fuel prices
and lower air travel in the carbon reduction cases lower
the demand for wide-body aircraft, which have higher
efficiencies in terms of seat-miles per gallon than do
narrow-body aircraft. In addition, near-term aircraft
technologies that can improve fuel efficiency are limited,
and they are not expected to be cost-effective even in the
1990-3% case. Among the six advanced aircraft technologies available by 2010, only weight-reducing materials and ultra-high-bypass engines, which are currently
in use, are expected to penetrate the market (Table 12);
and only the ultra-high-bypass engine technology is
projected to achieve significant penetration (more than
90 percent) by 2010, and then only in the 1990-3% case or
the high technology sensitivity case described below.
NewAircraftStockAircraft0102030405060708090Seat-MilesperGallonReference1990+9%LowTechnology1990+9%1990+9%HighTechnologyFigure 62. Projected New and Stock Aircraft Fuel
Efficiency, 2010
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FREEZE09.D080798A, FD09ABV.
D080398B, and HITECH09.D080698A.
Table 12. Projected Penetration for Selected
Advanced Technologies for Aircraft,
2010
(Percent of New Sales)
Technology Reference 1990+9%
1990+9%
High
Technology
Ultra-High-Bypass
Engines............ 9 90 77
Weight-Reducing
Materials ........... 85 85 96
Advanced
Aerodynamics....... 0 0 96
Thermodynamics .... 0 0 56
Source: Office of Integrated Analysis and Forecasting, National Energy Mod-
eling System runs KYBASE.D080398A, FD09ABV.D080398B, and
HITECH09.D080698A.
Freight Trucks, Rail, and Shipping
The projected demand for distillate fuel, used primarily
for freight trucks and rail, is also lower in the carbon
reduction cases than in the reference case in 2010Ñby 2
percent in the 1990+24% case, by 4.9 percent in the
1990+9% case, and by 8.3 percent in the 1990-3% case
(see Figure 61). Distillate fuel prices are projected to be
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
15 cents per gallon higher in the 1990+24% case, 37 cents
per gallon higher in the 1990+9% case and 68 cents
higher in the 1990-3% case. These increases are larger
than those projected for gasoline because of the higher
carbon content of distillate fuel.
Higher fuel prices do not result in as much change in
travel and efficiency for freight trucks and rail as they do
for light-duty vehicles. Because of the slow turnover in
the stock of freight trucks and rail and the high power
requirements of the engines used to move freight, fuel
savings are limited. The main source of reductions in
distillate fuel use is the response to overall lower economic activity and demand for goods by 2010 in the carbon reduction cases, leading to lower freight travel for
both trucks and rail. Lower demand for goods in the
1990+24%, 1990+9% and 1990-3% cases results in levels
of freight truck travel that are 1.3 percent, 2.4 percent
and 4.9 percent lower, respectively, in 2010 than projected in the reference case. Declines in coal consumption and production also lead to further cuts in rail travel
as described below.
The potential for improvement in fuel economy for
freight trucks is also limited. In the reference case, the
fuel efficiency of new freight trucks is projected to
increase by only 0.6 percent per year between 1996 and
2010. Even with higher distillate fuel prices in the 19903% case, the efficiency for new freight trucks improves at
an annual rate of only 0.8 percent. As a result of the
lower demand for goods and slower turnover in the
stock of freight trucks projected in the 1990+9% case
relative to the reference case, there is almost no difference in the projected average stock efficiencies for the
two cases in 2010 (Figure 63).
NewFreightTrucksStockFreightTrucks0246810MilesperGallonReference1990+9%LowTechnology1990+9%1990+9%HighTechnologyFigure 63. Projected New and Stock Freight Truck
Fuel Efficiency, 2010
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FREEZE09.D080798A, FD09ABV.
D080398B, and HITECH09.D080698A.
The number of advanced technologies available for
freight trucks is relatively small. Those with the greatest
potential are advanced aerodynamics, the turbo-
compound diesel engine, and the LE-55 heat engine,
with expected marginal fuel efficiency improvements of
approximately 25, 10, and 17 percent, respectively
(Table 13). In all the carbon reduction cases, the
advanced aerodynamics technology is projected to
achieve the greatest efficiency improvements and
highest penetration rates for both medium-and heavy-
duty trucks. The turbo-compound diesel engine and the
LE-55 heat engine do not penetrate the market until after
2010, except in the high technology sensitivity cases.
In percentage terms, the projections for rail and ship
freight travel in 2010 show the sharpest reductions
relative to the reference case in the carbon reduction
cases. Rail freight travel is 9 percent, 23 percent, and 32
percent lower in 2010 in the 1990+24%, 1990+9%, and
1990-3% cases than in the reference case. Since more than
40 percent of rail travel is for coal transportation, the
lower rail travel in the carbon reduction cases is
primarily due to the projected reductions in coal
production of 20 percent, 52 percent, and 71 percent in
the 1990+24%, 1990+9%, and 1990-3% cases relative to
the reference case. Domestic freight travel by ship is
projected to be 3 percent, 6 percent, and 10 percent lower
in the three cases than in the reference case. Domestic
shipping is not expected to be affected as adversely by
the decline in coal production as is rail traffic; however,
with lower demand for goods and industrial production
Table 13. Projected Penetration of Selected
Technologies for Freight Trucks, 2010
(Percent of New Sales)
Technology Reference 1990+9%
1990+9%
High
Technology
Medium Trucks
Improved Tires
and Lubricants ........ 0 0 0
Electronic
Engine Controls ....... 0 0 5
Advanced
Drag Reduction ........ 23 34 45
Turbo Compound Diesel . 0 2 8
LE-55 Heat Engine ..... 0 0 13
Heavy Trucks
Improved Tires
and Lubricants ........ 0 3 98
Electronic
Engine Controls ....... 0 4 98
Advanced
Drag Reduction ........ 100 100 100
Turbo Compound Diesel . 1 1 35
LE-55 Heat Engine ..... 0 0 10
Source: Office of Integrated Analysis and Forecasting, National Energy Mod-
eling System runs KYBASE.D080398A, FD09ABV.D080398B, and HITECH09.
D080698A.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
in the carbon reduction cases, domestic shipping is also
projected to be lower.
Like freight truck and rail travel, shipping is affected
more by the impacts of carbon prices on travel and shipping requirements than by the direct impacts of higher
fuel costs. High-carbon residual fuel has the largest projected price increases of all the transportation fuels, with
increments of 19 cents per gallon in the 1990+24% case,
46 cents in the 1990+9% case, and 84 centsÑalmost 100
percentÑin the 1990-3% case relative to the prices projected for 2010 in the reference case.
Approximately 15 to 17 percent of the drop in total fuel
consumption in 2010 in the carbon reduction cases is
attributed to aircraft, 6 to 7 percent to freight trucks, 4 to
6 percent to rail engines, and 1 percent to marine
engines. The relative energy consumption shares for the
major transportation modes and fuels do not vary
significantly across the cases (Table 14).
Table 14. Projected Fuel Consumption Shares in
the Transportation Sector by Fuel and
Travel Mode, 2010
(Percent of Total)
Projection Reference
1990
+24%
1990
+9%
1990
-3%
Fuel
Gasoline .......... 58 58 57 56
Distillate .......... 19 19 20 21
Jet Fuel ........... 16 16 16 16
Residual .......... 4 4 4 5
Alternative Fuels .... 3 3 3 3
Travel Mode
Light-Duty Vehicles. . 57 56 56 55
Freight Trucks...... 17 19 20 20
Aircraft............ 16 16 16 16
Rail .............. 2 2 2 2
Marine............ 6 6 6 7
Source: Office of Integrated Analysis and Forecasting, National Energy Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.
D080398B, and FD03BLW.D080398B.
Both freight rail and domestic shipping efficiencies are
projected to remain at reference case levels in the carbon
reduction cases. Stock turnover will virtually cease,
because rail ton-miles traveled are lower by 32 percent in
2010 in the 1990-3% case than in the reference case, and
domestic shipping travel is 10 percent lower. Also, with
the loss in revenue associated with the projected lower
levels of travel, efficiency improvements will be difficult
to achieve.
Alternative-Fuel Vehicles
According to consumer surveys, alternative-fuel vehicle
sales are dependent on vehicle price, the cost of driving
per mile, vehicle range, fuel availability, and commercial
availability. In 2010, alternative-fuel vehicle sales as a
percent of light-duty vehicle sales are projected to
increase to 11.98 percent in the 1990+24% case, 12.07 percent in the 1990+9% case, and 12.10 percent in the 19903% case from 11.91 percent in the reference case. The
projected market shares for alternative-fuel vehicles are
higher in the carbon reduction cases primarily because
higher fuel prices would encourage consumers to take
advantage of the higher fuel efficiencies and lower costs
of driving projected for some alternative-fuel vehicles
relative to gasoline vehicles. In addition, as the fuel efficiency of alternative-fuel vehicles improves, their driving range will increase.
Although alternative-fuel vehicle sales increase in percentage terms relative to the reference case in 2010, the
actual number of alternative-fuel vehicles sold is
expected to be smaller in the carbon reduction cases as a
result of projected declines in light-duty vehicle sales
overall. In the reference case alternative-fuel vehicle
sales are projected to be approximately 1.79 million
vehicles in 2010, whereas sales range between 1.68 and
1.75 million vehicles in the 1990+24%, 1990+9%,
1990+9% , and 1990-3% cases. Similar results are projected for alternative-fuel consumption as a percentage
of total transportation fuel use in 2010. Although the
projected cost of driving per mile is lower for some
alternative-fuel vehicles than for gasoline vehicles in
some of the carbon reduction cases, it would still be
more costly to drive an alternative-fuel vehicle than a
gasoline vehicle. The purchase prices for most
alternative-fuel vehicles still would be higher than those
for conventional gasoline-powered vehicles, and additional driving costs would be incurred as the result of
lower vehicle range and limited availability of fuel. Also,
with higher projected fuel prices, vehicle-miles traveled
are expected to be reduced for all vehicles, including
those that use alternative fuels. Finally, the higher efficiencies of alternative-fuel vehicles would lower their
total fuel consumption.
Sensitivity Cases
To examine the effects of technology improvements on
energy use and prices, two sensitivity cases were analyzed for the transportation sector. The 1990+9% low
technology sensitivity case was designed to hold average new vehicle fuel efficiencies at their 1998 levels
throughout the forecast period. The implication is that
stock turnover and travel reductions would have to
compensate for the lack of fuel efficiency improvements
in order to meet the carbon reduction targets. The
1990+9% high technology sensitivity case was designed
to illustrate the effects of advanced fuel-saving technologies on transportation fuel efficiency, fuel consumption,
and carbon emissions. This sensitivity case generally
assumes that the costs of new technologies will be
reduced, the marginal fuel efficiency benefits will be
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Mass Transit and Carpooling
An issue for the transportation sector is whether the ratifi-Available statistics support the contention that the lower
cation of the Kyoto Protocol by the United States will lead levels of vehicle-miles traveled associated with the carbon
to increased use of mass transit and carpooling. Automo-reduction cases do not necessarily imply increased use of
bile transportation is a major contributor to air pollution mass transit. According to the American Public Transit
and greenhouse emissions, and a cutback in this area Association, all forms of mass transit in terms of
would be desirable. U.S. transportation patterns make passenger-miles decline during periods of high fuel
this unlikely, however, in spite of the fact that the carbon prices.d Transit rail passenger-miles, which include light
reduction cases in this analysis project higher gasoline and heavy rail travel, declined by nearly 10 percent from
prices and lower levels of vehicle-miles traveled. 1973 to 1974 and by 5 percent from 1979 to 1981, even
The United States consumes far more energy per capita though real gasoline prices concurrently rose by 28 per-
for transportation than any other developed country, cent during both periods. Similar trends occurred in com
with U.S. passenger travel dominated by the automobile. muter rail, which experienced declines of almost 8
In 1990, about 86 percent of passenger-miles were percent from 1980 to 1982. Between 1979 and 1982, transit
accounted for by automobiles, and mass transit bus passenger-miles declined by 7 percent and intercity
accounted for less than 4 percent. The U.S. mass transit bus travel by 1 percent, while real gasoline prices
system includes buses, light rail, commuter rail, trolleys, increased by 15 percent. A counter example is the period
subways, and an array of services such as van pools, sub-from 1973 to 1974, when transit bus use rose by 11 persidized taxis, dial-a-ride services, and shared minibus cent, and intercity bus passenger-miles increased by 5
and van rides. Most cities of over 20,000 population have percent. That period was unique, however, because gaso
bus systems, and buses on established routes with set line was often either unavailable or required waits of up
schedules account for over half of all public transit pas-to several hours in gas station lines.
senger trips. About 70 percent of all public transit trips in Carpooling trends, according to the U.S. Census Bureau,
1990, however, were in the 10 cities with rapid rail sys-have declined from approximately 20 percent of the
tems; 41 percent were in New York City and its suburbs.a workforce in 1980 to just over 13 percent in 1990.e The
More recent statistics show that, as of 1995, mass transit National Personal Transportation Survey has reported
accounted for only 0.8 percent of total fuel consumption similar trends in vehicle occupancy rates, which indicate
in the transportation sector.b that from 1977 through 1990, vehicle occupancy rates
One reason for the low usage of mass transit in the United have declined in commuting to and from work, from 1.30
States and the concentration of use in major cities is urban to 1.14 person-miles per vehicle mile.f These occupancy
development that has decreased the importance of his-rates correspond to about one-third of total vehicle-miles
toric central business districts (CBDs). Peak trips in gen-traveled.
eral, and work trips in particular, have become diffuse in Because travelers do not take into account such externaliboth origin and destination and thus not easily served by ties as reducing greenhouse gas emissions when making
mass transit. In 1980 only 9 percent of the workers in their transportation decisions, and past gasoline price
urban areas and only 3 percent of workers living outside increases do not seem to have had an impact, it is unlikely
the central city were employed in the CBDs.c (In Europe, that mass transit and carpooling will increase in the
where population densities are much higher, access to the United States without policy intervention factors such as
workplace is much easier.) Other factors that work higher gasoline taxes and urban and transportation plan-
against mass transit in the United States are a past history ning that facilitates access to workplaces. There are differ-
of low gasoline prices, rising income levels, increasing ing opinions as to the role these factors could play in
numbers of women in the workforce with needs to drop shaping travel patterns. If history, geography, income,
off and pick up children at child care facilities, a move and demographics are the primary determinants of travel
toward less standardization of work hours, and premi-patterns, policy may play only a minor role in changing
ums placed on personal independence and time saved by energy use; but if instruments of public policy are pridriving rather than making use of mass transit. The same mary travel determinants, then there is a large potential
factors affect the use of carpooling. for policy to reduce energy useg and alter mass transit and
carpooling patterns.
aU.S. Congress, Office of Technology Assessment, Saving Energy in U.S. Transportation, OTA-ETI-589 (Washington, DC, July 1994),
pp. 5-6.
bS. Davis, Transportation Energy Databook No. 17, prepared for the Office of Transportation Technologies, U.S. Department of Energy
(Oak Ridge, TN: Oak Ridge National Laboratory, August 1997), p. 2-12.
cU.S. Congress, Office of Technology Assessment, Saving Energy in U.S. Transportation, OTA-ETI-589 (Washington, DC, July 1994),
pp. 5-6.
dAmerican Public Transit Association, 1994-1995 Transit Fact Book (Washington, DC, February 1995), pp. 106-107.
eS. Davis, Transportation Energy Databook No. 17, prepared for the Office of Transportation Technologies, U.S. Department of Energy
(data provided by the Journey-to-Work and Migration Statistics Branch, Population Division, U.S. Bureau of the Census) (Oak Ridge,
TN: Oak Ridge National Laboratory, August 1997), p. 2-12.
fFederal Highway Administration, National Personal Travel Survey: 1990 NPTS Databook, Vol. II, Chapter 7 (Washington, DC,
November 1993).
gU.S. Congress, Office of Technology Assessment, Saving Energy in U.S. Transportation, OTA-ETI-589 (Washington, DC, July 1994),
pp. 5-6.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
higher, and the advanced technologies will be commercially available at earlier dates than in the reference case
or the carbon reduction cases.52
Higher projected carbon prices in the low technology
sensitivity case lead to higher prices for all transportation fuels. In 2010, average fuel prices in the transportation sector are projected to be 14 percent higher in the
1990+9% low technology case than in the 1990+9% case.
Gasoline prices are projected to be about 19 cents per
gallon higher, jet fuel prices 21 cents per gallon higher,
distillate fuel prices 22 cents per gallon higher, and
residual fuel prices 26 cents per gallon higher.
Both fuel efficiency and travel are lower in the low technology case than in the 1990+9% case. Higher fuel prices
would affect travel both directly and through their secondary impacts on the general levels of macroeconomic
activity, disposable income, and freight movement. Of
all travel modes, vehicle-miles traveled by light-duty
vehicles are the most responsive to the higher fuel prices
in the 1990+9% low technology case, with a 5.1-percent
reduction from the projected level in the 1990+9% case in
2010. Air travel is reduced by a similar percentage, 5.5
percent, whereas smaller reductions are projected for
freight, rail, and domestic shipping travel (0.8 percent,
3.1 percent, and 0.9 percent, respectively). Total projected fuel consumption in 2010 is higher in the low technology case than in the 1990+9% case, because fuel
efficiency does not improve as rapidly.
With lower carbon prices and lower fuel prices in the
1990+9% high technology sensitivity case, more travel is
expected than in the 1990+9% case. Despite the higher
travel projection, however, more rapid improvements in
new vehicle and stock fuel efficiencies result in lower
fuel consumption in the high technology case, with
higher fuel efficiencies outweighing the projected
increases in vehicle-miles traveled that result from lower
projected fuel prices. Average transportation fuel prices
in 2010 are 9.6 percent lower in the 1990+9% high technology sensitivity case than in the 1990+9% case. Gasoline prices are projected to be 14 cents per gallon lower in
2010, jet fuel prices 13 cents per gallon lower, distillate
fuel prices 14 cents per gallon lower, and residual fuel
prices 16 cents per gallon lower.
Comparing across the travel modes, light-duty vehicles
hold the greatest potential for reducing fuel consumption and carbon emissions with more rapid technology
advances (Figure 64). Not only do light-duty vehicles
Light-DutyVehiclesTrucksAircraftRailMarineTotal0123QuadrillionBtu1990+9%
LowTechnology1990+9%1990+9%
HighTechnologyFigure 64. Projected Reductions From Reference
Case Projections of Transportation
Sector Fuel Consumption in High and
Low Technology Sensitivity Cases, 2010
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FREEZE09.D080798A, FD09ABV.
D080398B, and HITECH09.D080698A.
consume more fuel in total than the other vehicle types
(more than 56 percent of all transportation fuel use in
1996), they also have the greatest potential for advanced
technology penetration. In the 1990+9% high technology
sensitivity case, light-duty vehicles are projected to
account for 65 percent of the reduction in transportation
fuel use relative to the 1990+9% case, compared with 20
percent for trucks, 11 percent for aircraft, 4 percent for
rail, and 1 percent for marine.
Fuel-saving technologies for conventional light-duty
vehicles in the high technology case are assumed to have
approximately 50 percent lower marginal technology
costs and 30 percent higher marginal fuel efficiency
improvements than those for gasoline vehicles. All conventional technologies achieve lower sales penetration
rates in the high technology case than in the 1990+9%
case, due to lower fuel prices (Table 11); however,
because the marginal fuel efficiencies are also higher
than in the 1990+9% case, the total fuel efficiency
improvement is larger in the high technology case.
With lower marginal costs and earlier introduction dates
in the high technology sensitivity, most new aircraft
technologies reach significantly higher penetration rates
than in the 1990+9% case with reference technology
(Table 12). The penetration rate for ultra-high-bypass
52High technology assumptions were derived from the following sources: light-duty vehicle conventional technology attributes from J.
DeCicco and M. Ross, An Updated Assessment of the Near-Term Potential for Improving Automotive Fuel Economy, American Council for an
Energy-Efficient Economy (Washington, DC, November 1993); light-duty alternative fuel vehicle cost and performance attributes from U.S.
Department of Energy, Office of Transportation Technologies, Program Analysis Methodology: Final ReportÑQuality Metrics 98 Revised (Washington, DC, April 1997); freight trucks from U.S. Department of Energy, Office of Transportation Technologies, OHVT Technology Roadmap
(Washington, DC, October 1997), and conversations with Frank Stodolsky, Argonne National Laboratory, and Mr. Suski, American Trucking Association; air from conversations with Glenn M. Smith, National Aeronautics and Space Administration.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
engines is lower in the high technology case, because
they are partially displaced by advanced thermodynamic engines. Substantial fuel efficiency improvements
result from the penetration of weight-reducing materials, advanced aerodynamics, and advanced thermodynamic engines, which can potentially achieve efficiency
improvements of 15 percent, 18 percent, and 20 percent,
respectively.
Fuel efficiency for new freight trucks rises by more than
1 mile per gallon by 2010 in the high technology case
relative to the 1990+9% case, primarily because of the
penetration of the turbo compound diesel, LE-55 heat
engine, improved tires and lubricants, and electronic
engine controls on heavy-duty trucks (Table 13). Both
advanced engine technologiesÑthe turbo compound
diesel and LE-55 heat engineÑare diesel technologies,
which improve fuel economy by 10 percent and 23
percent, respectively.
The high technology case assumes that the U.S. Department of Energy Office of Transportation Technologies
program goals53 for alternative-fuel vehicle cost and performance improvements will be met. Generally these
program goals include a reduction of 50 to 66 percent in
the marginal price difference between comparable gasoline vehicles and electric or electric hybrid vehicles, and
a 75-percent reduction in the difference for fuel cell vehicles. Fuel efficiency improvements are assumed to be
230 to 300 percent greater for electric and electric hybrid
vehicles and 250 percent greater for fuel cell vehicles
than for gasoline vehicles. These fuel efficiency
improvements are also assumed to result in travel
ranges that are 57 percent greater for electric hybrid
vehicles and 20 percent greater for fuel cell vehicles than
the range for similar sized gasoline vehicles. Total
alternative-fuel vehicle sales in the 1990+9% high technology case in 2010 are projected to make up almost 19
percent of all light-duty vehicle sales, compared with
just over 11 percent in both the reference and 1990+9%
cases. The projected shares for different alternative-fuel
vehicle types are shown in Table 15.
In order for alternative-fuel vehicles to displace large
quantities of gasoline use, they must penetrate the
market early enough to replace gasoline vehicles and
Table 15. Projected Alternative-Fuel Vehicle
Shares of New Light-Duty Vehicle Sales
by Type in the High Technology Cases,
2010
(Percent)
Vehicle Type Sales Share
Flex-Fuel Methanol and Ethanol ........ 9.1
Dedicated Methanol and Ethanol........ 2.1
Electric ............................ 1.2
Hybrid Electric/Gasoline............... 1.3
Hybrid Electric/Diesel................. 1.6
Bi-Fuel CNG and LPG ................ 1.0
Dedicated CNG and LPG.............. 2.5
Fuel Cell Gasoline ................... 0.02
Fuel Cell Methanol ................... 0.01
Diesel Direct Injection ................ 2.1
CNG = compressed natural gas. LPG = liquefied petroleum gas
(propane).
Source: Office of Integrated Analysis and Forecasting, National Energy Modeling System run KYBASE.D080398A.
then sustain high sales volumes. Displacement of gasoline may be limited, however, because the vast majority
of the projected increase in alternative-fuel vehicle sales
consists of alcohol flexible-fuel vehicles, which are
expected to have only slightly higher fuel efficiencies
than gasoline vehicles. They will also use only 15percent blends of E85 and M85 and will more frequently
be consuming gasoline than the alternative fuel.
For alternative-fuel vehicles to maintain a larger share of
the vehicle market, they will need to have lower costs,
higher performance, and earlier availability dates than
projected in this analysis. Simultaneously, higher fuel
prices will be needed to send market signals to both
consumers and vehicle producers. The high technology
case indicates both of these points: fuel-saving
technology becomes available and is purchased in 2005,
but its advantage is quickly offset by reductions in
gasoline consumption, which lead to lower gasoline
prices. Consequently, as fuel prices begin to decline after
2008, consumers tend to demand higher performance
and larger vehicles, and manufacturers respond by
designing and producing larger, more profitable
models, such as sport utility vehicles.
53U.S. Department of Energy, Office of Transportation Technologies, Program Analysis Methodology: Final ReportÑQuality Metrics 98
(Washington, DC, April 16, 1997).
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
4. Electricity Supply
Introduction
This chapter discusses the electricity supply side options
under various domestic carbon emissions reduction cases, particularly the 24-percent-above-1990
(1990+24%), 9-percent-above-1990 (1990+9%) and 3percent-below-1990 (1990-3%) cases. The impacts on
electricity sector fuel use, capacity expansion and retirement decisions, electricity prices, and carbon emissions
are discussed. In addition, the results of sensitivity cases
incorporating alternative assumptions about improvements in technology costs and performance, the potential role for new nuclear power plants, and reducing
impacts on the coal industry are also discussed. The
effects of demand-side decisions (i.e., consumer appliance choices and usage, as discussed in Chapter 3) that
would reduce the demand for electricity are also considered.
During the approximately 100-year history of the
electricity supply industry, the key fuels used to meet
the ever-increasing demand for electricity have changed
as new generating technologies have emerged and fuel
prices varied (Figure 65). Beginning with small
hydroelectric facilities just before the turn of the century,
the industry then turned to fossil fuels. Among the fossil
fuels, coal has almost always played a major role in U.S.
electricity generation, and it remains the dominant fuel
today. Oil and natural gas use has varied, depending on
their respective prices. In fact, concerns about future oil
and natural gas prices contributed to the emergence of
nuclear power plants in the 1960s. In todayÕs market,
coal-fired power plants produce just over half of the
electricity used in the United States, nuclear plants 19
percent, natural gas plants 14 percent, and hydroelectric
plants about 10 percent. The remaining 7 percent comes
from oil-fired plants and plants using other fuels such as
municipal solid waste, wood, and geothermal and wind
power.
In the reference case, which does not include the Kyoto
Protocol, the power generation sector is expected to
become more energy-efficient over the next 20 years as
new, more efficient power plants are built. At the same
time, however, dependence on fossil fuels, especially
natural gas and coal, is expected to increase, leading to
significant growth in power plant carbon emissions.
Coal is expected to remain the dominant fuel as existing
plants are used more intensively, but generation from
195019601970198019902000201020200,0001,0002,0003,0004,0005,000BillionKilowatthoursCoalGasOilNuclearHydroRenewable/OtherHistoryProjectionsFigure 65. Electricity Generation by Fuel in the
Reference Case, 1949-2020
Note: Data on nonutility generation are not available for years before
1989, but it was small. In 1989, nonutility generation accounted for
6 percent of total U.S. electricity generation.
Sources: History: Energy Information Administration, Annual Energy Review
1997, DOE/EIA-0384(97) (Washington, DC, July 1998). Projections: Office of
Integrated Analysis and Forecasting, National Energy Modeling System run
KYBASE.D080398A.
natural gas is expected to increase rapidly, with gas-
fired plants making up the vast majority of new capacity
additions. Of the major non-carbon-based fuels,
hydroelectric generation is expected to change very
little, and nuclear generation is expected to decline as
older, more costly plants are retired. Looked at another
way, while the efficiency of the generation sector,
expressed as the amount of energy in terms of British
thermal units (Btu) needed to produce each kilowatt-
hour of electricity, is expected to improve, increasing
dependence on fossil fuels will lead to more rapid
growth in electricity sector carbon emissions than in
electricity sales (Figure 66). Without the improvement in
efficiency, growth in fossil fuel use would match the
growth in fossil-fired generation.
Although the costs of non-carbon-based generating
technologies have fallen, they still are not widely competitive with fossil fuel technologies. As a result, the
most economical options available to electricity suppliers for meeting the demand for electricity over the next
20 years are existing coal plants and new natural gas
plants. In 1995, the average operating cost of coal-fired
power plants was 1.8 cents per kilowatthour. Only
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Figure 66. Projections of Electricity Sales, Carbon
Emissions, Fossil Fuel Use, and
Fossil-Fired Generation, 1997-2020
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System run KYBASE.D080398A.
66 percent of their maximum potential output was
needed, however, to meet the 1996 level of demand.
Over the next 20 years, as the demand for electricity
grows, the utilization of coal-fired plants is expected to
approach 80 percent. For new capacity additions, the
low capital costs and high operating efficiencies of
natural-gas-fired combined-cycle plants make them the
most economical choice for most uses.
Electricity suppliers have a variety of options available
for reducing their carbon emissions. The degree to
which each of the options is employed will depend on
the level of reduction required and the resultant carbon
price (i.e., the market value of a Òcarbon emissions permitÓ) that evolves in the marketplace. Many of the
options may require a significant financial incentive before they become economically attractive. Among the key
carbon reduction options available to electricity suppliers are reducing the use of relatively carbon-intensive
power plants (particularly coal-fired plants), increasing
the use of less carbon-intensive technologies (mainly
natural-gas-fired plants), the use of Òcarbon-freeÓ technologies (i.e. wind, solar, biomass, geothermal, and
nuclear), improving the operating efficiencies of existing
plants, and investing in demand-side technologies that
reduce electricity consumption.
In the short run, before a large number of new plants can
be built, power suppliers will have to reduce carbon
emissions by increasing the use of less carbon-intensive
plants. For example, in todayÕs market, most oil and
natural gas steam plants are not used very intensively
because of their relatively high operating costs. If carbon
reduction efforts are made, however, their use is
likely to increase, because they produce less carbon per
kilowatthour than do coal-fired plants. In the longer run,
power suppliers are more likely to turn to new, less
carbon-intensive or carbon-free plants.
In this analysis, electricity producers are assumed to
have 15 new generating technologies to choose from
when new resources are needed, or when it is no longer
economical to continue operating existing plants (Table
16). The lead times in the tables represent the time
needed for site preparation and construction. Environmental licensing may take longer in some cases. The
first-of-a-kind costs represent the cost of building a plant
when the technology first becomes available, which tend
to be relatively high until experience is gained with the
technology. The nth-of-a-kind costs represent costs for
technologies when they have matured. For technologies
that are already considered mature, the two costs will be
the same. Investors in the generation market are
assumed to make their decisions by reviewing each technologyÕs current and future capital, operations and
maintenance, and fuel costs. Both current and expected
future costs are considered, because generating assets
require considerable investment and last many years.
Therefore, developers are assumed to evaluate the costs
of building and operating a plant for 30 years when
making their decisions.54 If the Kyoto Protocol is
enacted, developers will also have to consider the relative level of carbon emissions from each technology, as
well as the expected carbon prices. Depending on the
carbon price, the economic decision could be tilted
toward technologies that emit less carbon per unit of
electricity produced.
Overall, because of the relatively wide variety of options
available to them, electricity suppliers are expected to
account for a disproportionately large share of projected
carbon reductions. Nationally, to meet an emissions target 9 percent above 1990 levels, overall carbon emissions
in 2010 would have to be reduced by 18 percent from
their projected level in the reference case, which is 33
percent above 1990 levels. But in order to meet the target, emissions from the electricity sector in the 1990+9%
case are reduced by 39 percent in 2010 relative to the reference case (Figure 67). The situation is similar in the
1990-3% case: electricity sector carbon emissions in 2010
are 54 percent lower than the reference case level. The
reduction in carbon emissions is projected to be accomplished through a combination of fuel switching,
improvements in end-use efficiency, and improvements
in generator efficiency (Figure 68).
In the carbon reduction cases, carbon emissions in the
electricity sector are projected to begin falling even
before the enactment of the Kyoto Protocol, because
power plant developers are assumed to consider future
costs in their investment decisions. As the implementation date of the Kyoto Protocol approaches, it is assumed
54Capital costs are assumed to be recovered over the first 20 years of this period.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Table 16. Cost and Performance Characteristics of New Fossil, Renewable, and Nuclear
Generating Technologies
Technology
Size
(MW)
Lead
Time
(Years)
First
Electricity
Date
Overnight
Capital Costa
(1996 Dollars
per kWh) Variable
O&M
(1996 Mills
per kWh)
Fixed O&M
(1996
Dollars
per kW)
Heat Rate
(Btu per kWh) Carbon
Emissions
(Pounds
per MWh)
First-
of-a-
Kind
nth-
of-a-
Kind
First-
of-a-
kind
nth-
of-a-
Kind
Pulverized Coal
(95% Scrubber) ........... 400 4 2001 1,079 1,079 3.25 22.5 9,585 9,087 519
Advanced Coal
(IGCC) .................. 380 4 2001 1,833 1,206 1.87 24.2 8,470 7,308 417
Oil/Gas Stream
(Conventional) ............ 300 2 1998 991 991 0.5 30.0 9,500 9,500 296
Combined-Cycle
(Conventional, F-Frame) .... 250 3 2000 440 440 2.0 15.0 8,030 7,000 250
Combined-Cycle
(Advanced, G-& H-Frame) . . 400 3 2000 572 400 2.0 13.8 6,985 6,350 198
Combustion Turbine
(Conventional) ............ 160 2 1999 325 325 5.0 4.0 11,900 10,600 330
Combustion Turbine
(Advanced Turbine System) . 120 2 1999 458 320 5.0 5.7 9,700 8,000 249
Fuel Cell
(Molten Carbonate) ........ 10 2 2003 2,189 1,440 2.0 14.4 6,000 5,361 167
Nuclear (Evolutionary
Advanced Reactor) ........ 1,300 5 2010 2,356 1,550 0.4 55.0 10,400 10,400 0
Biomass ................. 100 4 2005 2,243 1,476 5.2 43.0 8,911 8,224 0
Geothermalb .............. 50 4 1996 NA 2,025 0.0 95.7 32,391 NA 0
Municipal Solid Wastec ..... 30 1 1995 6,403 5,289 5.4 0.0 16,000 16,000 0
Solar Thermald ............ 100 3 2000 2,903 e1,910 0.0 46.0 NA NA 0
Solar Photovoltaic ......... 5 2 1997 4,556 e3,185 0.0 9.7 NA NA 0
Wind.................... 50 3 1997 1,235 965 0.0 25.6 NA NA 0
aOvernight capital cost plus project contingencies.
bBecause geothermal cost and performance parameters are specific for each of the 51 sites in the database, the value shown is an average for the
capacity built in 2000.
cBecause municipal solid waste does not compete with other technologies in the model, these values are used only in calculating the average
costs of electricity.
dSolar thermal is assumed to operate economically only in Electricity Market Module regions 2, 5, and 10-13, that is, West of the Mississippi River,
because of its requirement for significant direct, normal insulation.
eCapital costs for solar technologies are net of (reduced by) the 10 percent investment tax credit.
kW = kilowatt. kWh = kilowatthour. MW = megawatt. MWh = megawatthour. NA = not available. O&M = operations and maintenance costs.
Sources: Most values are derived by the Energy Information Administration, Office of Integrated Analysis and Forecasting from analysis of reports and discussions with
various sources from industry, government, and the National Laboratories, with the following specific sources: Solar ThermalÑCalifornia Energy Commission
Memorandum, Technology Characterization for ER94 (August 6, 1993). PhotovoltaicÑElectric Power Research Institute, Technical Assessment Guide, EPRI-TAG
1993. Municipal Solid WasteÑEPRI-TAG 1993.
that developers will incorporate their expectations of unprecedented historically. Even during the 1960s and
carbon prices into their plans for new capacity additions, 1970s, when nuclear generation grew rapidly, the
and that more lower-carbon generating capacity will be change in fuel use patterns was not as dramatic as would
brought on line than would have been in the absence of be required in this case. In the 1990+24% case, the shift is
the expected carbon reduction mandate. less pronounced, but coal-fired generation still is
projected to be 17 percent lower in 2010 and 40 percent
lower in 2020 than in the reference case. Across the
carbon reduction cases, the projections show aTrends in Fuel Use consistent shift away from coal to natural gas and
and Generating Capacity renewables for electricity generation. In addition,
nuclear generation remains near current levels, and the
To reduce power plant carbon emissions in the 1990+9% demand for electricity falls as the carbon reduction goal
case, the mix of fuels used to produce electricity is tightens (Figure 70).
expected to change significantly from historical patterns
(Figure 69). The change required is possible, but it will The shift away from coal-fired generation occurs
be challenging. For example, the shift required to because coal accounts for such a large share of power
stabilize carbon emissions 9 percent above 1990 levels is plant carbon emissions. In 1996, coal-fired power plants
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
1995200020052010201520200100200300400500600700800MillionMetricTons1990+9%
Reference1990+24%
1990-3%
Figure 67. Projections of Carbon Emissions From
the Electricity Supply Sector, 1996-2020
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.
D080398B, and FD03BLW.D080398B.
1995200020052010201520200100200300400500600MillionMetricTonsFuelSwitchingGenerationEfficiencyDemandReductionsFigure 68. Projected Reductions in Carbon
Emissions From the Electricity Supply
Sector, 1990-3% Case, 1996-2020
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A and FD03BLW.D080398B.
Figure 69. Electricity Generation by Fuel, 1990+9%
Case, 1949-2020
Note: Data on nonutility generation are not available for years before
1989, but it was small. In 1989, nonutility generation accounted for
6 percent of total U.S. electricity generation.
Sources: History: Energy Information Administration, Annual Energy Review
1997, DOE/EIA-0384(97) (Washington, DC, July 1998). Projections: Office of
Integrated Analysis and Forecasting, National Energy Modeling System run
FD09ABV.D080398B.
CoalGasOilNuclearHydroelectricBiomassWindGeothermalOtherDemandReductions05001,0001,5002,0002,500-500-1,000BillionKilowatthoursReference1990+24%1990+9%1990-3%
Figure 70. Electricity Generation by Fuel, 2010
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.
D080398B, and FD03BLW.D080398B.
produced an estimated 92 percent of the carbon emissions in the power generation sector. In the reference
case, that share is expected to be 86 percent in 2010; and
in 2020, even though natural-gas-fired generation grows
rapidly, coal plants still are expected to account for 81
percent of total carbon emissions from the electricity sector. Per unit of fuel consumed (Btu), coal plants emit
nearly 80 percent more carbon than do natural gas
plants, and the difference is even greater per megawatthour of electricity generated (Table 17). New natural
gas combined-cycle plants are much more efficient than
existing coal plants, requiring less than two thirds the
amount of fuel (in Btu) to produce a kilowatthour of
electricity. As a result, per megawatthour of electricity
produced, existing coal plants release nearly 3 times as
much carbon into the atmosphere as do the most efficient new natural gas plants.
Coal
Generation
In the carbon reduction cases, the projected decreases
in coal-fired electricity generation are dramatic. In the
1990+24%, 1990+9%, and 1990-3% cases, coal-fired
generation in 2010 is expected to be 18 percent, 53
percent, and 75 percent lower, respectively, than in the
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Table 17. Carbon Emissions From Fossil Fuel Generating Technologies
Technology
Heat Rate
(Btu per
Kilowatthour)
Carbon Emissions
Pounds per
Million Btu
Pounds per
Megawatthour
Coal-Fired Technologies
Existing Capacity ................................... 10,000 57 571
New Capacity Additions .............................. 9,087 57 519
Advanced Coal Technology ........................... 7,308 57 418
Natural-Gas-Fired Technologies
Conventional Turbine ................................ 10,600 32 336
Advanced Turbine .................................. 8,000 32 253
Existing Gas Steam ................................. 10,300 32 326
Conventional Combined-Cycle......................... 7,000 32 222
Advanced Combined-Cycle ........................... 6,350 32 201
Fuel Cell .......................................... 5,361 32 170
Source: Energy Information Administration, Office of Integrated Analysis and Forecasting.
reference case (Figure 71). In 2020, the differences from
the reference case are even larger: 41 percent in the
1990+24% case, 77 percent in the 1990+9% case, and over
96 percent in the 1990-3% case. In 1990-3% case, coal-
fired generation is virtually eliminated. Coal plants
simply are not very economical when carbon prices are
high.
Such reductions in coal use would come at a cost.
Although they are major carbon emitters, existing coal
plants are very economical, and their operating costs
have been falling (Figure 72). Under more stringent
emissions reduction targets, however, with rising
carbon prices, the economics of coal-fired generation
would change (Table 18). For a power supplier deciding
whether to continue operating an existing coal plant,
build a new coal plant, build a new natural-gas-fired
combined-cycle plant, or convert an existing coal-fired
plant to natural gas, continued operation of the coal
plant would be a clear winner in the absence of a carbon
price. As the carbon price rises, however, the new
natural gas plant looks more attractive. In the
hypothetical example, assuming a 70-percent capacity
factor for the four types of plant, it would make sense to
shut the coal plant down and build a new natural gas
plant at a carbon price of approximately $100 per metric
ton of carbon.55 Assuming a 30-percent capacity factor,
the crossover point would be closer to $200 per metric
ton of carbon. In this hypothetical example, the carbon
prices that would induce power suppliers to retire
existing coal plants are high, because the operating costs
of most existing coal plants are low. In reality, the
crossover point would vary from plant to plant.
Generating Capacity
In all the carbon reduction cases, significant amounts of
coal capacity are expected to be retired (Figure 73). In
general, the projected changes in the mix of generating
2000200520102015202005001,0001,5002,0002,500BillionKilowatthoursReference1990+24%1990+9%1990-3%
Figure 71. Projections of Coal-Fired Electricity
Generation, 2000-2020
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.
D080398B, and FD03BLW.D080398B.
capacity parallel the changes in fuel use. As the domestic
carbon reduction requirement becomes more stringent,
more coal capacity is retired and more natural gas and
renewable plants are built (Figure 74). In the 1990+24%
and 1990+9% cases, there is 3 percent and 10 percent less
coal-fired capacity by 2010, and 13 percent and 36 percent less by 2020. Approximately two-thirds of the existing coal-fired capacity is projected to be retired by 2020
in the 1990-3% case. The net result is that the share of
capacity accounted for by coal plants declines from
around 40 percent in 1996 to just over 29 percent in 2010
and to slightly over 11 percent in 2020 in the 1990-3%
case.
One possible effect of the projected coal plant retirements is that some of the plants may be shut down
before their total investment costs are recovered. Such
55In NEMS, the capacity factor for a particular plant type is determined by its operating costs. The values presented here are for illustration only.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Table 18. Hypothetical Examples of Levelized Plant Costs at Various Carbon Prices
(1996 Cents per Kilowatthour)
Plant Type
Carbon Price (1996 Dollars per Metric Ton)
0 50 100 150 200 250 300
70-Percent Capacity Factor
Existing Coal-Fired...................... 1.64 2.92 4.21 5.49 6.78 8.06 9.35
New Coal-Fired ........................ 3.67 4.91 6.16 7.40 8.65 9.89 11.14
New Gas-Fired Advanced Combined-Cycle . . 3.04 3.53 4.02 4.52 5.01 5.50 6.00
Coal-to-Gas Conversion.................. 3.45 4.19 4.94 5.68 6.42 7.16 7.91
30-Percent Capacity Factor
Existing Coal-Fired...................... 1.92 3.21 4.49 5.78 7.06 8.35 9.63
New Coal-Fired ........................ 6.69 7.93 9.18 10.42 11.67 12.91 14.16
New Gas-Fired Advanced Combined-Cycle . . 4.23 4.72 5.22 5.71 6.21 6.70 7.19
Coal-to-Gas Conversion.................. 3.90 4.64 5.38 6.12 6.87 7.61 8.35
Source: Energy Information Administration, Office of Integrated Analysis and Forecasting.
1981198319851987198919911993199505101520253035401995MillsperKilowatthourFuelNonfuelFigure 72. Operating Costs for Coal-Fired
Electricity Generation Plants, 1981-1995
Source: Form FERC-1, ÒAnnual Report of Major Electric Utilities, Licensees,
and Other.Ó
200020052010201520200100200300400GigawattsReference1990+24%1990+9%1990-3%
Figure 73. Projections of Coal-Fired Generating
Capacity, 2000-2020
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.
D080398B, and FD03BLW.D080398B.
unrecovered costs would be stranded. Most coal plants
are fairly old, however, and their construction costs have
already been recovered. On the other hand, some plant
owners could suffer losses because plants they expected
to be profitable might no longer be profitable when
carbon prices are imposed.
Natural Gas
Generation
The story for natural gas generation is the opposite of
that for coal (Figure 75). As the requirement to reduce
carbon emissions tightens and the associated carbon
price rises, natural-gas-fired generation becomes more
economical than coal-fired generation. In 2010 and
beyond, electricity generation from natural gas is
between 17 percent and 76 percent higher in the carbon
reduction cases than in the reference case projections.
Overall, between 1996 and 2020, natural gas generation
increases by almost 500 percent in the most stringent
carbon reduction cases, and even in the 1990+24% case it
CoalGasandOilHydroOtherRenewablesNuclear0100200300400500GigawattsReference1990+24%1990+9%1990-3%
Figure 74. Electricity Generation Capacity by Fuel,
2010
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.
D080398B, and FD03BLW.D080398B.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
2000200520102015202005001,0001,5002,0002,5003,000BillionKilowatthoursReference1990+24%1990+9%1990-3%
Figure 75. Projections of Natural-Gas-Fired
Electricity Generation, 2000-2020
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.
D080398B, and FD03BLW.D080398B.
19952000200520102015202005001,0001,5002,0002,5003,000BillionKilowatthoursFigure 76. Natural-Gas-Fired Electricity
Generation, 1990-3% Case, 1996-2020
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System run FD03BLW.D080398B.
is more than 30 percent higher than in the reference case
by 2020. Although it may be expensive to stop using
low-cost coal plants, replacing them with efficient
natural gas combined-cycle plants reduces carbon
emissions per kilowatthour of electricity generated by
nearly two-thirds.
The rate of increase in natural-gas-fired generation
varies over the 24-year projection period (Figure 76).
When carbon emission limits are first imposed in 2005,
there is rapid growth in natural gas generation, both
because the rising carbon price makes existing natural
gas plants more economical than existing coal plants
and because new natural gas plants are added quickly.
After the initial shift to natural gas, the growth in natural
gas generation continues, but at a slower rate. In the later
years of the projection, natural gas generation does not
increase as rapidly, because carbon-free renewable
technologies become economical as the demand for
electricity grows and natural gas prices increase.
In the carbon reduction cases, power plant use of natural
gas (excluding industrial cogeneration) is projected to
rise from roughly 3 trillion cubic feet in 1996 to between
8 and 12 trillion cubic feet in 2010 and between 12 and 15
trillion cubic feet in 2020. The projected increase in
demand for natural gas in the electricity sector
contributes to higher gas prices overall. As a result, only
small increases are projected for gas demand in other
sectors for the less stringent cases. In the more stringent
cases, gas demand in the other sectors (excluding
industrial) actually declines. For example, in the
1990+9% case, electricity sector gas use in 2010 is 57
percent higher than projected in the reference case, but
total gas consumption is only 10 percent higher (see
Chapter 5 for a discussion of natural gas supply).
Generating Capacity
There is only a little variation in the projections of total
natural-gas-fired generating capacity across the carbon
reduction cases. On the other hand, there are differences
in the types of natural gas plants projected to be built
(Figure 77). In the more stringent carbon reduction
cases, with higher carbon prices, the mix of natural gas
plants shifts from relatively inefficient simple natural
gas turbines and older steam plants to more efficient
combined-cycle facilities. The trend toward more efficient gas-fired technologies would be even stronger in
the 1990-3% case without the significant reduction in
electricity demand that is projected relative to the reference case (see below, Figure 84).
SteamCombinedCycleTurbine050100150200GigawattsReference1990+24%1990+9%1990-3%
Figure 77. Projections of Natural-Gas-Fired
Electricity Generation Capacity, 2010
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.
D080398B, and FD03BLW.D080398B.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
A critical question is whether new natural gas capacity
can be built in sufficient quantity and in the right places
to reduce carbon emissions to the levels required by the
Kyoto Protocol. For example, in the 1990-3% case, the
amount of capacity, mostly natural gas, projected to be
built in some years far exceeds the amount of capacity
built in any year since 1983. The average amount of
generating capacity brought on line each year since 1983
has been around 10 gigawatts (33 typical plants).56 The
peak year was 1985, when just under 22 gigawatts of
capacity was added. In the 1990-3% case, annual
additions are projected to exceed 28 gigawatts (93
typical plants) in some years.
Some gas-fired plants are expected to be built to meet
growth in demand, but most are likely to replace retiring
coal plants. From 2008 to 2020, the projected additions of
generating capacity in the 1990-3% case average 24 gigawatts annually, with just over 28 gigawatts in 2009. This
level of construction is high but not unprecedented. It is
actually less than the amount of capacity that was built
annually during the 1970s, when the demand for electricity was growing at more than twice the rate projected
in the reference case.
Given time and forewarning, the natural gas plant
design and construction industry should be able to meet
the challenge presented in the carbon reduction cases;
however, the prices for new gas-fired facilities might rise
above those used in this analysis. In addition, the situation could be exacerbated by the fact that many other
countries may also be turning to natural gas in order to
reduce their carbon emissions.
Not only will a large number of new natural gas plants
have to be built, they will also have to be built in the right
places. TodayÕs electricity transmission system is constructed around major load and supply centers, connecting major cities to major power plants. The location of
power plants is critical to the reliability of the electricity
supply system. If, as expected, a large number of existing coal plants are retired to reduce carbon emissions,
many of the new gas plants will have to be built at the
locations of the coal plants they replace, in order to
maintain the reliability of the system. (Biomass and
wind plants must be built where their resources are
available.) The alternative would be to reconfigure the
transmission system to accommodate new plant locations,57 an undertaking that might require additional
investment.
One option for adding new natural-gas-fired capacity
would be to modify existing coal-fired plants to burn
natural gas instead of coal. This option, however, may
not prove to be economical. Generally, there are two
approaches for converting a coal plant to burn gas. The
first is simply to modify the existing coal boiler so that it
can be fired with natural gas. From a mechanical perspective this is not terribly difficult or expensive. The
required plant modifications would be expected to cost
$70 to $80 per kilowatt of capacity, mainly for new burners and gas handling equipment (compressors, metering
station, distribution headers, etc.). In terms of performance, there would be a small loss of efficiency, 2 to 5 percent, if gas were burned in a boiler originally designed to
burn coal.58
The main problem with this approach to plant conversion is the relative thermal inefficiency of existing coal
plants. The majority of older coal plants consume
between 10,000 and 10,500 Btu of fuel for each kilowatthour of electricity they produce,59 as compared with
6,500 to 7,500 Btu of fuel input for each kilowatthour of
electricity produced by a new gas-fired combined-cycle
plant. Existing coal plants are economical because the
fuel is inexpensive, not because they are thermally efficient.
As described above (see Table 18), in the absence of
required carbon emissions reductions, existing coal-
fired plants are the most economical option for electricity generation. Conversion of existing plants from coal
to gas is not the most economical option if the plant is to
be used at a high capacity factor. If the price of carbon
emissions rises, however, continuing to run the existing
coal plant becomes less economical. Assuming a 70percent capacity factor and a carbon price of $100 per
metric ton, it would make sense to abandon the plant
(not the site) and build a new gas-fired combined-cycle
plant. At a lower capacity factor, the carbon price would
have to be higher before the operational cost savings
from the greater efficiency of a new combined-cycle
plant would offset its higher capital costs (Table 18).
The second approach to using gas in an existing coal
plant would be to ÒrepowerÓ it by converting it into a
natural gas combined-cycle plant. This approach would
result in higher plant efficiency, but it would also be
much more expensive than the first approach. In a typical repowering, the coal handling system and the boiler
are replaced with new combustion turbines and a heat
56Depending on the technology type, new power plants differ tremendously in size, from a few kilowatts for the smallest distributed
photovoltaic technologies to 500,000 kilowatts (500 megawatts) or more for the largest new coal and nuclear technologies. Throughout this
report, when a number of typical plants is provided, a 300-megawatt average plant size is used.
57See Energy Information Administration, ÒAn Exploration of Network Modeling: The Case of NEPOOL,Ó in Issues in Midterm Analysis
and Forecasting 1998, DOE/EIA-0607(98) (Washington, DC, July 1998), for a discussion of the impact of plant location on reliability and pricing.
58Cost and performance impact estimates provided by Parsons Engineering.
59Energy Information Administration, Form EIA-860, ÒAnnual Electric Generator Report.Ó
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
recovery boiler. The only significant part of the plant
that is maintained is the original turbine generator. This
approach can be attractive at some facilities, but it is not
without problems. New combined-cycle plants are
packaged systems. The turbines, heat recovery boiler,
and turbine generator are designed to work smoothly
together for optimal efficiency. Because many older
coal-fired plants were custom designed and built, they
do not always come in standard sizes or configurations
or with standard operational parameters. If such facilities are to be repowered, additional work will be
required to integrate the system components. Given that
for a typical combined-cycle plant the steam turbine
generator accounts for between 10 and 22 percent of the
capital cost of the plant,60 the additional work could easily drive the cost of repowering beyond what it would
cost simply to replace the plant with a new, more efficient packaged combined-cycle plant.
Renewable Fuels
In the carbon reduction cases, U.S. electricity suppliers
are expected to turn to renewable energy resources later
in the projection period to meet the demand for
electricity while reducing carbon emissions. Wind,
biomass, geothermal, solar, and hydropower resources
generally are thought to have less environmental impact
than fossil fuels; they are domestically available; and in
some instances they have begun to penetrate U.S.
electricity markets. Significant growth in the use of
nonhydroelectric renewable resources for electricity
generation is expected to accompany efforts to reduce
carbon emissions (Figure 78).
200020052010201520200100200300400500600BillionKilowatthoursReference1990+24%
1990+9%
1990-3%
1990-7%
Figure 78. Projections of Nonhydroelectric
Renewable Electricity Generation,
2000-2020
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.
D080398B, FD03BLW.D080398B, and FD07BLW.D080398B.
The largest increases in renewable generation are expected after 2010 in the most stringent carbon reduction
cases (Table 19). For this reason, the results of the 7percent-below-1990 (1990-7%) case are also discussed in
this section. Before 2010, nonhydroelectric renewable
technologies generally are not competitive with new
natural gas plants, but their costs are expected to fall
over time. With higher carbon prices, these technologies
can be expected to play a significant role in reducing carbon emissions. In the reference case little growth in generation from renewables is expected. In the carbon
reduction cases, nonhydroelectric renewable generation
is 1.1 to 1.7 times the reference case level in 2010 and 1.5
to 4.8 times the reference case level in 2020.
Because of the lack of market experience with renewable
technologies other than hydropower, there is considerable uncertainty about the costs of developing them on
the scale that would be needed for large carbon emission
reductions. It is also unclear whether electric system reliability can be maintained if large quantities of wind or
solar, which have intermittent output, are developed.
Although some environmental objections have been
raised against some renewables, including negative
effects on animal life, destruction of habitat, and damage
to scenery and recreation, these effects are small in comparison with the alternatives. While wind and biomass
technologies are expected to be the most important
renewable technologies used to reduce carbon emissions, othersÑincluding geothermal, conventional
hydroelectric, and solar power plantsÑmay also play a
role (Table 19).
Wind
Among the nonhydroelectric renewable fuels, biomass
and wind technologies are expected to make the most
significant contributions to carbon emission reductions.
Projected growth in the wind and biomass industries,
together with the natural gas industry, would at least
partially offset the impacts of declines in the coal industry. The biomass industry in the United States today is
small, but it could see large growth. Similarly, the wind
industry, estimated to employ 30,000 to 35,000 people
worldwide in 1995, could increase several times over in
the most stringent carbon reduction cases. In some
regions, wind is projected to provide a significant share
of electricity supply. However, the ability of wind
resources to meet large-scale U.S. electric power needs
reliably and cost-effectively is uncertain. Wind power is
an intermittent technology, available only part of the
time during a day or season. As a result, EIA assumes
that the maximum contribution of wind power will be
limited to 12 percent of any regionÕs total annual generation requirements (excluding cogeneration) to avoid
reliability problems that larger shares might cause.
60Electric Power Research Institute, Technical Assessment Guide. The steam turbine and auxiliary systems account for 10 percent of the
plant. If the boiler can also be used, this figure rises to 22 percent.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Table 19. Projected U.S. Electricity Generation From Renewable Fuels
(Billion Kilowatthours)
Projection
2000 2010 2020
Refer-
ence
Refer-
ence
1990
+24%
1990
+9%
1990
-3%
1990
-7%
Refer-
ence
1990
+24%
1990
+9%
1990
-3%
1990
-7%
Electricity Generators
Conventional Hydropower..... 310.3 313.0 313.0 313.0 317.4 321.9 313.2 313.1 313.1 317.7 322.4
Geothermal ................ 17.2 16.8 18.0 21.7 29.9 30.4 19.9 25.1 33.4 47.2 53.3
Municipal Solid Waste........ 22.8 27.0 27.0 26.8 26.5 26.5 29.8 29.8 29.7 29.8 29.9
Wood and Other Biomass ..... 8.2 8.7 17.6 21.0 34.7 36.4 8.7 22.5 83.1 244.4 305.1
Solar Thermal .............. 0.9 1.2 1.2 1.2 1.2 1.2 1.5 1.5 1.5 1.5 1.5
Solar Photovoltaic ........... 0.1 0.6 0.6 0.6 0.7 1.0 1.4 1.4 1.4 1.8 2.3
Wind ..................... 5.7 6.2 11.2 24.7 35.7 48.9 8.7 43.6 108.3 123.4 142.8
Subtotal ................. 365.2 373.5 388.6 409.0 446.1 466.2 383.2 437.0 570.5 765.9 857.2
Cogenerators
Municipal Solid Waste........ 2.2 2.3 2.3 2.3 2.3 2.3 2.3 2.3 2.3 2.3 2.3
Biomass................... 41.2 47.3 47.4 46.9 45.9 45.6 48.9 50.2 50.2 50.4 50.5
Subtotal ................. 43.5 49.6 49.7 49.2 48.2 47.9 51.2 52.5 52.5 52.7 52.8
Total Renewable Generation .. 408.7 423.1 438.3 458.2 494.3 514.1 434.4 489.5 623.1 818.5 910.0
Total Electricity Generation ... 3,716.8 4,267.6 4,144.0 3,929.7 3,712.6 3,641.7 4,648.2 4,422.3 4,282.7 4,160.2 4,105.1
Renewable
Share of Generation (Percent) . . 11.0 9.9 10.6 11.7 13.3 14.1 9.3 11.1 14.5 19.7 22.2
Nonhydroelectric Renewable
Share of Generation (Percent) . . 2.6 2.6 3.0 3.7 4.8 5.3 2.6 4.0 7.2 12.0 14.3
Source: Office of Integrated Analysis and Forecasting, National Energy Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.D080398B,
FD03BLW.D080398B, and FD07BLW.D080398B.
In the reference case, wind remains a minor contributor
to both total renewable energy and total electricity supply through 2020 (Table 19), accounting for just 2 percent
of generation from renewables and far less than 1 percent of total generation. In the carbon reduction cases, its
contribution grows. In the 1990+9% case, generation
from wind resources reaches 25 billion kilowatthours in
2010 and 108 billion kilowatthours in 2020, accounting
for nearly 17 percent of renewable generation and 2.5
percent of all U.S. electric power. In the 1990-3% and
1990-7% cases, with greater carbon reduction requirements, U.S. reliance on wind power is expected to be
higher, particularly after 2010. Generation from wind
power reaches 36 billion kilowatthours by 2010 in the
1990-3% case and increases even more thereafter, reaching 123 billion kilowatthours in 2020. In the 1990-7% case
it rises to 10 percent of renewable generation in 2010 and
16 percent (143 billion kilowatthours) in 2020, accounting for more than 3 percent of all electric power output.
In terms of generating capacity, wind accounts for more
than 11 percent of all renewables capacity in 2010 in the
1990-3% case and 26 percent of all renewables capacity
in 2020 in the 1990-7% case (Table 20). Again, however,
wind-powered capacity remains a relatively small share
of overall U.S. electricity generating capacity, in no case
exceeding 6 percent of the total. Wind power is already
entering some U.S. markets, and hundreds of megawatts
of new wind capacity is expected to enter U.S. service
before 2000. In the carbon reduction cases, wind power
expands rapidly (Figure 79). The projection for wind
capacity in 2005 in the 1990+9% case exceeds the
reference case projection for 2020, and in 2020 it is more
than 38 gigawatts. The wind capacity projections for
2020 are 44 gigawatts in the 1990-3% case and 51
gigawatts in the 1990-7% caseÑmore than 14 times the
reference case forecast.
The importance of wind power varies from region to
region. Whereas wind capacity today is concentrated in
200020052010201520200102030405060GigawattsReference1990+24%
1990+9%
1990-3%
1990-7%
Figure 79. Projections of Wind-Powered Electricity
Generation Capacity, 2000-2020
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.
D080398B, FD03BLW.D080398B, and FD07BLW.D080398B.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Table 20. Projected U.S. Electricity Generation Capacity From Renewable Fuels
(Gigawatts)
Projection
2000 2010 2020
Reference
Reference
1990
+24%
1990
+9%
1990
-3%
1990
-7%
Reference
1990
+24%
1990
+9%
1990
-3%
1990
-7%
Electricity Generators
Conventional Hydropower..... 79.39 79.78 79.78 79.80 80.74 81.84 79.78 79.79 79.80 80.78 81.92
Geothermal ................ 3.02 2.80 2.98 3.51 4.68 4.75 3.02 3.77 4.95 6.94 7.81
Municipal Solid Waste........ 3.40 4.02 4.01 3.99 3.95 3.95 4.42 4.42 4.41 4.43 4.44
Wood and Other Biomass ..... 1.64 1.76 1.80 2.70 4.93 5.32 1.76 2.74 11.95 35.27 43.99
Solar Thermal .............. 0.36 0.44 0.44 0.44 0.44 0.44 0.54 0.54 0.54 0.54 0.54
Solar Photovoltaic ........... 0.02 0.22 0.22 0.22 0.27 0.39 0.56 0.56 0.56 0.71 0.91
Wind ..................... 2.55 2.75 4.47 9.44 13.19 18.17 3.52 15.87 38.08 44.06 51.37
Subtotal ................. 90.39 91.77 93.71 100.10 108.20 114.85 93.60 107.68 140.29 172.72 190.97
Cogenerators
Municipal Solid Waste........ 0.44 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45
Biomass................... 6.08 6.70 6.68 6.60 6.48 6.44 6.84 6.96 6.93 6.93 6.94
Subtotal ................. 6.52 7.14 7.13 7.05 6.92 6.89 7.29 7.41 7.38 7.38 7.39
Total Renewable Capacity .... 97 99 101 107 115 122 101 115 148 180 198
Total Electricity Capacity ..... 803 916 895 921 945 944 1,008 972 965 958 949
Renewable
Share of Capacity (Percent) .... 12.07 10.80 11.26 11.64 12.19 12.90 10.01 11.84 15.30 18.79 20.91
Nonhydroelectric Renewable
Share of Capacity (Percent) .... 2.18 2.09 2.35 2.97 3.64 4.23 2.09 3.63 7.03 10.36 12.27
Source: Office of Integrated Analysis and Forecasting, National Energy Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.D080398B,
FD03BLW.D080398B, and FD07BLW.D080398B.
a few placesÑprincipally California, with smaller
amounts in Texas and MinnesotaÑin the carbon reduction cases, wind power development is expected to
occur in most regions west of the Mississippi River, as
well as in New England. Wind plants do not penetrate
heavily in most parts of the East and Southeast, where
resources are limited. For example, in the 1990-3% case,
more than 70 percent of all wind capacity in 2010 is
projected to be in the West, with three-quarters of the
remainder in the Upper Midwest. Still, wind power supplies only around 2 percent of generation in the Upper
Midwest, the Northwest and California and nearly 10
percent in the Southwest in 2010 in the 1990-3% case. On
the other hand, in the 1990-7% case, wind accounts for
significant shares of total generation in 2020 in some
regions.
Large-scale wind power development faces significant
uncertainties with regard to reliability, technology costs,
and resource development costs. Concerns about reliability center around the intermittent nature of wind. In
some areas, winds are highly predictable and coincident
with daily or seasonal electric power demands. By
nature, however, winds are rarely steady, are in various
degrees unpredictable (intermittent), and may occur at
times of low demand. As a result, wind power requires
the availability of other capacity to back it up. In addition, the variation in output from wind plants can stress
distribution and transmission lines as well as other generating equipment. The upper limit on the amount of
wind capacity that can be handled economically on a
given system is unknown. Various studies suggest a
very wide range of possibilities, but the highest value
achieved for a single hour in the United States is 8 percent.
In Europe, wind power development has grown rapidly
in recent years. In 1997, for example, Germany surpassed the United States in total wind capacity and
became the first nation to exceed 2,000 megawatts of
capacity. In Denmark, wind capacity exceeded 1,100
megawatts in 1997 and could approach 10 percent of the
nationÕs electricity generation by 2005 if planned expansion occurs. In Spain total wind capacity exceeded 450
megawatts at the end of 1997. In all three nations, additional wind capacity additions are planned over the next
5 years.
The rapid wind development in Europe is being encouraged by relatively high electricity prices and government subsidies. Under German law, wind power
producers are reportedly paid the equivalent of 9 to 10
cents per kilowatthour (90 percent of the residential
retail price). Prices paid to wind developers are reported
to be up to 9 cents per kilowatthour in Denmark and
about 8 cents per kilowatthour in Spain. Those prices are
much higher than U.S. wholesale electricity prices,
which typically are 2 to 4 cents per kilowatthour. Nevertheless, the European record suggests that power systems can support a larger share of wind than they have
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
in the United States to date and that, if prices are high
enough, capacity can be added fairly rapidly.61
A second issue is the considerable uncertainty surrounding the future cost of wind turbines. Installed capital costs for wind turbines and associated equipment
have fallen over the past 20 years and are expected to
continue falling, particularly if large numbers of turbines are built. The costs are near $1,000 per kilowatt of
wind capacity today, and they are projected to be below
$800 per kilowatt early in the 21st century and to
approach $600 per kilowatt by 2020 in the most stringent
carbon reduction cases. With no known manufacturing
barriers to large increases in factory production capacity
for wind turbines, the industry should be able to meet
the production levels called for in the carbon reduction
cases, given sufficient lead times. Of course, it is impossible to say with certainty that the projected cost declines
will occur. This analysis does adjust for the cost effects of
short-term bottlenecks in identifying sites, permitting
projects, manufacturing equipment, and installing projects, but the actual effects of rapid large-scale expansion
are not known.
While there appear to be large wind resources in many
regions, the costs of developing some of the sites may be
high. In general, wind power costs are expected to
increase as the best natural resources are consumed and
less-favored sites enter service. Lower quality sitesÑ
including those on steep, rocky, or sharply varied
surfaces, those in more difficult environments (excessive
cold, moisture, dirt, insects, or storms), and those with
less useful winds (unpredictable, ill-timed, sharply
varying, too fast)Ñcould have much higher costs than
more favorable sites. Moreover, in most regions only a
portion of the total potential is likely to be economical.
The possible stress on wind resources (and therefore
costs) can be seen by comparing projections of wind
capacity with EIAÕs estimates of ÒeconomicÓ resourcesÑidentified as those available at capital costs no
more than double the baseline projection (Figure 80). In
the 1990-7% case, eight regions consume a third or more
of ÒeconomicÓ wind resources, and three regions exceed
that portion of supply, including California. In those
regions, more expensive wind resources are developed
in the most stringent carbon reduction cases. Little is
known about the actual costs at these levels of resource
use.
The costs of transmission interconnections and of
upgrading existing distribution and transmission
networks are also expected to increase as the penetration
of wind resources grows. As projects are developed at
greater distances from existing lines, the costs of new
ECAR1ERCOT2MAAC3MAIN4MAPP5NY6NE7FL8SERC9SPP10NWP11RA12CNV13020406080100120140160PercentFigure 80. Projected Shares of Most Economical
Wind Resources Developed by Region,
1990-7% Case, 1996-2020
Note: ECAR = East Central Area Reliability Coordination Agreement
Region; ERCOT = Electric Reliability Council of Texas; MAAC = Mid-
Atlantic Area Council; MAIN = Mid-America Interconnected Network;
MAPP = Mid-Continent Area Power Pool; NY = New York Power Pool;
NE = New England Power Pool; FL = Florida subregion of the South-
eastern Electric Reliability Council; STV = Southeastern Electric Reli-
ability Council excluding Florida; SPP = Southwest Power Pool; NWP =
Northwest Pool subregion of the Western Systems Coordinating
Council; RA = Rocky Mountain and Arizona-New Mexico Power Areas;
CNV = California-Southern Nevada Power Area.
Source: Office of Integrated Analysis and Forecasting, National Energy Mod-
eling System run FD07BLW.D080398B.
interconnections will increase. In addition, the costs of
upgrading existing local distribution networks, both to
transmit the electricity generated from wind power and
to offset the destabilizing local effects of varying power
flows, will increase.
Finally, market competition for land with good wind
resources is also likely to increase the future costs of
extensive wind power development. Other urban or
agricultural uses may compete for some locations. Public opposition to wind project development on environmental, cultural, and recreational grounds may also
grow as large numbers of wind facilities are built.
Because excellent wind resources tend to occur in highly
visible places, such as along ridges and other natural
projections, preferred sites often serve other cultural,
scenic, or religious purposes, and they may not be made
available for wind power development. For example, it
remains to be seen whether the development of 170
square miles in Texas (about 0.1 percent of the land area)
for the wind capacity that would be needed to meet the
2020 projections in the 1990-7% case would be acceptable to the StateÕs inhabitants.
61American Wind Energy Association, International Wind Energy Capacity Projections (Washington, DC, April 1998).
62Only 6 percent of the estimated wind resources in region 5 (including Minnesota, Iowa, and the Dakotas) are used in the 1990-7% case;
however, the remaining resources are not economically accessible to other regions.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Biomass
Unlike wind plants, which are intermittent, biomass
plants operate continuously. Biomass currently is being
used to supply energy for power generation and in the
industrial, transportation, and residential sectors. The
largest amount of biomass is used in the paper and lumber industries, where residue is burned to produce both
electricity and steam (cogeneration). Biomass is also
used to produce ethanol for fuel in the transportation
sector, and wood is burned for residential heating.
Current biomass consumption in the electricity sector,
excluding cogeneration, is limited to a few inefficient
wood-burning generating units and a small amount of
cofiring at coal plants. Newer technologies, primarily
several types of gasification combined-cycle units, are in
the demonstration phase in the United States and are
expected to be commercially available by 2005. Such
units would be somewhat more expensive than current
technology, but they are expected to be more than twice
as efficient. They can use a variety of fuel sources, such
as wood and wood residues, several types of energy
crops, and crop residues. Without a carbon price, these
facilities currently are not competitive with new natural
gas or coal plants. However, using biomass in the production of electricity produces no net carbon emissions.
The carbon emitted during biomass combustion
approximates the carbon sequestered during the growth
of the trees or crops that are burned. As a result, it is an
attractive option for complying with the Kyoto Protocol.
In the 1990+24% case, biomass generation increases only
slightly from the levels projected in the reference case. In
the 1990+9% case, however, biomass generation is projected to reach 68 billion kilowatthoursÑ21 percent
above the reference case projectionÑin 2010 and 133 billion kilowatthoursÑmore than double the reference
case projectionÑin 2020. In the 1990-3% case, biomass
generation is projected to be 81 billion kilowatthours in
2010Ñ44 percent above the reference caseÑand 295 billion kilowatthoursÑ5.0 times the reference caseÑin
2020. And in the 1990-7% case, biomass generation
exceeds the reference case projection by about 47 percent
in 2010 and by 6.2 times in 2020. In each of these cases,
biomass is allowed to contribute up to 5 percent of a coal
plantÕs fuel input, but because coal plant usage declines
rapidly as the carbon price increases, the contribution
from cofiring is limited.
With biomass resources projected to play such a major
role in meeting electricity needs in the carbon reduction
cases, a critical question is whether the projected levels
of reliance on biomass would be realistic. To answer that
question, it is necessary to examine the components of
the biomass resource. Biomass resources are diverse and
potentially much larger than the amounts projected to
be developed even in the most stringent carbon
reduction cases in this analysis (Figure 81).
ECARERCOTMAACMAINMAPPNYNESERC/FLSPPNWPRACNV0102030405060GigawattsReference1990+24%1990+9%1990-3%
ResourceFigure 81. Estimated Biomass Resource
Availability and Projected Generating
Capacity in 2020 by Region
Note: ECAR = East Central Area Reliability Coordination Agreement
Region; ERCOT = Electric Reliability Council of Texas; MAAC = Mid-
Atlantic Area Council; MAIN = Mid-America Interconnected Network;
MAPP = Mid-Continent Area Power Pool; NY = New York Power Pool;
NE = New England Power Pool; FL = Florida subregion of the South-
eastern Electric Reliability Council; STV = Southeastern Electric Reli-
ability Council excluding Florida; SPP = Southwest Power Pool; NWP =
Northwest Pool subregion of the Western Systems Coordinating
Council; RA = Rocky Mountain and Arizona-New Mexico Power Areas;
CNV = California-Southern Nevada Power Area.
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.
D080398B, FD03BLW.D080398B, and FD07BLW.D080398B.
Biomass materials are derived from a variety of sources,
including urban wood waste, mill residues, forest residue, agricultural residue, and energy crops grown specifically for combustion. Urban wood waste includes
tree trimmings, construction and demolition debris, and
discards such as crates and pallets. (Some of these materials are currently being used to make recycled products
or as fuel, and the resource data used for this analysis
exclude those quantities.) Mill residues are the sawdust
and scrap from sawmills, pulp mills, and wood product
facilities. Many mill residues are consumed on site, but
some are accumulated in stockpiles or sent to landfills,
often at a cost to the producer. Forest residues are, generally, material that is too low grade to be used for other
products. They include branches, dead trees, unmarketable species, and cull trees from commercial forests. The
alternative to its use as a fuel is to leave it in the forest.
Agricultural residues include a wide variety of materials. The greatest quantities (and the only amounts
included in this analysis) are from wheat straw and
cornstalks. Only a small amount is currently used as
fuel, most being left in the field. It is assumed here that
only 40 percent of all agricultural residues would be
available for use as fuel, with the rest continuing to be
left in the field. What the above types of residues have in
common is that they are very inexpensive at the source.
On the other hand, the cost of gathering and delivering
them to a power plant, compared with the cost of coal,
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
usually makes them too expensive for use in electricity
generation under current economic conditions.
Energy crops involve dedicated operations that would
likely require long-term agreements between growers
and conversion plant operators. The primary energy
crops are willow, poplar, and switchgrass, each with distinct growing areas and conditions. Energy crops differ
from residues in that it is the cost of growing them, not
collection, that dominates their total costs.
Agricultural lands can be divided into croplands, pasturelands, and Conservation Reserve Program (CRP)
acreage. The total U.S. agricultural land supply is
approximately 960 million acres, of which about one-
third is now used for field crops. In some instances,
energy crops can be grown on poor quality land that has
no other use. The amounts of agricultural land assumed
to be available for energy crops in the resource data used
for this analysis include all the CRP acreage, 20 percent
of the cropland, and 10 percent of the pastureland. However, even in the cases that project the highest levels of
biomass use, the total amount of land needed for energy
crops would be about 10 to 12 million acres, which is in
the range of the yearly fluctuations of U.S. cropland
planted. Thus, the question of competition for land does
not appear large. As fossil fuel prices rise in the more
stringent carbon reduction cases, the value of biomass
fuels would also rise, making energy crops more attractive economically.
There may be competition between the use of land for
biomass energy crops and its use for tree planting to
increase carbon sequestration. In terms of the amount of
carbon sequestered or emissions avoided per acre of
land used, displacing a new gas-fired plant with a
biomass-fired plant would have about the same impact
as planting trees. For example, the U.S. Environmental
Protection Agency estimates that planting 1 acre of trees
on marginal land would sequester 0.6 to 1.6 metric tons
of carbon annually.63 In comparison, if a new biomass
power plant displaced a new gas-fired plant, an estimated 1.3 metric tons of carbon emissions would be
avoided per acre of land used.64 The comparison would
not be as close if the generation displaced were from a
coal-fired power plant, which would emit roughly 3
metric tons of carbon in producing the same amount of
electricity that a biomass plant would generate from 1
acre of crops. The critical issue in the land use decision
between tree planting and energy crops will be the relative economics of the two choices. If sequestration
proves to be more economical, fewer biomass plants
may be built than projected in this analysis. Instead of
building a biomass plant, a developer could simply
build a gas-fired plant and also grow enough trees to offset the carbon emissions from the plant.
It is assumed in this analysis that energy crops will not
become economical until new integrated gasification
combined-cycle (IGCC) plants are available in 2005 and
after. The current technology for biomass plants, using
stoker boilers, is inefficient and uneconomical. The
newer IGCC technology is now being tested, and it is
expected to be vastly superior to the current technology
in terms of both efficiency and emissions. Most of the
experience with the IGCC technology has been in
Europe, particularly in Scandinavia. Sydkraft, the
second-largest utility in Sweden, has been operating a 6megawatt wood-fired IGCC plant in Varnami, Sweden,
since 1994. Finland has a 30-megawatt unit operating on
wood waste, as well as several smaller peat-fired gasification units with a combined capacity of 50 megawatts.
There are several other demonstration plants that total
about 5 megawatts of capacity worldwide. Future plans
include 12 megawatts of capacity in Italy (Bioelettrica), 8
megawatts in the United Kingdom, and 32 megawatts in
Brazil. In addition, a number of refineries are currently
operating IGCC plants that burn petroleum coke.
In the United States, the most advanced IGCC project is
operated by the Vermont Department of Public Works
in cooperation with utilities in the State, the U.S. Department of Energy, the U.S. Environmental Protection
Agency, and the U.S. Agency for International Development. The system, which gasifies waste wood and wood
chips from a dedicated poplar tree farm, is just beginning operation, with a design capacity of 15 megawatts.
The project is being used to demonstrate the economics
of the technology. In addition, a privately owned 7.5megawatt unit fueled with various wood, paper, and
industrial wastes began operating in the Midwest in
June 1998, and a 75-megawatt alfalfa-fired unit is
planned for operation in 2001 in Minnesota.
As shown in Table 21, the potential resource base for biomass from all sources amounts to approximately 15
quadrillion Btu annually, roughly enough to meet 15
percent of todayÕs U.S. energy needs if fully developed.
Even in the most stringent carbon reduction case, however, only about 15 percent of the resource, about 2.3
quadrillion Btu, is projected to be used. The region that
shows the greatest projected growth in biomass consumption is the Southeast, followed by the Midwest.
The Southeast has ample supplies of both forests and
cropland. In the Midwest, the land suitable for energy
crops is vast, although energy demand there is low. The
63U.S. Environmental Protection Agency, Climate Change Mitigation Strategies in the Forest and Agriculture Sectors (Washington, DC, June
1995), p. ES-5.
64This estimate was derived from the following assumptions: biomass yield 6 tons per acre, biomass heat content 17,000,000 Btu per ton,
biomass plant heat rate 8,000 Btu per kilowatthour, gas plant heat rate 7,000 Btu per kilowatthour, and natural gas carbon content 14,400
metric tons per trillion Btu.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
region that comes closest to reaching a limit on available
resources is Florida, which has high electricity demand
and limited biomass resources. The West is the area that
uses biomass the least, because land suitable for energy
crops is limited, and other resources, including other
renewables, are more plentiful.
Table 21. U.S. Biomass Resources
Biomass Resource
Quantity Available
in 2020
(Quadrillion Btu)
Price Range
(1996 Dollars
per Million Btu)
UrbanWoodWaste... 0.2 0-3
MillResidues........ 0.8 1-4
ForestResidues...... 6.5 3-4
CropResidues....... 0.9 2-3
EnergyCrops........ 6.5 1-3
Total ............. 15.0 Ñ
Source: Urban Wood Waste and Mill Residues: Antares Group, Inc. Forest
and Crop Residues: Oak Ridge National Laboratory. Energy Crops: Oak Ridge
Energy Crop County Level Database (December 20, 1996).
Biomass Limitation. Because of concerns about the
ability of the biomass energy business to develop as
rapidly as would be required to meet the capacity and
generation projections in the most stringent carbon
reduction cases in this analysis, a special sensitivity case
was analyzed, assuming that no new biomass capacity
would be built. All other assumptions were same as
those in the 1990-7% carbon reduction case. In the
sensitivity case, the projected carbon price was
approximately $39 per metric ton higher in 2020 than in
the 1990-7% case, with smaller increments in 2010 and
2015.
Without additional biomass capacity, new natural gas
capacity for electricity generation was projected to be
about 43 gigawatts higher than in the 1990-7% case in
2020, making up 212 billion kilowatthours of the 295 billion kilowatthours of generation ÒlostÓ from biomass.
Most of the remaining decrement was balanced out by
lower demand resulting from higher projected electricity prices that stemmed from the higher carbon price.
Natural gas prices at the wellhead were also projected to
be higher in the biomass limitation sensitivity case, by
about $0.13 per thousand cubic feet in 2020 as compared
with the projected price in the 1990-7% case.
Geothermal
Although it is a more limited resource than biomass or
wind, geothermal energy has the potential to contribute
to the goal of carbon emission reductions. Only hydrothermal resources west of the Rocky Mountains are considered in this analysis. The technologies represented
for new generating capacity are dual-flash and binary
cycle plants, both of which are currently available. The
existing dry-steam capacity at The Geysers is expected
to decline as the resource continues to be depleted.
Although few domestic orders for new geothermal
plants are being placed, the U.S. geothermal industry
remains viable because of activity with foreign projects,
such as those in Indonesia and the Philippines. Under
the Kyoto Protocol, the large U.S. resources, which are
costly to develop because of their inaccessibility, could
be brought within economic reach. Although little new
capacity has been built in the United States in recent
years, studies have estimated that more than 27 gigawatts of new capacity could be developed from currently identified resources and as much as 50 gigawatts
when potential unidentified resources are included.65
In the reference case, geothermal electricity generation is
projected to be 17 billion kilowatthours in 2010 and 20
billion kilowatthours in 2020. In the 1990+9% case, geothermal generation is projected to increase to 22 and 33
billion kilowatthours in those years, levels that are 29
percent and 68 percent, respectively, above the reference
case projection. In the 1990-3% case, geothermal generation increases to 30 billion kilowatthours in 2010 and 47
billion kilowatthours in 2020. In the reference case, 280
megawatts of new capacity is added by 2010, more than
80 percent of which is built in the Northwest and the
remainder in California. In the 1990-3% case, roughly 60
percent of the projected new capacity is built in the
Northwest, 35 percent in California, and the remainder
in the Southwest. These levels are within estimates of the
potential for geothermal development by the California
Energy Commission (CEC) and the Northwest Power
Planning Council (NPPC). The CEC found more than 3
gigawatts of potential66 and the NPCC nearly 4 gigawatts of potential in an optimistic case.67
Municipal Solid Waste
Electricity generation from municipal solid waste facilities is not expected to increase beyond the reference case
levels of 29 billion kilowatthours in 2010 and 32 billion
kilowatthours in 2020, regardless of the carbon reduction target assumed. The economics of these facilities are
driven primarily by waste disposal costs (landfill tipping fees), rather than their energy production. After rising in the 1980s, tipping fees have stabilized, and they
are not expected to increase significantly. Moreover,
efforts to reduce carbon emissions could actually reduce
the waste stream available for combustion because of
greater emphasis on reusable products, reduced use of
packaging materials, and recycling. In addition to their
high cost, municipal solid waste facilities are expected to
65Energy Information Administration, Geothermal Energy in the Western United States and Hawaii: Resources and Projected Electricity Genera
tion Supplies, DOE/EIA-0544 (Washington, DC, September 1991).
66California Energy Commission, Technical Potential of Alternative Technologies (December 2, 1991).
67Northwest Power Planning Council, Northwest Power in Transition: Opportunities and Risk, 96-5 (March 13, 1996).
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
be at a disadvantage in the electricity generation market
because of the carbon emissions produced from the
petroleum-based portion of the waste stream (primarily
plastics), local resistance to their operation, and other
environmental factors.
Solar
A variety of photovoltaic (PV) configurations serve U.S.
electricity markets. Grid-connected PV can be (1) large
central station units greater than 1 megawatt, (2) smaller
distribution-level units less than 1 megawatt, and (3)
individual end-user units, usually much less than 20
kilowatts. Off-grid PV always serves individual end
usesÑfor remote buildings, pumps, signals and communications devices and for lightingÑwhere the costs
of grid interconnection are high. EIA forecasts include
only grid-connected power.
PV is expected to grow steadily over the forecast period,
as experience grows and costs decline. In general,
increases in electricity prices should imply increasing
opportunities for PV technologies. In the reference case,
an increase in U.S. grid-connected PV is projected, from
just over 10 megawatts in 1996 to 560 megawatts in 2020.
No change from reference case levels is expected in the
1990+24% case. In the 1990-3% and 1990-7% cases, grid-
connected PV capacity increases more rapidly, exceeding 700 and 900 megawatts by 2020, respectively.68
Off-grid PV applications, currently estimated to grow by
less than 10 megawatts a year, should expand much
more quickly if electricity prices rise, particularly if individual consumers shoulder the full costs of interconnection in locations that are difficult to serve. Furthermore,
as costs decline, experience grows, and world demand
increases, global markets for U.S. PV outputÑalready
absorbing nearly two-thirds of U.S. productionÑshould
also enjoy robust expansion. As a result, U.S. production
of PV is likely to expand even more rapidly than domestic PV consumption.
Despite the optimistic outlook for PV in cases indicating
increasing electricity pricesÑand despite expected large
drops in PV costsÑthe technology is not expected to
become a large component of U.S. electricity supply
through 2020. In most instances, central station fossil,
nuclear, and other renewable sources will remain far
less costly than PV over the forecast period.
Even in the 1990-7% case, central station PV is expected
to remain more expensive than alternatives through
2020 in all regions. In the most favorable areas, such as
the Southwest, where central station PV costs are projected to decline to around 9 cents per kilowatthour after
2012, electricity generation costs for natural-gas-fired
advanced combined-cycle plants are expected to be
much lower, around 6 cents per kilowatthour including
the carbon price, and to provide power more reliably
and for a much greater proportion of the demand cycle.
As a result, no new central station PV capacity is
expected to be built on a cost-competitive basis.
Distributed PV units less than 1 megawatt are likely to
succeed in small numbers in limited circumstances, and
they are included, along with small end-user units, in
EIA forecasts for grid-connected PV growth. Distributed
PV may become competitive where the combination of
excellent insolation, transmission or distribution line
congestion, and unavailability of natural-gas-fired
capacity make PV a cost-effective option. Such combinations, however, are expected to be infrequent.
As costs drop and experience grows, end-user sited PV
may grow more rapidly, but it is not expected to become
a general source of end-user electricity supply. More
individual instances should occur in which delivered
peak power can be cost-effectively supplied by grid-
connected PV, such as where peak-time distribution line
congestion and difficulty in siting new lines raise the
costs of power from central station plants. Overall, however, PV is expected to remain costly for almost all applications that could use grid-connected power.
Smaller-scale PV units purchased by retail consumers
are likely to cost even more than utility-scale PV. Moreover, grid-using PV consumers could incur some fixed
costs of the transmission and distribution system to
which they remain connected. And to the extent that
utilities incur additional costs from the presence of end-
user PVÑsuch as for protecting lines and personnel
from intermittent and unexpected electricity flowsÑ
users could incur additional costs. As a result, utilities
may be unwilling to pay full retail rates for electricity
purchases from end-user PV units.
Unlike PV, which uses solar energy to create electricity
directly, solar thermal technologiesÑincluding trough,
central receiver, and dish StirlingÑconvert solar energy
to heat and then to electricity in generating units (usually turbines). The 360 megawatts of trough units built in
California in the 1980s constitute almost all the solar
thermal units operating today. No additional trough
units are planned at this time. One central receiver unit,
the 10-megawatt Solar II, is currently being tested. No
commercial-scale central receiver units are in operation
or planned. Dish-Stirling units are in relatively early
testing stages, with only a few kilowatts operating.
Unless breakthroughs are forthcoming, solar thermal
appears unlikely to make a notable contribution to U.S.
electricity supply, even in the most demanding carbon
68Increasesin PV capacity were determined exogenously to reflect small, distributed, and end-userapplications. Central-station PV was
allowed to compete with other central station generating technologies.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
reduction cases. Solar thermal suffers a number of disadvantages. Cloud cover and humidity weaken the required (direct) solar radiation sufficiently to eliminate
all but the drier Western regions from consideration,
and where solar conditions are best the water volumes
needed for steam production are in shortest supply.
In addition, the technology currently has both high capital costs and limited availability. The facilities cannot
operate many hours without storage, but adding energy
storage fields to compensate for non-peak solar hours
means significant additional capital costs. As a result,
central station solar thermal generation is not expected
to penetrate U.S. markets significantly before 2020.
Hydropower
Under currently expected circumstances, little additional hydroelectric power is likely to be available to
meet U.S. carbon emission reduction targets. Conventional hydroelectricity is the major source of renewable
electricity today, supplying about 80 percent of renewable generation and nearly 10 percent of all U.S. electric
power in 1996. However, the combination of few additional sites, high capital costs, reduced Federal support,
and changing national water-use priorities away from
electricity and toward environmental improvementsÑincluding for fish, habitat preservation, and recreationÑsharply limit the potential for expansion of
U.S. hydropower capacity, whether or not carbon reduction measures are required.
In the reference case, U.S. conventional hydroelectric
power stays virtually unchanged over the forecast period, annually providing about 313 billion kilowatthours.
Because both overall electricity generation and use of
other renewables increases, the hydropower shares of
both renewable and total generation decline. In 2020,
conventional hydropower is projected to provide about
72 percent of U.S. renewable electricity generation and
less than 7 percent of total generation.
Increasing carbon reduction requirements are projected
to increase reliance on other renewables but have little
effect on hydropower. In the 1990+9% case, total renewable generation in 2020 is nearly 44 percent greater than
in the reference case, but hydroelectricity remains
unchanged. Despite much greater reliance on renewables in the 1990-3% and 1990-7% cases, U.S. conventional hydroelectric power increases only slightly. Even
in the 1990-7% case, hydroelectric generation in 2020 is
less than 3 percent above the reference case projection.
The increases that are projected in this case are primarily
from new units at existing dams rather than the addition
of new dams. As a consequence, by 2020, conventional
hydroelectric generation slips to second place, below
biomass, providing about 35 percent of total renewable
electricity generation.
Nuclear
Nuclear generation is expected to be higher in the carbon reduction cases than in the reference case. In the reference case, more than half of the nuclear plants existing
today are expected to be retired when their licenses
expire. The economics of the retirement versus life
extension decision will change, however, if significant
reductions in carbon emissions are required.
To simulate this decision process, an approach was
developed for evaluating the economic choice of continuing to operate a nuclear plant or retiring it and building a replacement plant. Essentially it was assumed that
as nuclear plants age their components will eventually
need to be replaced. At that point, the component replacement costs and the plantÕs continuing operating
costs can be compared to the costs of building and operating another type of generator. Because it is impossible
to predict when component replacement costs will be
incurred for a particular plant, it was assumed for the
sake of simplicity that all nuclear plants would need
refurbishment at 30 years and again at 40 years of life.
The 30-year point represents the point at which many
existing plants are expected to require turbine generator
replacements, and the 40-year point represents the point
at which plants will have to be prepared for continued
operation after their 40-year operating licenses expire.
Even in the 1990+24% case, where the projected carbon
price is much less than that in the 1990-3% case, it would
be economical to incur the 30-year component
replacement cost and continue operating most nuclear
plants. For some plants, however, it would not be
economical to continue operation after 40 years. With
the higher carbon prices in the 1990-3% case, almost all
existing nuclear plants would be maintained and
continue their current electricity generation levels
throughout the projection period (Figure 82). The
difference in electricity generation projections between
the reference and 1990-3% cases is greater for nuclear
than for any other non-carbon-based fuel (see Figure 70).
In the absence of that increment in nuclear generation,
greater reliance on natural gas and nonhydroelectric
renewables would result in even higher generating
costs.
In the 1990-3% case, additional generation from nuclear
plants operating beyond 40 years offsets approximately
30 to 40 million metric tons of carbon emissionsÑ
approximately equal to the difference between the
carbon targets in the 1990-3% and 1990-7% cases. Thus,
in the absence of the projected nuclear plant life
extensions, projected electricity prices in 2010 in the
1990-3% case would be some 5 percent higher,
equivalent to the 2010 price projection in the 1990-7%
case.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
2000200520102015202002004006008001,000BillionKilowatthoursReference1990+24%1990+9%1990-3%
Figure 82. Projections of Nuclear Electricity
Generation, 2000-2020
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.
D080398B, and FD03BLW.D080398B.
20002005201020152020020406080100120GigawattsReference1990+24%1990+9%1990-3%
Figure 83. Projections of Nuclear Electricity
Generation Capacity, 2000-2020
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.
D080398B, and FD03BLW.D080398B.
The higher projections for nuclear electricity generation
in the carbon reduction cases would have implications
for nuclear waste disposal. The projected impact is not
significant through 2010, but in 2020 cumulative spent
fuel discharges from nuclear units would be 6 percent
and 9 percent higher than the reference case projection
in the 1990+9% and 1990-3% cases, respectively. The
spent fuel calculations assume that all spent fuel will be
removed from a reactor when it is retiredÑa greater
amount than would be discharged during a normal year
of operation. Thus, even greater differences would be
seen if spent fuel projections were calculated over the
entire lifetime of all nuclear units.
Nuclear capacity varies significantly across the carbon
reduction cases (Figure 83) not because new nuclear
plants are built but because existing plants are
maintained and life-extended. In the 1990+9% case, the
carbon price makes it economical to maintain almost 75
percent of existing U.S. nuclear power capacity
throughout the projection period, so that the projected
capacity in 2020 is 26 gigawatts higher than in the
reference case. With higher carbon prices in the 1990-3%
case, it would be economical to keep 86 percent or more
of the existing nuclear capacityÑroughly 40 gigawatts
more than in the reference caseÑoperating through
2020.
Demand Reduction
Electricity usage decisions by consumers, as discussed
in Chapter 3, would also play a large role in reducing
electricity sector carbon emissions (Figure 84). Even in
the 1990+24% case, consumers would be expected to
reduce their electricity consumption by 4 percent in 2010
and 6 percent in 2020 relative to the levels of
consumption projected in the reference case. When a
200020052010201520200-200-400-600-800BillionKilowatthours1990+24%1990+9%1990-3%
Figure 84. Projected Changes in Electricity Sales
Relative to the Reference Case,
2000-2020
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.
D080398B, and FD03BLW.D080398B.
more stringent carbon reduction target is assumed in the
1990-3% case, consumer usage decisions are more
important. In this case, lower demand for electricity
accounts for a large share of the reduction in electricity
sector carbon emissions.
Electricity Prices
While electricity suppliers do have options available for
reducing their carbon emissions, it will take financial
incentives to encourage them to implement them. In
turn, this will have an impact on average electricity
prices. In all the cases discussed in this analysis, with the
exception of the competitive pricing cases described
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
below, electricity prices are based on average costs in all
regions except California, New York, and New England.
It is assumed that competitive prices, based on marginal
costs, will be phased in over time in those three
regions.69 In other words, the total costs of producing
and delivering electricity to consumers are divided by
the amount of electricity sold to calculate the average
prices. In the carbon reduction cases, electricity production costs include the projected carbon prices. A discussion of competitive electricity markets is provided
below.
In all the carbon reduction cases, projected electricity
prices are higher than reference case prices beginning in
2005 as the carbon targets are phased in (Figure 85). The
highest prices are projected between 2008 to 2012. In
subsequent years, as new renewable plants become
more economical and the financial incentives needed to
ensure their development moderate, electricity prices
are expected to decline. In 2009, average electricity
prices in the 1990-3% case could be as much as 82 percent
higher than in the reference case. The higher prices
would lead to higher consumer bills. In 2010, residential
consumers would pay $10, $23, and $36 more per month
on average in the 1990+24%, 1990+9%, and 1990-3%
cases, respectively, than the $70 average monthly bills
projected in the reference case.
Regionally, the price impact would be greatest in those
regions where generation currently is dominated by
coal-fired power plants. Particularly hard hit would be
the midwestern ECAR and MAPP regions, where coal-
fired generation accounts for 89 and 70 percent of total
1995200020052010201520200246810121996CentsperKilowatthour1990+9%
Reference1990+24%
1990-3%
Figure 85. Projections of Electricity Prices,
1996-2020
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.
D080398B, and FD03BLW.D080398B.
generation, respectively. In the 1990+9% case, efforts to
reduce carbon emissions could lead to an increase of as
much as 71 to 78 percent in the price of electricity in the
two regions between 2008 and 2010 relative to the prices
projected in the reference case. Nationally, prices in the
1990+9% case in 2008 are only 50 percent higher than in
the reference case.
The impact on prices could be greater in a more competitive market. The results shown in Figure 85 are based on
prices calculated as they have been in the regulated electricity market over the past 50 to 60 years.70 This may not
be appropriate in the near future. The U.S. electric industry is in the midst of a major change in its regulatory
pricing structure. Historically, prices have been based
on the average cost of producing and delivering electricity to the customer, but in a competitive market this will
not be the case.
In a competitive market, prices will be based on the
operating costs of the last plant needed to meet demand.
On a typical hot summer day, generating plants are
brought on line as the demand for electricity grows. Initially, the lowest cost plants (in terms of operating costs)
are brought on line, but as consumer needs grow, more
costly units are started. At any given time, the price for
power will equal the cost of operating the highest cost
unit supplying powerÑthe Òmarginal unit.Ó The operating costs for a typical plant include fuel and operations
and maintenance costs and, in a carbon reduction case,
the carbon price. Because carbon prices would be
included in the operating costs of the marginal plant,
they would have a direct impact on the competitive
price of electricity. In a regulatory pricing environment
the effect of the carbon price would be smaller, because
the operating costs for plants with lower carbon emissions would be averaged in with the costs for units with
higher emissions.
In this analysis, when higher carbon prices are projected,
end-use electricity prices are higher under marginal cost
(competitive) pricing than they would be under average
cost (regulated) pricing (Figure 86). The effect of
marginal cost pricing on electricity prices increases with
the level of the carbon price. Because the effect is
relatively minor in the less stringent carbon reduction
cases, the 1990-3% case is examined. In this illustration,
the higher prices in the early years under marginal cost
pricing cause consumers to reduce their electricity use,
resulting in lower generation requirements. Consequently, it is easier for suppliers to meet the carbon
reduction goals, and the carbon price is lower than it
would be under average cost pricing (Figure 87),
although the competitive electricity price remains
higher than the average electricity price.
69See Energy InformationAdministration, Annual Energy Outlook 1998, DOE/EIA-0383(98) (Washington, DC, December 1997), for a discussion of competitive pricing.
70In all cases the California, New York, and New England regions are treated as competitive.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
An easy way to see the impact of the carbon price is to
look at the impact it has on the types of plants that will
set the marginal price of power. A carbon permit system
would change the plants that set the market price of
electricity in a competitive pricing environment. In a
carbon reduction case assuming competitive pricing, the
order in which plants are used would differ from that in
a corresponding reference case. The coal-fired plants
that traditionally serve as baseload generators would be
more expensive than the other fossil fuel plants or non-
carbon-based technologies (renewables and nuclear) in
the competitive pricing carbon reduction case. Therefore, they would be dispatched last and set the marginal
price more often.
200020052010201520200246810121996CentsperKilowatthourMarginal(Competitive)
PricingAverage(Reference)
PricingFigure 86. Projected Electricity Prices in Regulated
and Competitive Electricity Markets,
2000-2020
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs FD03BLW.D080398B and FD03COMP.D080698C.
Figure 88 shows the fraction of time in which each
technology would set the margin in a reference
competitive case and in a 1990-3% competitive case. In
2010, even though total coal-fired generation is much
lower in the 1990-3% case, the amount of time that coal
units set the marginal price is greater than in the
reference competitive case. In both cases, the marginal
plant type shifts from generally older, existing plants
(coal and other fossil steam) in 2010 to newer units
(combined cycle and combustion turbine) in 2020.
Because the carbon price would have a greater impact on
plants with higher emissions, the carbon reduction case
favors more efficient technologies. Thus, in 2020, the
marginal price is most often based on the cost of a new
combustion turbine in the reference case, but new
combined-cycle units set the marginal price more
frequently in the 1990-3% competitive case.
Changing electricity trade patterns are also expected to
affect electricity prices. Although no new construction of
interregional transmission lines is assumed in this
analysis, changes in economy trades still occur. Economy trades take place whenever there is capacity available in a neighboring region that is cheaper than the cost
of the marginal plant that would be needed in the home
region. For example, in the reference competitive case,
Region 1Ñthe East Central Area Reliability Coordination AgreementÑis a net exporter of power, because it
has a large amount of coal capacity that can be operated
inexpensively. In the 1990-3% competitive case, as a
result of the carbon price, coal-fired capacity is more
expensive to operate than other technologies. In this
case, Region 1 becomes a net importer of electricity,
finding generation from neighboring regions less expensive than electricity from its coal-fired units. Because
the marginal cost of generation in a given region
200020052010201520200501001502002503003501996DollarsperMetricTonMarginal(Competitive)
PricingAverage(Reference)
PricingFigure 87. Projected Carbon Prices in Regulated
and Competitive Electricity Markets,
2000-2020
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs FD03BLW.D080398B and FD03COMP.D080698C.
Reference1990-3%Reference1990-3%
020406080100PercentofTotalCoalOil/GasSteamOil/GasTurbineOil/GasCombinedCycle20102020Figure 88. Projected Percentage of Time for
Different Plant Types Setting
National Marginal Electricity Prices,
2010 and 2020
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYCOMP.D080598A and FD03COMP.D080698C.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
is the cost after economy trades are made, changes in
trade patterns directly affect competitive prices. Figure
89 shows the fraction of time in which a trade is
responsible for setting the marginal price in each region
in 2020.
ECAR1ERCOT2MAAC3MAIN4MAPP5NY6NE7FL8SERC9SPP10NWP11RA12CNV13020406080100PercentofTotalReference1990-3%
Figure 89. Projected Percentage of Time for
Interregional Trade Setting Marginal
Electricity Prices, 2020
Note: ECAR = East Central Area Reliability Coordination Agreement
Region; ERCOT = Electric Reliability Council of Texas; MAAC = Mid-
Atlantic Area Council; MAIN = Mid-America Interconnected Network;
MAPP = Mid-Continent Area Power Pool; NY = New York Power Pool;
NE = New England Power Pool; FL = Florida subregion of the
Southeastern Electric Reliability Council; STV = Southeastern Electric
Reliability Council excluding Florida; SPP = Southwest Power Pool;
NWP = Northwest Pool subregion of the Western Systems Coordinat-
ing Council; RA = Rocky Mountain and Arizona-New Mexico Power
Areas; CNV = California-Southern Nevada Power Area.
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYCOMP.D080598A and FD03COMP.D080698C.
Sensitivity Cases
Technological Progress
The development and market penetration of new technologies for consumer use (new air conditioners, furnaces, refrigerators, etc.) and for supplier use (new
generation, transmission, and distribution equipment)
will have a significant impact on the feasibility and costs
of meeting the Kyoto Protocol targets in the U.S. electricity sector. All the carbon reduction cases in this analysis
include substantial improvements in technology, mainly as a function of market penetration. For example, in
the reference case the cost of new advanced combined-
cycle plants declines from a starting point of $572 per
kilowatt to $400 per kilowatt, a 30-percent improvement. In addition, the thermal efficiency of the same
technology improves by roughly 10 percent. The situation is similar for wind plants, the cost of which falls
from around $1,000 per kilowatt to under $750
per kilowatt. It is possible that further improvements
might occur; however, it is impossible to predict to what
degree a concerted effort to reduce carbon emissions
might stimulate the development of new technologies or
reduce the costs of existing ones.
As described in Chapter 2, to look at the potential
impacts of technological innovation, development, and
market penetration, a set of low (currently available)
technology and high technology sensitivity cases were
developed. In the 1990+9% low technology case, the new
generating options available were limited to technologies available in 1998. In the 1990+9% high technology
case, cost and performance characteristics were
assumed to improve at rates consistent with those used
in the high technology sensitivity cases in the Annual
Energy Outlook 1998.
The performance and cost data used in the high technology cases are considered optimistic but not unreasonable. In addition, two new plant types, coal gasification
with carbon sequestration and natural gas combined
cycle with carbon sequestration were made available
beginning in 2010 in the high technology case. The
uncertainty involved in selecting aggressive cost and
performance values for different technologies is considerable. Thus, the results of these sensitivity cases should
not be viewed as indicating which technologies are most
promising but, rather, as indicative of the extent to
which technological innovation might lower the costs of
meeting carbon emission reduction targets.
The key result of the high technology cases is that if new,
more efficient, lower cost technologies evolve, the cost of
meeting the Kyoto Protocol targets could be lowered
significantly. The most important of the generating technologies appears to be the advanced natural gas combined cycle; however, as pointed out above, this is a
product of the high technology assumptions, and it is
impossible to say which technology might progress the
most.
Figure 90 shows the average heat rate (number of Btu
needed to generate each kilowatthour of electricity) for
all natural-gas-fired generating plants. Even in the low
technology case, the average heat rates for natural gas
plants are expected to improve significantly. The
improvement is greater in the 1990+9% case and even
greater in the 1990+9% high technology case.
The effects of assuming lower and higher rates of
technological progress on electricity prices in the carbon
reduction cases are significant. For example, in 2010,
projected electricity prices in the 1990+9% low
technology case are more than 70 percent higher than
those in the reference case (Figure 91). In the 1990+9%
case and the 1990+9% high technology sensitivity case,
they still are higher than in the reference case, but by
only 49 and 36 percent, respectively. In 2020 the price
difference remains quite high in the low technology case
but is only 45 percent and 13 percent in the 1990+9% and
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
1990+9% high technology cases, respectively. Neither of
the carbon sequestration technologies penetrates the
market in the 1990+9% high technology case, because
the projected carbon price is relatively low, and other
high-technology options are more attractive.
Nuclear Power
One carbon-free technology around which there is considerable uncertainty is new nuclear power plants. Currently nuclear power accounts for 20 percent of the
power produced in the United States; however, no new
nuclear power plants have been ordered since 1978, and
the last one to come on line was Watts Bar 1 in 1996. In
recent years, the overall performance of existing plants
has improved dramatically (although several older units
19952000200520102015202002,0004,0006,0008,00010,00012,000BtuperKilowatthour1990+9%LowTechnology1990+9%
1990+9%HighTechnologyFigure 90. Projections of Average Heat Rates for
Natural-Gas-Fired Power Plants in High
and Low Technology Cases, 1996-2020
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs FREEZE09.D080798A, FD09ABV.D080398B, and
HITECH09.D080698A.
1995200020052010201520200246810121996CentsperKilowatthour1990+9%HighTechnologyReference1990+9%
1990+9%LowTechnologyFigure 91. Projected Electricity Prices in High and
Low Technology Cases, 1996-2020
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FREEZE09.D080798A, FD09ABV.
D080398B, and HITECH09.D080698A.
were retired before their 40-year operating licenses
expired). In addition, manufacturers are now working
on designs for a new generation of nuclear power plants,
which are expected to be safer and less costly. As with
any new technology the first few newly designed units
are likely to be quite expensive, but costs should fall as
manufacturers and regulators gain experience with
them.
A special sensitivity case was used in this analysis to
examine the possible impacts of new nuclear power
plants on the carbon reduction cases. Because new
nuclear plants are not economical in the 1990+9% case,
this sensitivity was analyzed against the 1990-3% case.
The 1990-3% nuclear sensitivity case assumes a carbon
emissions target 3 percent below 1990 levels and new
nuclear plant costs about 8 percent lower than the costs
typically associated with the early units of new technologies, with rapidly declining costs as the new technology penetrates the market.
In the 1990-3% nuclear sensitivity case, about 40
gigawatts of new nuclear capacity is built, mostly in the
later part of the projection period (Figure 92). With
higher carbon prices and lower initial construction costs,
the new plants are becoming competitive with other
generating technologies. Nuclear electricity generation
in the 1990-3% nuclear sensitivity case is only 9 billion
kilowatthours higher than in the 1990-3% case in 2010
but is 248 billion kilowatthours higher in 2020.
As discussed above, increases in nuclear capacity and
generation will result in greater amounts of spent
nuclear fuel discharged from nuclear generating units.
The waste must ultimately be moved to a permanent
storage facility. The 1990-3% nuclear sensitivity case
results in a 15-percent increase in projected cumulative
20002005201020152020020406080100120140GigawattsReference1990-3%1990-3%NuclearSensitivityFigure 92. Projections of Nuclear Generating
Capacity in the 1990-3% Nuclear
Sensitivity Case, 2000-2020
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FD03BLW.D080398B, and
NUKE03LC.D081298A.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
spent fuel discharges by 2020, relative to the reference
case.
The future of nuclear power in the United States is
uncertain. Indeed, it may depend on the extent to which
limits are set on carbon emissions in response to the
Kyoto Protocol. The reference and carbon reduction
cases in this analysis assume no new nuclear construction, for several reasons. One is concern about the future
of nuclear waste disposal. The Nuclear Waste Policy Act
of 1982 directed the U.S. Department of Energy (DOE) to
begin accepting spent fuel for permanent disposal in
1998. As yet, however, no permanent waste storage site
is available, and most of the waste is still being stored
on-site by the utilities that operate nuclear power plants.
The current schedule projects 2010 as the earliest that the
proposed site at Yucca Mountain could begin accepting
waste. Given the history of schedule slippage in the
waste repository project, new investors may not commit
to new nuclear power construction until they are certain
that DOE will be prepared to handle the waste. In
addition, public concerns about the safety of both
plant operations and waste disposal will need to be
addressed. The publicÕs association of nuclear power
with its weapons origin, along with highly publicized
accidents at Three Mile Island and Chernobyl, have
heightened safety concerns. Public opposition can cause
delays in project approval, adding risk to investments in
nuclear power.
Another uncertainty is the cost of new nuclear construction. If another nuclear reactor is built in the United
States, it will be one of several new designs that have
been approved by the U.S. Nuclear Regulatory Commission (NRC). Two evolutionary designs have received
final approval from the NRC, and one Òpassively safeÓ
design is still being reviewed. The nuclear industry
hopes that creating relatively few, standardized designs
will bring down construction costs and reduce the time
needed to build future plants. However, past experience
suggests that there will be considerable uncertainty until
the first new units have actually been completed. No
nuclear plant operating in the United States today was
built at its initial estimated cost or schedule. Instead, all
faced both cost overruns and delays in completion.
There is also uncertainty about the useful lifetimes of
currently operating nuclear reactors. In recent years, a
number of nuclear plants have been permanently shut
down well before their license expiration dates, mainly
because of the availability of more economical generation. Operating a nuclear unit for a full 40 years (the
license life) will generally require additional capital
expenditures over the last 10 to 15 years of the plantÕs
life. Whether or not it is economical to incur such costs
will depend on factors specific to each plant, such as
location, other types of generation available, and fuel
prices.
If limitations are placed on carbon emissions in the
future, the relative costs of electricity generation could
shift in favor of nuclear power. This analysis assumes
that license renewal for nuclear plants will be considered, if economical, in all cases with restrictions on
carbon emissions. Operators of nuclear power plants
that are economical will renew the plant licenses,
incurring the costs assumed to be necessary to prepare
the plant for an additional 20 years of operation. In 1998,
two utilitiesÑBaltimore Gas & Electric and Duke
PowerÑfiled applications to renew the operating
licenses of existing plants, the Calvert Cliffs units in
Maryland and the Oconee plant in South Carolina. The
approval process is likely to be lengthy for the first few
plants, but as the NRC develops a standard review
process, more utilities may consider license renewal a
viable option.
Reducing the Impact on the Coal Industry
Coal is the most carbon-intensive fuel used for electricity additional fuel switching results as new capacity is built
production. The carbon emission rate for coal is 78 per-to replace electricity from existing coal units.
cent higher than that for natural gas, which has the lowest Historically, electric utilities have accounted for most of
rate among the fossil fuels. Consequently, carbon reduc-the coal consumption in the United States. Therefore, fuel
tion strategies are expected to affect coal more than other switching to reduce carbon would seriously affect the coal
energy sources. Because of their heavy reliance on coal, industry. In the 1990+9% case, utility coal use in 2020 is
electricity generators have historically produced more projected to be 78 percent lower than in the reference case.
carbon than the other energy sectors. In 1996, more than In the 1990-3% case, coal consumption for electricity pro-
one-third of U.S. carbon emissions resulted from electric-duction would be nearly eliminated in 2020. Absent sig
ity production. nificant changes in other sectors, continued use of coal in
Reductions in carbon emissions in the electricity sector the electricity generation sector is not economical in the
are expected to occur primarily as a result of switching 1990+9% case. Substantially lower coal use would likely
from coal to fuels with lower emission rates, such as natu-have dramatic impacts on mining employment, as fewer
ral gas and renewables. Initially, fuel switching occurs miners would be needed, and on the railroads, whose
mostly by changing the utilization of existing capacity. transportation of the coal used in power plants would
That is, coal-fired plants are operated less frequently decline dramatically.
and gas-fired units are used more extensively. Later on, (Continued on page 94)
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Reducing the Impact on the Coal Industry (Continued)
In the carbon reduction cases, the projected utilization
rates for coal-fired generating capacity are much lower
than the rates at which they have traditionally been oper-
ated. Many coal plants are designed as baseload capacity
that operates almost continuously because they cannot be
restarted quickly or efficiently. The low utilization rates
in the carbon reduction cases are more typical of peaking
or reserve capacity, which is run infrequently. It is unclear
whether coal plants, particularly the larger units, can be
operated either efficiently or economically in this manner.
For purposes of energy security, it may be advisable to
maintain a broad portfolio of fuel options, including some
coal. Coal is the largest domestic energy source and
accounts for most of U.S. energy exports. In contrast,
imports already represented over half of oil supplies in
1996, and imports are projected to make up more than 15
percent of natural gas supplies by 2020 in the reference
case. Consequently, fuel switching from coal to gas would
increase U.S. dependence on foreign energy sources.
Renewable technologies, such as wind and biomass, are
relatively new, and the projected capacity in the carbon
reduction cases far exceeds existing capacity, particularly
in the 1990-3% case.
With these issues in mindÑthe impacts on the coal and
railroad industries, efficient operation of generating
units, and energy securityÑa coal sensitivity case was
prepared that maintained a share of the coal-fired elec-
tricity generation that would otherwise be lost. For the
1990+9% case, the carbon price for coal was adjusted, on a
Btu basis, to be equivalent to that for natural gas. Because
the utilization rates for coal-fired and gas-fired capacity
are determined by the delivered prices and operating effi-
ciencies for the respective fuels, the impact on coal in the
sensitivity case was significantly reduced. Although coal
use would still be lower because of reduced electricity
demand and higher renewable capacity levels, utilization
rates for coal units would more closely resemble current
levels, because the adjustment effectively maintains the
historical cost advantage of coal over natural gas.
The key result of the 1990+9% coal sensitivity case is that
subsidizing some portion of the coal industry would
make it more difficult to reach carbon emission reduction
targets, significantly raising both the carbon price and the
price of electricity (see figures below). In the 1990+9% coal
sensitivity case, the projected carbon price in 2010 is 124
percent higher than the carbon price in the 1990+9% case,
and the price of electricity is 6 percent higher. (The impact
on electricity prices is dampened by the reduced carbon
price for coal users.) The impact on fossil fuel prices other
than coal is also large. In 2020, the differences from the
1990+9% case are 126 and 5 percent, respectively. In con-
trast to the impact in the 1990+9% case, the reduction in
coal use in the sensitivity case is significantly moderated.
By 2020, the reduction in coal consumption by electricity
producers would be only 41 percent relative to the refer-
ence case projection.
19952000200520102015202001002003004001996DollarsperMetricTon1990+9%CoalSensitivity1990+9%
Projected Carbon Prices in the Coal Sensitivity
Case, 1996-2020
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs FD09ABV.D080398B and HICOAL09.D080998B.
1995200020052010201520200246810121996CentsperKilowatthour1990+9%CoalSensitivity1990+9%
ReferenceProjected Electricity Prices in the Coal Sensitivity
Case, 1996-2020
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FD09ABV.D080398B, and
HICOAL09.D080998B.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
5. Fossil Fuel Supply
The impacts on fossil fuel suppliers of policies to limit
carbon emissions will depend on how much carbon is in
each type of fuel: the more carbon in the fuel, the more
severe the impact. If the Kyoto Protocol carbon emissions reduction targets were imposed, the U.S. coal and
oil industries would see lower consumption and production than in the reference case, which does not incorporate the Protocol, whereas the natural gas industry
would expand. Natural gas wins out over coal and oil in
the carbon reduction cases used for this analysis,
because its carbon content per British thermal unit (Btu)
is only 55 percent of that for coal and 70 percent of that
for oil. As a result of higher natural gas consumption
and lower oil and coal consumption, carbon emissions
from natural gas are projected to be higher in the carbon
reduction cases, while emissions from oil and coal are
lower.
Natural Gas Industry
Natural gas is a clean, economical, widely-available fuel
used in more than 58 million homes and more than 60
percent of the manufacturing plants in the United States.
Almost one-quarter of the energy consumed in the
United States comes from natural gas. Most of the natural gas consumed in the United States is produced
domestically from wells in the central part of the Nation.
Gas is transported from the Central United States by
pipelines throughout the country and becomes more
expensive the farther it must be shipped. Yet natural gas
is generally cheaper than oil products, though more
expensive than coal on the basis of heating values.
In 1996 the combustion of natural gas produced 318 million metric tons of carbon emissions in the United States,
about one-fifth of the U.S. total. The industrial sector
was responsible for the biggest share of those emissions,
about 45 percent, followed by residential, commercial,
and electricity generation in order of magnitude. Twelve
years from now, if no carbon reduction measures are put
in place, emissions from natural gas combustion are
expected to be about 100 million metric tons higher than
they were in 1996. Even though the projected emissions
are higher in 2010, the natural gas share of total emissions increases only slightly from 1996.
Natural gas consumption, production, imports, and
prices are all expected to rise in the reference case.
Natural gas consumption increases more rapidly than
consumption of any other major fuel in the reference
case from 1996 to 2010. Natural gas use increases in all
sectors, but consumption by electricity generators more
than doubles to take advantage of the high efficiencies of
combined-cycle units and the low capital costs of combustion turbines. By 2010 the generating capability of
combined-cycle plants increases more than sixfold, and
the generating capability of combustion turbines more
than doubles. More than four-fifths of the new consumption is supplied by increased domestic production.
The remainder comes from increased imports, primarily
from Canada.
Two-thirds of the production increase between 1996 and
2010 is expected to come from onshore resources in the
lower 48 States; the rest is expected to come from Alaska
and lower 48 offshore resources. More production
comes from onshore lower 48 resources, because
roughly 75 percent of current proved reserves are
located onshore, and continued technology improvements make development of the vast onshore unconventional resources more economical. Wellhead prices rise
moderately in the reference case through 2010, reflecting
increased consumption and its impact on resources, as
each type of production progresses from larger, more
profitable fields to smaller, less economical ones.
Policies designed to reduce carbon emissions would
boost natural gas consumption, production, imports,
and prices, principally because natural gas consumption
would displace coal consumption in the electricity supply sector. In response, gas production and imports
would increase, pushing up prices. In the 3-percentbelow-1990 (1990-3%) case, for example, the natural gas
share of the U.S. energy market is projected to increase
from 24 percent in 1996 to 35 percent in 2010, compared
with an increase of only 2 percentage points in the reference case. Following the imposition of a carbon price,
higher prices for natural gas eventually would bring gas
into competition with conservation (i.e., demand reduction) and renewable fuels, slowing the growth of gas
consumption and prices.
Natural Gas Consumption
Natural gas plays a key role in the transition to lower
carbon emissions, because it allows fuel users to
consume the same number of Btu, while emitting less
carbon. Thus, one strategy for fuel users seeking to
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
quickly reduce coal use is to increase gas use. Natural
gas consumption is expected to rise more rapidly in all
the carbon reduction cases than in the reference case,
driven by rising consumption in the electricity supply
sector (Figure 93). Although electricity generators
would produce less electricity in the carbon reduction
cases than in the reference case, they would consume
more natural gas, because relatively high-carbon coal
would be replaced with relatively low-carbon natural
gas. In the 9-percent-above-1990 (1990+9%) case, where
the projected carbon price is relatively low, natural gas
steadily replaces coal; but in the 1990-3% case, with a
higher carbon price, renewable sources of generation
begin to compete successfully with natural gas after
2010.
1995200020052010201520200510152025TrillionCubicFeetOther1990+24%
1990+9%
1990-3%
ElectricityGeneratorsReferenceReference1990+24%
1990+9%
1990-3%
Figure 93. Natural Gas Consumption, 1996-2020
Note: Other uses are for residential, commercial, industrial, and
transportation consumption.
Source: Office of Integrated Analysis and Forecasting, National Energy Mod-
eling System runs KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.
D080398B, and FD03BLW.D080398B.
The projections for natural gas use in the residential,
commercial, industrial, and transportation sectors are
almost always lower in the carbon reduction cases than
in the reference case, because those sectors have
significantly less opportunity to switch from higher-
carbon fuels to lower-carbon natural gas. In the
residential and commercial sectors there is very little
coal use, and most oil consumption occurs in areas
where natural gas pipelines are limited. In the industrial
sector, under the best circumstances, gas consumption
can only hold its own in the carbon reduction cases, as
some boilers switch from coal to gas. In the transportation sector gas has difficulty competing because of the
limited range of compressed natural gas vehicles. As a
result, consumption of natural gas in these sectors is
reduced from the reference case levels because of higher
natural gas prices, which lead to conservation and the
penetration of more efficient technologies.
The pattern of total gas consumption differs in the
carbon reduction cases, depending on the carbon price
(Figure 93). Higher carbon prices, as in the 1990-3% case,
lead to a quick surge in natural gas consumption when
the carbon price takes effect in 2005 and gas gains an
advantage over coal for electricity generation. Later in
the forecast the increase in gas consumption in the 19903% case is moderated, as renewables on the supply side
and energy efficiency gains on the demand side begin to
cut into the natural gas market. Moderate carbon prices
in the 1990+9% case result in a steadier rise in natural gas
consumption, ultimately to higher levels in 2020 than
those expected in the 1990-3% case, because natural gas
prices are not high enough to induce significant levels of
conservation or competition from renewables. Low
carbon prices in the 24-percent-above-1990 (1990+24%)
case lead to an even slower, 1.8 percent annual rise in
consumption, from 1996 to 2020.
From 1950 to the late 1980s, electricity generators were
third among the major users of natural gas, after industrial and residential users. In the late 1980s, they began to
slip into fourth position, after commercial users, where
they are today. When oil and coal prices were declining
in the late 1980s, gas prices were fairly constant. As a
result, oil and coal took a larger share of the growing
electricity generation market while gas use remained
flat. Gas consumption continued to grow in the commercial sector, however, eventually surpassing electricity
sector consumption.
In the future, supply to electric generators is expected to
become more important to the gas industry. In the reference, 1990+24%, and 14-percent-above-1990 (1990+14%)
cases, electricity generators become the second largest
consumers of natural gas, behind the industrial sector,
by 2010. In the higher priced carbon reduction cases,
they become the largest consumers of natural gas by
2010. Consumption of natural gas for electricity generation is projected to reach 12.2 trillion cubic feet in 2010 in
the 1990-3% case, more than 5 trillion cubic feet higher
than in the reference case and more than four times the
1996 level (Figure 93). Electricity generators can be
expected to take a greater interest in natural gas pipeline
capacity expansion by investing in some projects or by
making long-term contracts. Pressure to merge gas and
electricity companies could mount as the advantage of
arbitraging the two markets becomes apparent. Electricity generators might also increase their direct ownership
of natural gas resources or make long-term contracts
with producers in efforts to reduce price volatility.
Natural Gas Production
In most of the carbon reduction cases examined here,
natural gas production, in response to higher consumption and prices, is higher than it is in the reference case
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
projections throughout the forecast period. Production
patterns across the cases are similar to the consumption
pattern: the 1990-3% case shows a sharper rise immediately after 2005, whereas the 1990+9% case shows a
steadier but ultimately higher rise after 2011, and the
1990+24% case is slightly above the reference case. In
2010, production is projected to be 26.2 trillion cubic feet
in the 1990-3% case, 25.9 trillion cubic feet in the
1990+9% case, and 24.1 trillion cubic feet in the
1990+24% case.
The imposition of carbon reduction targets in 2005
causes a sharp increase in natural gas production, due
largely to increased consumption by electricity
generators. The largest production increase is projected
in the 7-percent-below-1990 (1990-7%) case (Figure 94),
because competing coal prices rise faster than in any
other case. The projected increase in natural gas
production between 2005 and 2006 is 1.75 trillion cubic
feet in the 1990-7% case, compared with only 0.39 trillion
cubic feet in the reference case.
Figure 94. Increases in Natural Gas Production,
1983-1984 and 2005-2006
Sources: History: Energy Information Administration, Annual Energy Review
1997, DOE/EIA-0384(97) (Washington, DC, July 1998). Projections: Office of
Integrated Analysis and Forecasting, National Energy Modeling System runs
KYBASE.D080398A, FD24ABV.D080398B, FD1998.D080398B, FD09ABV.
D080398B, FD1990.D080398B, FD03BLW.D080398B, and FD07BLW.
D080398B.
Historically, the largest 1-year increase in gas production was 1.37 trillion cubic feet between 1983 and 1984
(Figure 94). However, in 1984 production was recovering to levels that already had been reached in 1982, and
production in both 1983 and 1985 was down from the
previous year. In contrast, the levels expected in 20052007Ñwhile not unlikelyÑhave never before been
reached. Increasing natural gas consumption during the
initial phases of a carbon emissions reduction program
may be the biggest challenge facing the oil and gas
industry, and careful planning will be required.
Sufficient natural gas resources are available, however,
and infrastructure can be made available, if the price is
right.
All the carbon reduction cases would require more natural gas wells to be drilled to reach the expected higher
production levels. In 1996 about 9,100 successful gas
wells were drilled. In the reference case, some 12,000 are
expected by 2010. The largest annual increase required
in any of the carbon reduction cases is less than 700
wells. A 700-well increase could easily be handled by the
drilling industry, considering that the number of successful gas wells increased by more than 2,000 from 1996
to 1997, when prices increased from $1.55 in 1995 to
$2.23 in 1997. The stimulating effect of prices on drilling
can also be seen in the 1990-3% case, which projects the
highest number of gas wells in 2010, because gas wellhead prices are only a few cents below the 1990-7% case
and oil wellhead prices are higher.
Although the number of available drilling rigs has been
declining since 1982, price increases are a powerful
incentive for increased drilling and the purchase of new
drilling equipment. The number of available drilling
rigs increased by almost 14 percent annually between
1974 and 1982Ñfrom 1,767 to 5,644Ñas natural gas
prices more than quadrupled in real terms.71 About
1,600 drilling rigs were available in the United States in
1996. To support the increased drilling in the carbon
reduction cases, the number of available drilling rigs is
also expected to rise, especially between 2005 and 2010,
when 2-percent increases in rig construction are projected in some years. Given the historical response to rising prices, rig availability is unlikely to be a problem in
the carbon reduction cases.
Increased drilling produces higher reserves in the carbon reduction cases than the reference case, but not until
after 2010. Initially, increased consumption of natural
gas depresses reserves in the carbon reduction cases,
compared with the reference case projection, because
production exceeds reserve additions. After 2010, however, natural gas reserves in the carbon reduction cases
begin to exceed reserves in the reference case, pulled up
by the higher prices. In all the cases, reserves peak late in
the forecast and then begin to decline. The peak year for
reserves is important, because a decline in reserves indicates that production is exceeding reserve additions.
When that happens, wellhead prices tend to rise because
of depletion effects. Reserves peak later in the higher
carbon price cases, as higher wellhead prices sustain
drilling and discoveries over a longer period. In the
1990-3% and 1990+9% cases, reserves peak in 2018, compared with 2013 in the reference case and 1990+24%
case. The highest peak is projected in the 1990-7% case,
at 195.5 trillion cubic feet of reserves in 2018. Projections
71T.A. Stokes and M.R. Rodriguez, Ò44th Annual Reed Rig Census,Ó World Oil (October 1996).
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
of reserve levels depend on the assumed levels of natural gas resources and, as such, are highly uncertain, particularly in the offshore regions of the lower 48 States.
In general, increased reserves indicate that a mineral
industry is well prepared to serve its customers;
however, reserves must be placed in the context of
production to gauge their real adequacy. Reserve-toproduction (RP) ratios provide a measure of the
adequacy of reserves. In this analysis, RP ratios
generally are projected to fall faster in the carbon
reduction cases than in the reference case, because
production exceeds replacement of reserves (Figure 95).
The path of RP ratios over the forecast is heavily
influenced by the production path. When production
increases steeply in the 1990-3% case the RP ratio drops
steeply, whereas in the 1990+24% case the RP ratio drops
more steadily to lower ultimate levels. In 1996, the RP
ratio for natural gas was 8.3. In the reference case, it is
projected to fall to 6.4 in 2020. In the 1990+24% case, the
RP ratio in 2020 is slightly lower than in the reference
case and is at the lowest level of any year in the forecast.
In the 1990-3% and 1990+9% cases, the RP ratio in 2020
exceeds the reference case projection (Figure 95). Thus,
when a higher carbon price is projected, the adequacy of
natural gas reserves improves relative to that projected
in the reference case, because higher gas prices are
expected to lead to more reserve additions.
Most types of natural gas production are projected to be
higher in the carbon reduction cases than in the
reference case, with the exception of associated-
dissolved (AD) and Alaskan gas. AD gas production is a
function of oil production, which is expected to be lower
in the carbon reduction cases than in the reference
case (see the ÓOil ProductionÓ section below). While
19901995200020052010201520200.60.70.80.91.0Index,1990=1.01990+24%
1990-3%
Reference1990+9%
HistoryProjectionsFigure 95. Index of Natural Gas Reserve-to-
Production Ratios, 1990-2020
Sources: History: Energy Information Administration, U.S. Crude Oil, Natural
Gas, and Natural Gas Liquids Reserves 1996, DOE/EIA-0216(96) (Washington,
DC, November 1997), and preceding reports. Projections: Office of Integrated
Analysis and Forecasting, National Energy Modeling System runs KYBASE.
D080398A, FD24ABV.D080398B, FD09ABV.D080398B, and FD03BLW.
D080398B.
increasing in all cases, AlaskaÕs production of natural
gas is expected to be lower in the carbon reduction cases,
because the market for Alaskan gas is limited mostly to
the State. Electricity generators in Alaska are already
more heavily dependent on natural gas than coal, and
their opportunities to switch from coal to gas are limited.
So, electricity generators reduce gas consumption.
Although not included in this analysis, the market for
Alaskan natural gas could grow through increased
exportation of liquefied natural gas, manufacturing of
liquids from natural gas (the Fischer-Tropsch process),
increased industrial manufacturing, or methanol
manufacturing.
Employment in the oil and gas industries generally has
fallen in recent years, as oil production has declined and
productivity has increased. According to the U.S.
Bureau of Labor Statistics, employment in the oil and
gas extraction industries declined from 400,000 employees in 1988 to 322,000 in 1996, a reduction of approximately 20 percent. Over the same period, total oil and
gas production dropped from 34.9 quadrillion Btu to
33.0 quadrillion Btu, a reduction of only 5 percent. Rising productivity accelerated the decline in employment
relative to the decline in production.
In the reference case, higher natural gas production is
projected to more than offset lower oil production, leading to a total oil and gas production level of 38.6 quadrillion per year Btu by 2020. Although employment in the
oil and gas industries is not included in the projections
for this analysis, it is reasonable to expect that the
increase in production would at least reduce the rate of
decline in employment. In the 1990+9% case, total oil
and gas production in 2020 is projected to be 2.1 quadrillion Btu (about 5 percent) higher than in the reference
case, despite a reduction of 0.5 quadrillion Btu in oil production. Thus, the projection for the 1990+9% case
implies that there would be more workers in the natural
gas industry in 2020.
The patterns of U.S. natural gas production projected in
the carbon reduction cases differ among the six onshore
and three offshore producing regions, depending on
consumption and available resources. In the largest
producing regions, the Rocky Mountain and Gulf Coast
onshore and Gulf Coast offshore, production rises
throughout the forecast in the reference case, because
significant amounts of resources are located in those
regions, and technology improvements make more of
the resources available for production in the projection
periodÑparticularly, unconventional resources and
conventional resources at depths greater than 10,000
feet. In the three medium-sized onshore regions,
production peaks during the forecast in the reference
case and declines as production becomes more costly. In
the two least productive regions, the West Coast and
Pacific offshore, production generally falls throughout
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Natural Gas Supply Issues
Uncertainty regarding estimates of the NationÕs natural jack up to $325 million for a deepwater semisubmersible.c
gas resources has always been an issue in projecting pro-Considerable training is needed to develop a workforce,
duction. Although this study relies on resource estimates and many people are reluctant to enter the workforce
made by the U.S. Geological Survey (USGS) and Minerals because of its cyclical history and their consequent fear of
Management Service (MMS), some uncertainty sur-future layoffs. In addition, there are concerns about the
rounds those estimates. Although many analysts believe adequacy of the infrastructure to move gas from offshore
that the USGS estimates are too high, an April 1998 study drilling platforms to the shore.
by the Gas Research Institute (GRI)a contends that the To address these uncertainties, several studies are being
industry has Òsignificantly underestimatedÓ the growth undertaken. For example, former Secretary of Energy
potential of existing fields and should look to the Midcon-Federico Pea commissioned the National Petroleum
tinent, onshore Gulf Coast, East Texas, and San Juan Basin Council (NPC) to undertake a study of whether the indusfor reserve growth. GRI has increased its reserve esti-try will be able to respond to meet projected demands,d
mates for those areas but maintains that assessing the and the Natural Gas Supply Association (NGSA) is work-
actual amounts remains a difficult task. Uncertainty is a ing on a report that will analyze whether the industry can
particular problem in the offshore area (which the indus-meet increased demand projections without increasing
try hopes will provide significant supplies) because not wellhead prices.e
much historical data is available for offshore production.
Not all of the industryÕs original hopes may be realized, Royalty issues are also of concern. The Assistant Secretary
however. For example, the sub-salt area, which until of the Interior for Land and Minerals Management, Rob-
recently was regarded as a promising supply source, is no ert L. Armstrong, raised the issue of a possible increase in
longer considered to be as promising. the deepwater royalty rate to 16.67 percent from 12.5 per
cent after the current Òroyalty holiday.Ó Although the
Another concern about supply availability is access to
proposal has not been supported by Congress, the uncer
public land for drilling. Drilling moratoria have placed
tainty about royalty relief that stems from any talk about
offshore areas in the eastern Gulf of Mexico, North Caro
changes could place a damper on investment.
lina, and California off limits, and drilling is limited in
some areas of the West because of concern about emis-Despite the above concerns, considerable investment is
sions. Substantial resources in the Arctic National Wild-being made in the industry. According to Arthur Anderlife Refuge (ANWR)b are also restricted from drilling, senÕs 10th annual ÒU.S. Oil & Gas Industry Outlook Sur-
although the current inability to market natural gas from vey,Ó executives of most U.S. exploration-and-production
northern Alaska renders the accessibility issue moot. companies plan to increase spending in 1998.f As an
In addition to concerns about supply availability, there is example, Shell has recently announced plans to spend
widespread speculation in the industry as to whether the nearly $1 billion to develop three oil-and-gas fields in the
deepwater Gulf of Mexico.g
level of production that would be needed to meet the
hefty increases in demand projected in various carbon Clearly, there are conflicting opinions throughout indusreduction scenarios could be achieved, given current try as to whether steep increases in production can be
worldwide shortages of offshore rigs and skilled person-achieved in a timely fashion, even with significant
nel. Virtually every available offshore rig was in use increases in wellhead prices. In order for this to happen,
throughout 1997, and capacity expansion has been lim-the industry first needs to be confident that the demand
ited by uncertainty surrounding the actual demand for will be there, so that the necessary investments in infra-
new rigs. The lead time for construction of new rigs is 2 to structure, rigs, drilling, and manpower development can
3 years, and costs range from $115 million for a 350-foot be made in time.
aAssessment and Characterization of Lower-48 Oil and Gas Reserve Growth, prepared by Energy and Environmental Analysis, Inc. for
the Gas Research Institute (Chicago, IL, April 1998).
bANWR resources are not included in this analysis.
cÒSimmons: Offshore Rig Shortage Looms,Ó Oil and Gas Journal (April 27, 1998), p. 24.
dÒProducers Question Studies Showing Rising Gas Demand But Flat Prices,Ó Inside F.E.R.C.Õs Gas Market Report (May 15, 1998), p. 2.
eÒConcerned About Prices, NGSA To Throw Shadow Over Rosy Supply Pictures,Ó Inside F.E.R.C (May 11, 1998), p. 7.
fÒE&P Companies Plan To Boost Spending Despite Variety of ConcernsÑStudy,Ó Inside F.E.R.C.Õs Gas Market Report (December 26,
1997), p. 9.
gÒShell To Spend $1 Billion To Develop Three Gulf Deep-Water Discoveries,Ó Inside F.E.R.C.Õs Gas Market Report (April 3, 1998), p. 9.
the forecast, as a small resource base precludes signifi-Natural Gas Imports
cant responses to higher prices. Regional production in
Natural gas imports are projected to be higher in all the
the carbon reduction cases is generally higher than in the carbon reduction cases than in the reference case, as the
reference case because prices are higher. In regions industry works to meet rising demands for natural gas.
where production peaks during the forecast, production In 2010, net natural gas imports are projected to be 4.7
tends to peak sooner in the carbon reduction cases, trillion cubic feet in the reference case and up to 5.7 trilbecause more gas is produced earlier. lion cubic feet in the carbon reduction cases. Net imports
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
are highest in the cases with high carbon prices, where
imports surge as the carbon prices are imposed and
remain at higher levels than those projected in the reference case. However, by the end of the forecast, the highest levels of net imports are expected in the 1990-3%
case, rather than the 1990-7% case, because consumption
is projected to be higher in the 1990-3% case.
In most of the carbon reduction cases, the majority of the
higher imports come from Canada in 2010, but in the
1990-7% and 1990-3% cases at least half of the increase
comes from Mexico. Even though Canada would be subject to its own carbon restrictions, it has a large enough
resource base to increase both domestic consumption
and exports. The Canadian Gas Potential Committee
estimated in 1997 that the Western Canada Sedimentary
Basin contained 263 trillion cubic feet of marketable
gas.72 In 2010 Natural Resources Canada projects Canadian natural gas consumption at 3.6 trillion cubic feet,
up 600 billion cubic feet from 1995.73 If carbon reduction
targets were imposed, CanadaÕs gas consumption
would likely be higher. For example, if gas consumption
in Canada were 10 percent higher in 2010 as a result of
carbon restrictions, as projected for the United States in
the most stringent carbon reduction cases, it would
reach 4.2 trillion cubic feet in 2010. Even at that level,
however, U.S. prices are expected to be high enough to
continue the flow of imports from Canada.
In the carbon reduction cases, Mexico is a net exporter of
natural gas to the United States in 2010, whereas it is a
net importer in the reference case. Mexico begins to
export gas to the United States in the carbon reduction
cases in response to higher consumption and higher
wellhead prices. Net imports of liquefied natural gas
(LNG) reach one-third of a trillion cubic feet annually in
all the carbon reduction cases but do so more quickly in
the cases with higher projected carbon prices.
Natural Gas Pipelines
Interstate natural gas pipeline capacity additions would
need to be higher in the carbon reduction cases than they
are in the reference case projections, but they are
expected to be manageable. In the reference case, cumulative additional natural gas pipeline capacity crossing
the 12 regions used for this analysis are projected to
increase to 52.5 trillion cubic feet of design capacity in
2010 from the 1996 capacity of 43.0 trillion cubic feet. The
most significant increase is projected from 1998 to 2001,
when capacity increases by 6.3 trillion cubic feet because
of increasing consumption in the Midwest and Northeast not because of carbon reduction policies. During
the 1998-2001 period, the Alliance pipeline is expected
to come down to the Midwest from Canada, and
the Maritimes/Northeast and Portland Natural Gas
Transmission System pipelines are expected to come
down from Sable Island in Canada to the northeastern
United States. After 2001, pipeline capacity is projected
to increase more gradually through 2010.
In the carbon reduction cases, the largest 1-year increase
in pipeline capacity after 2001 is seen from 2011 to 2012
in the 1990+9% and 1990+14% cases, when capacity
increases by 1.6 trillion cubic feet. The capacity increases
in this period are primarily out of Texas, Louisiana, and
Oklahoma, through the South, to the southern coastal
States in response to growing consumption. The largest
increase soon after imposition of the carbon price is from
2006 to 2007 in the 1990-3% case, when capacity is projected to increase by 1.4 trillion cubic feet. The increase is
mainly from west to east, from the Texas-Oklahoma-
Louisiana region to the Middle South.
Historically, the largest recent annual increase in pipeline capacity was 1.6 trillion cubic feet from 1991 to 1992,
partly because of the construction of four major pipelines into California from the Mountain States (Kern
River, Mohave, El Paso, and Transwestern) and two
major pipelines out of Canada (Great Lakes into the
Midwest and Iroquois into the New York/New England
area). In view of the historical and expected near-term
increases in capacity, capacity expansion is not likely to
be a problem in any carbon reduction scenario, as long
as pipeline requirements are known 2 to 3 years in
advance.
Natural Gas Prices
Natural gas prices are higher in the carbon reduction
cases than in the reference case, both at the wellhead and
at the burner tip. At the wellhead, higher production to
satisfy increased natural gas consumption, in the face of
increasingly expensive resources, boosts prices. At the
burner tip, adding carbon prices to resource costs could
more than double some end-use prices.
In the reference case, lower 48 wellhead natural gas
prices are projected to rise from $2.24 per thousand cubic
feet in 1996 to $2.33 in 2010 in 1996 dollars (Figure 96).
The 2010 wellhead prices are more than 40 cents per
thousand cubic feet or 19 and 29 percent higher in the
1990-3% and 1990+9% cases, which project higher
consumption and the use of increasingly expensive
resources. The highest wellhead prices in 2010 are seen
in the 1990-7% case at $3.03 per thousand cubic feet,
where carbon prices are highest in 2010.
The pattern of natural gas wellhead prices is similar to
the consumption and production patterns (see above).
In the reference case, prices rise gradually, but in the
carbon reduction cases prices rise quickly after a carbon
72Canadian Gas Potential Committee, Natural Gas Potential in Canada (Calgary: University of Calgary, 1997), Figure 1.2.
73Calculated from Natural Resources Canada, CanadaÕs Energy Outlook 1996-2020 (Ottawa: Natural Resources Canada, 1997), Annex C.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
price is imposed in 2005. In the cases with higher pro-slow the overall rate of growth in natural gas conjected carbon prices, gas prices rise more quickly, then sumption. When moderate carbon prices are projected,
flatten out as energy conservation on the demand side gas prices rise more steadily but ultimately reach higher
and renewable energy production on the supply side levels.
Natural Gas Pipeline Expansion
There are three ways of increasing pipeline capacity. The This may work to speed up the approval process. The secsimplest and least expensive is to increase throughput by ond is projected increases in natural gas demand, indeincreasing compression at compressor stations. The sec-pendent of the Kyoto Protocol. Demand growth is already
ond is through a process called Òlooping,Ó in which paral-anticipated to result from electric utility restructuring
lel pipe is laid next to existing pipe to increase capacity activities in a growing number of States, retirements of
along an existing route. The third, and most costly, is to nuclear facilities, and measures included in the Presibuild new pipe, usually entailing additional costs for land dentÕs Climate Change Technology Initiative (a $6.3 biland/or right-of-way. lion initiative), which will proceed regardless of the fate
Two key criteria must be met in order for an expansion of the Kyoto Protocol. If the anticipated increases in
project even to be proposed: (1) the existence of demand demand do materialize, they could provide the impetus
must be shown, and (2) the project must be proven to be for much of the capacity increase that would be needed in
financially viable. Four steps are needed to bring a project the event that the Kyoto Protocol is ratified.
to fruition: (1) an open season of 1 to 2 months during On the other hand, financial considerations are creating
which bids for the proposed capacity are solicited and some uncertainty about the responsiveness of the pipe-
received, (2) a planning stage of 3 to 5 months, (3) filing line industry. A major issue is whether the economic cliwith the Federal Energy Regulatory Commission (FERC) mate for investment will continue to be favorable.
for approval, with an average time of 15 months (ranging Pipeline owners are claiming that they currently face con-
from 5 to 18 months), and (4) an actual construction stage, siderable risk because of increased competition and the
which averages 6 to 9 months. Barring unforeseen delays, threat of capacity turnback. While the Natural Gas Supply
capacity can be added with a lead time of 2 to 3 years. Association (NGSA) contends that the FERCÕs current
Problems that can slow down the process include the fil-policy for determining pipeline returns on equity is fair
ing of environmental impact statements and acquiring and properly accounts for the risk faced by the pipeline
necessary approvals, and changes in market conditions industry, the Interstate Natural Gas Association of Amer(such as the changing market conditions that affected the ica (INGAA) contends that the CommissionÕs generic
Altamont project, which was approved in 1990 but still method artificially lowers allowed returns, and that rates
has not been constructed). FERC has seen a significant should be calculated on a case-by-case basis. Pipeline
increase recently in the number of comments and protests executives contend that the 12-to 13-percent average rate
received on proposed expansion projects. Another poten-of return for pipelines in 1996 was far lower than the 20
tial problem is competition between two pipelines for percent rate earned by most public companies.b In
expansion to serve the same market, such as the recent response to the industryÕs concerns, the FERC is currently
competition to move supplies from Western Alberta, evaluating possible changes in the method used to calcu-
Canada, into the Midwest. late pipeline returns. As even more risk is associated with
Greater increases in pipeline capacity than those pro-the levels of expansion forecast in the carbon reduction
jected in the carbon reduction cases are likely between cases, a key question is, ÒWho will assume the added
now and 2000. More than 116 expansion projects have riskÑutilities that need the gas, other consumers willing
already been proposed. For the 71 projects for which pre-to contract for gas, or the pipeline companies?Ó
liminary estimates are available, the estimated total costs Despite the obvious uncertainties, recent history shows
exceed $11 billion. In 2000 alone, $4.6 billion in expendi-that the industry can handle expansions of the same order
tures is anticipated, as several major projects may be com-of magnitude as those being projected as a result of the
pleted.a The added capacity is needed to provide access to Kyoto Protocol. Changes in the pipeline industry between
new and expanding production areas, such as Canada now and the time of the rapid capacity expansions that
and the deep offshore, and to accommodate shifts in are expected to be needed to support electricity suppliers
demand patterns, such as new demands for natural gas to after the enactment of carbon reduction targets will be key
replace electricity generation capacity lost as a result of to the industryÕs ability to respond. This is one of the
nuclear retirements. issues that the upcoming NPC study commissioned by
Although there is speculation within the industry as to former Secretary of Energy Federico Pea will be address-
whether the needed expansions can occur, two factors ing. Several other industry studies are underway to
support an optimistic outlook. The first is changes in evaluate the industryÕs ability to respond, including an
FERC policy, which now leans more toward letting the INGAA study that Òwill be looking at what needs to be
pipelines assume more risk rather than requiring firm done for the pipeline industryÓ to achieve a market of 30
contracts to be in place before approving an expansion. trillion cubic feet by 2010.c
aEnergy Information Administration, Office of Oil and Gas, EIAGIS Natural Gas Geographic Information System Natural Gas Proposed
Construction Database (Washington, DC, preliminary as of April 1998).
bÒNGSA: Return-on-Equity Fair Despite Protests by Pipelines,Ó Natural Gas Week (March 9, 1998), p. 6.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
1970198019902000201020200123451996DollarsperThousandCubicFeet1990+24%
1990-3%
Reference1990+9%
HistoryProjectionsFigure 96. Natural Gas Wellhead Prices, 1970-2020
Sources: History: Energy Information Administration, Annual Energy Review
1997, DOE/EIA-0384(97) (Washington, DC, July 1998). Projections: Office of
Integrated Analysis and Forecasting, National Energy Modeling System runs
KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.D080398B, and
FD03BLW.D080398B.
On a regional basis, access to end-use markets heavily
influences wellhead prices. Some of the lowest wellhead
prices are seen in the Rocky Mountain region, where
access to eastern markets is limited by pipeline
constraints. This is balanced by wellhead prices in the
two largest producing regions in this study, the Gulf
Coast onshore and offshore, which have prices slightly
above the national average in 1996. Wellhead prices are
currently higher in the Northeast region than any other,
where demand is significant and growing. Regional
prices are generally higher in the carbon reduction cases,
because of higher demand. Though more exaggerated,
the pattern of growth across regions is much the same as
in the reference case.
The projected end-use prices for natural gas in the
carbon reduction cases are double the prices in the
reference case at their peak in the most extreme cases.
The main components of end-use prices are the
wellhead price, the carbon price, and transmission and
distribution margins. On a percentage basis, residential
prices are the least affected by the imposition of carbon
prices, and the prices to electricity generators are the
most affected (the projected carbon price is almost the
same for both sectors, but gas prices are significantly
higher in the residential sector). In 1996, natural gas
prices for end users in the residential sector, which has
the largest number of end-use customers, were $6.37 per
thousand cubic feet. In the 1990-3% case, residential
prices are expected to peak in 2013 at $11.31 per
thousand cubic feet (in 1996 dollars), compared with
$5.71 in the reference case (Figure 97). The difference is
almost entirely attributable to the carbon price, which
adds $4.20 to residential gas prices in 2013. Wellhead
prices and transmission margins are also projected to be
higher, however, because of higher in total gas
consumption even though residential consumption is
197019801990200020102020024681012141996DollarsperThousandCubicFeet1990+24%
1990-3%
Reference19901990-7%
1990+9%
1990+14%
HistoryProjectionsFigure 97. Delivered Natural Gas Prices in the
Residential Sector, 1970-2020
Sources: History: Energy Information Administration, Annual Energy Review
1997, DOE/EIA-0384(97) (Washington, DC, July 1998). Projections: Office of
Integrated Analysis and Forecasting, National Energy Modeling System runs
KYBASE.D080398A, FD24ABV.D080398B, FD1998.D080398B, FD09ABV.
D080398B, FD1990.D080398B, FD03BLW.D080398B, and FD07BLW.
D080398B.
lower. In the residential sector, margins for distribution
services are higher because fixed costs must be spread
over a smaller consumption base. In the 1990-7% case,
residential prices are projected to peak at $12.10 per
thousand cubic feet in 2013, because this case has the
highest carbon prices. End-use prices in the carbon
reduction cases follow a pattern similar to the pattern of
carbon prices.
The story is much the same for the electricity supply
sector, where the most growth in consumption is
expected, except that the projected difference between
wellhead and end-use margins is much smaller (less
than 10 cents per thousand cubic feet) in the 1990-3%
case in 2010. The differences in margins is not as high as
in the residential sector because higher electric generator
consumption allows gas utilities to spread their fixed
costs over a larger volume of gas. In 1996, delivered
prices to electricity generators were $2.70 per thousand
cubic feet. At their peak in 2014, prices in the 1990-3%
case are projected to be $8.27 per thousand cubic feet,
compared with $3.05 in the reference case. As in the
residential sector, the higher the carbon price, the higher
the end-use price.
End-use prices for natural gas are affected by their distance from the sources of supply. End-use prices in the
Texas-Louisiana region are currently less than half of
prices in New England, for example. Although New
England currently has the highest average natural gas
end-use prices, prices are expected to be highest in the
Mid-Atlantic region in a few years, as new pipeline projects are completed into New England and as consumption for electric generation increases. Regional prices are
generally higher in the carbon reduction cases than in
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
the reference case, because of higher demand. They
show much the same pattern of growth as in the reference case.
Oil Industry
Oil is a larger source of energy than natural gas. Nearly
40 percent of U.S. energy comes from oil, most of which
is used to fuel our vehicles and industry. Gasoline and
diesel oil fuel more than 200 million vehicles, one for
every 1.3 people in the country. Almost half of our oil
was imported by tanker ship from Venezuela, Mexico,
Saudi Arabia, and other countries at a cost of more than
$60 billion in 1996. The rest is produced domestically,
mainly in Texas, Alaska, Louisiana, and California, and
shipped by pipeline and tanker. With the exception of
residual fuel oil, this easily moved, universally-available
liquid tends to cost more per Btu than other forms of
energy.
In 1996, oil combustion produced 621 million metric
tons of carbon emissions in the United States, over two-
fifths of the total and more than those produced from
burning coal. The transportation sector was responsible
for the major share of those emissions, almost three
quarters, followed by industrial and residential emissions in order of magnitude. In 2010, if no carbon reduction measures are put in place, emissions from oil
combustion are expected to be more than 130 million
metric tons higher than they were in 1996, although their
share of the total will be slightly lower.
U.S. oil consumption is expected to increase between
1996 and 2010 in the reference case, despite a projected
decline in domestic oil production. Most of the growth is
expected in the transportation sector, where oil consumption is projected to increase by almost 30 percent
from 1996 to 2010. About half the increase comes from
light-duty vehicle travel and more than 20 percent from
increased air travel. Oil use in the industrial sector is
projected to increase by about 15 percent between 1996
and 2010, with more than three-fifths of the increase
coming in refining and petrochemical feedstocks. As a
result of these increases, oilÕs share of the energy market
will increase slightly over time.
While petroleum production from conventional sources
in the lower 48 States and in Alaska is expected to fall
between 1996 and 2010, enhanced oil recovery and offshore production are expected to increase, but not
enough to prevent an overall decline. Net imports of
crude oil and petroleum products are projected to rise to
fill the gap between consumption and production. In the
reference case, almost three-fifths of the U.S. oil supply
in 2010 is projected to come from imports, with about
three-fourths of total imports entering the country in the
form of crude oil and the rest as finished or unfinished
products. Gross refinery margins are projected to
increase on the strength of increased refinery throughput and capacity expansion. End-use prices show little
change in the reference case, as increases in world oil
prices are balanced by assumed reductions in motor fuel
taxes. Federal taxes on gasoline and diesel fuel are
assumed to stay constant in real dollar terms, which
would mean a decline in nominal terms.
Policies aimed at reducing carbon emissions would lead
to lower consumption, production, imports, and refinery margins for the U.S. oil industry. On the other hand,
end-use prices and market share would be higher.
Higher end-use pricesÑreflecting new carbon pricesÑ
would reduce consumption in the carbon reduction
cases, lessening the need for domestic production and
foreign imports. Refinery margins in those cases would
be lower, because consumption of petroleum products
and expansion of refinery capacity are projected to be
lower than in the reference case. Despite the lower levels
of oil consumption projected in the carbon reduction
cases, oilÕs share of the energy market would be higher
as a result of an even larger drop in coal use. For
example, in the 1990-3% case, oil is projected to claim 41
percent of the domestic energy market in 2010 and coal
just 7 percent, as compared with their respective 38percent and 22-percent shares in 1996.
Oil Consumption
Oil consumption is expected to be lower in the carbon
reduction cases than in the reference case (Figure 98),
with most of the difference in the transportation sector.
Current petroleum product consumption is at about the
previous peak level of consumption reached 20 years
ago. In the reference case, consumption rises from 18.5
million barrels per day in 1996 to 22.5 million barrels per
day in 2010. In the carbon reduction cases, higher carbon
prices overwhelm lower crude oil prices and lead to
lower levels of oil consumption in 2010Ñ22.0 million
barrels per day in the 1990+24% case and 20.0 million
barrels per day in the 1990-3% case. Consumption in the
transportation sector is particularly affected. More than
65 percent of the difference between the reference and
the carbon reduction cases in 2010 is in the transportation sector.
In the reference case, petroleum consumption rises
throughout the forecast. Consumption also rises continually throughout the forecast in the carbon reduction
case with the lowest projected carbon prices, the
1990+24% case. In the other cases, consumption declines
during the 2005-2009 period after the carbon price is
imposed. The higher the carbon price, the greater the
decline in consumption. After 2009, consumption rises
in all cases through the rest of the forecast, because highway and air travel increase while carbon prices change
modestly.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
1970198019902000201020200510152025MillionBarrelsperDay1990+24%
1990-3%
Reference1990+9%
HistoryProjectionsFigure 98. Petroleum Consumption, 1970-2020
Sources: History: Energy Information Administration, Annual Energy Review
1997, DOE/EIA-0384(97) (Washington, DC, July 1998). Projections: Office of
Integrated Analysis and Forecasting, National Energy Modeling System runs
KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.D080398B, and
FD03BLW.D080398B.
Oil use in the transportation sector is expected to absorb
the largest share of the projected declines between 2005
and 2009, accounting for more than 85 percent of the
total drop in oil consumption in the three most stringent
carbon reduction cases, with smaller reductions in the
residential, commercial, and industrial sectors. In the
1990-3% case, transportation consumption falls from
14.2 million barrels per day in 2004 to 13.5 million
barrels per day in 2009, followed by a continuing
increase to 15.0 million barrels per day in 2020. During
the period of declining consumption, high carbon prices
produce rapid increases in transportation fuel prices.
After 2009, when consumption begins to rise, fuel prices
in the transportation sector are generally level or
declining, as the carbon prices decline.
Oil Production
U.S. oil production declines steadily throughout the
forecast both in the reference case and in the carbon
reduction cases, but lower consumption and diminishing oil reserves in the later years of the carbon reduction
cases lead to larger production declines. In the reference
case, crude oil production is projected to drop from 6.5
million barrels per day in 1996 to 5.9 million barrels per
day in 2010, compared with 5.8 million barrels per day in
the 1990+24% case and 5.7 million barrels in the
1990+9% and 1990-3% cases in 2010. The higher the carbon price, the lower is the crude oil price, the less is the
buildup in reserves, and the lower is oil production,
because the higher carbon prices overwhelm lower
crude oil prices.
Domestic oil drilling activity rises steadily in the reference case and in the least stringent carbon reduction
case. In the more stringent cases, drilling generally
increases, but declines are projected in the middle years
19901995200020052010201520200.00.51.01.52.02.5BillionBarrels1990+24%
1990-3%
Reference1990+9%
HistoryProjectionsFigure 99. Lower 48 Crude Oil Reserve Additions,
1990-2020
Sources: History: Energy Information Administration, U.S. Crude Oil, Natural
Gas, and Natural Gas Liquids Reserves 1996, DOE/EIA-0216(96) (Washington,
DC, November 1997), and preceding reports. Projections: Office of Integrated
Analysis and Forecasting, National Energy Modeling System runs KYBASE.
D080398A, FD24ABV.D080398B, FD09ABV.D080398B, and FD03BLW.
D080398B.
of the forecasts, when high carbon prices depress oil
prices. The lowest levels of drilling activity are seen in
the cases with the highest projected carbon prices, which
result in the lowest wellhead prices.
Despite the projections of increased oil drilling both in
the reference case and in the carbon reduction cases, oil
reserves are not expected to rise over the forecast period.
Declining reserves are projected in all the cases, because
reserve additions do not exceed production. For example, all the carbon reduction cases show reserve additions of only 1.9 billion barrels in 2005 (Figure 99), when
production is projected to be about 2.2 billion barrels for
the year. Thus, oil reserves decline. In the reference case,
higher oil prices sustain enough drilling for annual
reserve additions to peak at 2.0 billion barrels in 2009. In
the more stringent carbon reduction cases, however,
declining oil prices cause reserve additions to fall after
2005. The inability of the oil industry to replace reserves
has less effect on oil prices than the inability to replace
gas reserves has on gas prices, because oil prices are set
in a world market, and because the RP ratios for oil are
actually projected to rise.
Oil RP ratios, which are indicative of the industryÕs ability to sustain production, rise over the forecast both in
the reference case and in the carbon reduction cases, as
oil production falls more quickly than reserves. The RP
ratio in the reference case rises from 7.1 in 1996 to 7.3 in
2010. RP ratios in the carbon reduction cases are slightly
lower, because the low oil prices in the carbon reduction
cases depress reserve additions more than production.
Most types of oil production are projected to be lower in
the carbon reduction cases than in the reference case,
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
and most of the lower production in the carbon reduction cases is in lower 48 onshore conventional and
enhanced oil recovery productionÑthe two types of
production that are the most responsive to lower oil
prices. In 2010, conventional onshore lower 48 oil production is 90,000 barrels per day lower in the 1990-3%
case than in the reference case, 60,000 barrels per day
lower in the 1990+9% case, and 20,000 barrels per day
lower in the 1990+24% case. Enhanced oil recovery is
50,000 barrels per day lower in the 1990-3% case, 40,000
barrels per day lower in the 1990+9% case, and 10,000
barrels per day lower in the 1990+24% case.
Regionally, oil production is generally lower in the carbon reduction cases than in the reference case. It is significantly lower in the Southwest (western Texas and
eastern New Mexico), in the Rocky Mountains, and in
the offshore Gulf Coast, which are the largest producing
regions. In the 1990-3% case, for example, the projected
production in 2010 in each of these regions is 40,000 barrels per day less than projected in the reference case. In
the Midcontinent region (Kansas, Oklahoma, and
Arkansas), oil production is slightly higher in the more
stringent carbon reduction cases than in the reference
case, because increased drilling for gas in the carbon
reduction cases leads to more oil discoveries and greater
oil production; however, the peak difference is only
about 10,000 barrels per day.
Regional crude oil prices are most affected by the quality
of the crude oil. West Coast crude oil prices are generally
lower than prices in the rest of the Nation because the
density of West Coast crude oils is higher. Dense crude
oils contain less of the higher-valued light products, like
gasoline or diesel fuel, so their value is lower. Crude oil
prices are lower in the carbon reduction cases, but the
relationships among regional prices is the same as in the
reference case.
Oil Imports
The projections for net imports of crude oil and
petroleum products are lower in the carbon reduction
cases than in the reference case, because oil consumption
is projected to be lower, with domestic sources
providing a greater share of the NationÕs oil needs. As a
share of total consumption, net oil imports reach 59
percent in 2010 in the reference and 1990+24% cases but
only 54 percent in the 1990-3% case and 56 percent in the
1990+9% case. In all the cases, the projected import
levels are above current levels, which are the highest yet
recorded. The total value of net oil imports in 2010 is
$103 billion in the reference case but only $96 billion in
the 1990+24% case, $82 billion in the 1990+9% case, and
$70 billion in the 1990-3% case (Figure 100). Both values
are well below the 1980 peak of $138 billion (in 1996
dollars). Even in 2020, the total projected expenditures
for oil imports in the reference case are only $123 billion.
Net crude oil imports rise steadily throughout the
forecast in the reference case and in the 1990+24% and
1990+9% cases. In the 1990 stabilization, 1990-3%, and
1990-7% cases, however, net crude oil imports begin to
fall when the carbon price is first imposed, bottoming
out in 2009 before beginning to rise again. Imposition of
relatively high carbon prices causes oil consumptionÑ
and importsÑto fall temporarily in these cases.
Net petroleum product imports are affected more
strongly than crude oil imports in the carbon reduction
cases, because imported crude oil is generally more
valuable to U.S. refiners than imported products inasmuch as profits are maximized only at high rates of
refinery utilization. In the reference case, net product
imports rise from 1.1 million barrels per day in 1996 to
3.1 million barrels per day in 2010. In comparison, the
corresponding increases are only 70,000 barrels per day
in the 1990-3% case, 760,000 barrels per day in the
1990+9% case, and 1.64 million barrels per day in the
1990+24% case. In the reference case and in the less stringent carbon reduction cases, net petroleum product
imports exceed the historic 1973 peak of 2.8 million barrels per day at some time during the forecast, beginning
as early as 2009 in the reference case, for example.
In the two most stringent reduction cases, unlike the
other cases, product imports fall from 2004 through 2008
because of a decline in petroleum product consumption,
and net product imports stay below the historic peak
through 2020. In the 1990-7% case, net product imports
remain below even their 2004 peak of 2 million barrels
per day through 2020.
19801990200020102020020406080100120140Billion1996Dollars1990+24%
1990-3%
Reference19901990-7%
1990+9%
1990+14%
HistoryProjectionsFigure 100. Net Expenditures for Imported Crude
Oil and Petroleum Products, 1974-2020
Sources: History: Energy Information Administration, Monthly Energy
Review June 1998, DOE/EIA-0035(98/06) (Washington, DC, June 1998).
Projections: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD1998.
D080398B, FD09ABV.D080398B, FD1990.D080398B, FD03BLW.D080398B,
and FD07BLW.D080398B.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Petroleum Products
Consumption of almost all the individual petroleum
products is projected to be lower in the carbon reduction
cases than in the reference case, because higher prices
lead to lower demand. Gasoline consumption in 2010 is
3 percent lower in the 1990+24% case than in the reference case, 8 percent lower in the 1990+9% case, and 15
percent lower in the 1990-3% case, in direct response to
the projected carbon prices. Distillate, diesel, and jet fuel
consumption levels are also lower. Residual fuel is the
least affected, because it is projected to compete successfully with natural gas and coal in the industrial sector.
The projected consumption of residual fuel in 2020 is
actually higher in the 1990+9% case than in the reference
case because of higher industrial demand.
In 2010, the projected product shares of total petroleum
consumption are approximately the same in the reference, 1990+24%, 1990+9%, and 1990-3% cases: 43 percent
for gasoline, 18 percent for distillate, 11 percent for jet
fuel, 4 percent for residual fuel, and 24 percent all other
products. The gasoline and jet fuel shares are slightly
lower in the 1990-3% case, with slightly higher shares for
the other, mostly heavier products. Purely on the basis
of carbon content, consumption might be expected to
move away from the heavier products, which have more
carbon, and toward the lighter products; however,
sector-by-sector tradeoffs with conservation and with
other fuels are more critical to the shares. For example,
residual fuel oil consumption in the industrial sector in
2010 is higher in the 1990-7% case than in the reference
case, because the projected carbon price makes residual
fuel less expensive than coal.
Ethanol
Ethanol consumption is generally expected to be higher
in the carbon reduction cases than in the reference case
(Figure 101). The United States consumed 80,000 barrels
per day of ethanol in 1996 and is expected to consume
180,000 barrels per day in the reference case in 2010.
Consumption is generally higher in the carbon reduction cases because of the growth in inexpensive
cellulose-derived ethanol and because ethanol is exempt
from the addition of a carbon price. However, ethanol
consumption trends are quite complex because of
changing legislation, production, and tax patterns.
In 1996 almost all ethanol consumed was blended
directly into gasoline, but over the forecast period more
ethanol is expected to be converted into an intermediate
blending component or used in new types of alternative-
fueled vehicles. At present ethanol is blended into gasoline as an ÒoxygenateÓ for reformulated and high oxygenated gasoline; up to 10 percent ethanol is also
blended into traditional gasoline as a petroleum substitute. Oxygenates are used to reduce carbon monoxide
emissions, as in oxygenated gasoline, or reduce the
precursors of ozone pollution, as in reformulated gasoline. Besides ethanol, the other primary oxygenate is
methyl tertiary butyl ether (MTBE). One gallon of ethanol contains approximately twice the amount of oxygen
as one gallon of MTBE, but gasoline containing ethanol
cannot be transported in pipelines because ethanol has
an affinity for water, which limits its use as a blending
component. From 1996 to 2010 ethanol for blending is
expected to remain at about 80,000 barrels per day in the
reference case. In the more stringent carbon reduction
cases, ethanol for blending is expected to be significantly
higher; in the less stringent cases, it is expected to be
slightly lower, because ethanol is more economically
attractive when the carbon price is higher.
Similar to the methanol oxygenate MTBE, ETBE (ethyl
tertiary butyl ether), an ethanol oxygenate made from a
combination of ethanol and isobutylene, is expected to
become profitable in the next few years. The advantage
of ETBE over straight ethanol is that it can easily be
blended with gasoline and shipped by pipeline. In 2010
in the reference case, ethanol for ETBE production is
30,000 barrels per day. In the more stringent carbon
reduction cases, ETBE production is expected to be
slightly higher; in the less stringent cases, it is expected
to be slightly lower, because ethanol is more economically attractive when the carbon price is higher.
To further complicate matters, over the next few years,
flexible fuel vehicles are expected to begin burning a significant amount of 85 percent ethanol fuel (E85), as a
result of legislative mandates under the Energy Policy
19901995200020052010201520200100200300400500600MillionBarrelsperCalendarDay1990+24%
1990-3%
Reference1990+9%
HistoryProjectionsFigure 101. Consumption of Ethanol in the
Transportation Sector, 1992-2020
Sources: History: Energy Information Administration, Renewable Energy
Annual 1997, DOE/EIA-0603(97) (Washington, DC, October 1997).
Projections: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.
D080398B, and FD03BLW.D080398B.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Act of 1992.74 Around 2005, vehicles capable of burning
only ethanol are projected to begin making a significant
impact on the ethanol market, because they are expected
to have one-third longer range and slightly higher gas
mileage than flex-fuel vehicles. From 1996 to 2010, E85
consumption is expected to grow from less than 2,000
barrels per day to about 70,000 barrels per day in all
cases, because E85 demand is expected to be driven primarily by legislative mandates. E85 demand is slightly
higher in the less stringent carbon reduction cases,
because the price of ethanol is attractive; demand is
slightly lower in the more stringent carbon reduction
cases because overall fuel demand is lower.
The sources of ethanol are also expected to change over
time. At present ethanol is primarily derived from fermentation of corn. However, ethanol can also be made
from cellulose biomass such as agricultural crop residuals, switchgrass, and other agricultural wood crops. In
this analysis cellulose ethanol production was allowed
to begin in 2001 at 1,300 barrels per day, based on current construction plans. From 2006 forward, capacity for
cellulose-based ethanol is allowed to grow annually at
10,000 barrels per day for the reference case and 16,000
barrels per day for the carbon reduction cases.
Ethanol produced from non-fossil fuels receives a Federal tax credit of 54 cents per gallon. This is equivalent to
5.4 cents per gallon on gasoline blended with 10 percent
ethanol. (The credit is prorated for blends of less than 10
percent and applies to the ethanol used to make ETBE.)
The tax exemption is scheduled to decline to 51 cents a
gallon from 2000 to 2007 and is allowed to remain at 51
cents through the rest of the forecast. Because this tax
credit is in nominal dollars, inflation eats away about
half its value in real terms by 2020. In the carbon reduction cases, a carbon price is not added to ethanol or the
ethanol part of ETBE, because ethanol is produced with
a non-fossil-fuel feedstock. Any carbon emitted from
burning ethanol is assumed to be recovered when new
crops are planted. To prevent ethanol from receiving
both a tax credit and an advantage from not suffering an
added carbon price, ethanol is allowed to receive the
greater of the two; in some cases from 2005 to 2007 the
tax credit is greater.
In the carbon reduction cases, ethanol consumption in
some years is lower than in the reference case (Figure
101), because the carbon price causes the cost of corn-
based ethanol to increase and not enough inexpensive
cellulose-based ethanol is yet available. One of the costs
of corn production is diesel fuel. When the cost of diesel
fuel goes up because of the added carbon price in 2005,
the cost of ethanol rises. Higher ethanol prices make
MTBE more attractive than ethanol as an oxygenate. In
addition, declining oil prices and lower oil demand
work to slow increases in the price of MTBE, which is
usually made entirely from fossil fuels. Significant quantities of cellulose-based ethanol do not become available
until after 2005. Significant new demand for ethanol
does not appear until after 2010, when the absence of an
added carbon price in ethanol makes ethanol much
more attractive as a feedstock for gasoline production.
(Appendix A has additional information on the ethanol
supply assumptions.)
Petroleum Product Prices
The projected prices of petroleum products in the carbon
reduction cases are substantially higher than those in the
reference case projections. For example, in 2010 the
transportation sector gasoline price is 54 cents a gallon
higher in the 1990-3% case than in the reference case
(Figure 102). Gasoline prices are higher in cases with
higher carbon prices and lower in cases with lower carbon prices, and the prices of other petroleum products
follow the same pattern. The primary components of
petroleum product prices are the crude oil price, refinery processing, Federal and State taxes, carbon prices,
and distribution costs.
In effect, carbon prices cause greater increases in the
prices of fuels that have higher carbon contents. In the
1990+24% case, the carbon price in 2010 adds 21 cents
per gallon to the price of residual fuel oil but only 9 cents
per gallon to the price of liquefied petroleum gas; the
corresponding price increases projected for gasoline, jet
fuel, and distillate fuel oil are 16, 17, and 19 cents per
gallon.
19901995200020052010201520200501001502001996CentsperGallon1990+24%
1990-3%
Reference1990+9%
HistoryProjectionsFigure 102. Gasoline Prices in the Transportation
Sector, 1990-2020
Sources: History: 1990-1995: Energy Information Administration (EIA),
Petroleum Marketing Annual 1995, web site www.eia.doe.gov/oil-gas/pmal/
pmaframe.html (May 30, 1997). 1996: EIA, Petroleum Marketing Monthly,
DOE/EIA-0380(96/03-97/04) (Washington, DC, 1996-97). Projections: Office of
Integrated Analysis and Forecasting, National Energy Modeling System runs
KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.D080398B, and
FD03BLW.D080398B.
74Public Law 102-486, Oct. 24, 1996, Title III, Section 303; Title V, Sections 501 and 507.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
World oil prices and demand-side effects moderate to
some extent the higher prices resulting from the carbon
price. Higher product prices lead to reduction in
demand in all the carbon reduction cases, which reduces
world oil prices. Thus, the world oil price and demand
effects combine to relieve some of the pressure on
product prices that results from carbon prices (Table 22).
The only product with a positive demand-side effect in
the carbon reduction cases relative to the reference case
is E85 in the 1990-3% and 1990-7% cases (Table 22).
Demand for ethanol grows more rapidly in the cases
with higher projected carbon prices, because there is no
carbon price added to ethanol-based products. (Because
ethanol is made from renewable plant material, carbon
emitted from burning ethanol is assumed to be recovered when new crops are planted.) In 2010, the
projected demand for ethanol is 70 percent higher in the
1990-3% case than in the reference case. With the
projected growth of demand for ethanol in the 1990-3%
case, increasing supplies of inexpensive biomass-based
ethanol are made available, reducing projected price
increases in 2010.
Regional petroleum product prices in the carbon reduction cases reflect many of the same market patterns
that exist today. In general, the Northeast and Pacific
regions continue to have the highest priced petroleum
products in the reference case and the carbon reduction
cases (Figure 103). Prices in these regions remain
relatively high because State tax rates are higher and
supplies are limited. Limited refining capacity in the
Northeast region increases reliance on imports and
supplies brought in from other regions. In contrast, the
Pacific region is isolated from outside sources of supply
by geography and by environmental restrictions.
Geographically separated from the rest of the Nation by
the Rocky Mountains, California must rely heavily on its
own refinery production. In addition, the State of
California has the most restrictive environmental
regulations on gasoline and diesel in the country, which
Table 22. Components of Differential Petroleum Product Prices Relative to the Reference Case, 2010
(1996 Dollars per Gallon)
Fuel
1990+24% 1990+9% 1990-3%
Demand
Reduction
Carbon
Price Total
Demand
Reduction
Carbon
Price Total
Demand
Reduction
Carbon
Price Total
Gasoline ........... -0.02 0.16 0.14 -0.08 0.38 0.30 -0.15 0.69 0.54
Distillate ........... -0.04 0.19 0.15 -0.05 0.42 0.37 -0.13 0.81 0.68
Jet Fuel............ -0.02 0.17 0.15 -0.07 0.41 0.34 -0.13 0.76 0.63
Residual Fuel ....... -0.02 0.21 0.19 -0.04 0.50 0.46 -0.08 0.93 0.85
LPG .............. -0.02 0.09 0.07 -0.08 0.23 0.15 -0.13 0.42 0.29
E85 ............... -0.02 0.03 0.01 -0.08 0.05 -0.03 0.07 0.11 0.18
World Oil Price ...... -0.02 Ñ Ñ -0.05 Ñ Ñ -0.07 Ñ Ñ
Source: Office of Integrated Analysis and Forecasting, National Energy Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.D080398B, and
FD03BLW.D080398B.
1990-3%1990+9%1990+24%Reference1996PacificWestSouthCentralEastSouthCentralSouthAtlanticMiddleAtlanticNewEnglandEastNorthCentralWestNorthCentralMountan13014416218513412614015618011912613915517912012113515117511712513915518012313414816418813112413915417812412413815417812112513854179132Figure 103. Retail Gasoline Prices by Region, Average of All Grades, 1996 and 2010
(1996 Cents per Gallon)
Sources: 1996: Energy Information Administration, Form EIA-782A, ÒRefinersÕ/Gas OperatorsÕ Monthly Petroleum Product Sales Report,Ó and Form EIA-782B,
ÒResellersÕ/RetailersÕ Monthly Petroleum Product Sales Report,Ó and volume-weighted taxes estimated by the Office of Integrated Analysis and Forecasting. Projections:
Office of Integrated Analysis and Forecasting, National Energy Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.D080398B, and
FD03BLW.D080398B.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
result in additional processing costs and further limit
CaliforniaÕs sources of supply.
Refinery Industry
Like all energy-intensive U.S. industries, the refinery
industry would be adversely affected by policies aimed
at reducing the consumption of carbon-based fuels. U.S.
refiners would bear the burden of reducing refinery
emissions of greenhouse gases, and at the same time
demand for their primary products would decline.
Lower demand for petroleum products is expected to
slow the growth of the U.S. refinery industry. In the reference case, the combined distillation capacity of U.S.
refineries is projected to be 16.9 million barrels per day
in 2010, with a utilization rate of 95 percent. In comparison, in the 1990+24% and 1990-3% cases, the projections
for distillation capacity in 2010 are 16.8 and 16.5 million
barrels per day, respectively, with utilization rates of 95
and 93 percent. From 2010 to 2020, distillation capacity
grows in the carbon reduction cases in response to
increasing petroleum consumption. U.S. refiners are not
expected to recover all the investments in new capacity
made before 2003 in the 1990-3% case, because consumption drops off between 2005 and 2015. Thus, utilization drops off particularly in 2009 in the 1990-3% case.
Reduced utilization rates and product consumption
may have an adverse impact on smaller or less competitive refineries that cannot develop ways to increase
product margins or market share. In the 1990+9% and
1990+24% cases, utilization remains close to 95 percent
throughout the forecast, and investment continues to be
recovered.
Refinery fuel consumption in the carbon reduction cases
drops in direct response to declines in product consumption and crude oil input. Total petroleum consumption at refineries in 2010 is projected to be 143 and
310 trillion Btu lower in the 1990+9% and 1990-3% cases
than in the reference case. By 2020, however, compared
to the reference case, total petroleum consumption at
refineries is higher in the 1990+9% case because residual
fuel oil replaces natural gas and is lower in the 1990-3%
case because total consumption is lower.
Consumption of natural gas at refineries in the carbon
reduction cases drops off after 2010, because gas is projected to be more expensive than petroleum. The higher
price for natural gas causes petroleum fuel consumption
to rise. Late in the forecast LPG and residual fuel consumed at refineries are higher in the more severe carbon
reduction cases than in the reference case, because still
gas production and consumption are lower as a result of
lower crude inputs to refineries, and because higher
natural gas prices result from the higher demand for
natural gas. Refinery processing gain also follows the
petroleum product consumption and domestic refinery
production of products, with processing gains 4 percent
and 11 percent lower in the 1990+24% and 1990-3%
cases, respectively, than in the reference case in 2010.
Petroleum product margins (wholesale price minus
crude costs), which indicate the amount of revenue
received by refineries per gallon, are lower in the carbon
reduction cases than in the reference case, in response to
lower product consumption (Figure 104). In the
1990+24% case, margins for gasoline, distillate, diesel,
and jet fuel in 2010 are 4 to 11 percent lower than in the
reference case, and in the 1990-3% case they are 26 to 30
percent lower. Between 2010 and 2020 the margins for
gasoline, distillate, and diesel remain about the same,
and those for jet fuel increase slightly in the carbon
reduction cases, because of shifts in demand.
1995200020052010201520200.000.050.100.150.200.250.300.351996DollarsperGallon1990+24%
1990-3%
Reference1990+9%
Figure 104. Projected Wholesale Gasoline Margins,
1996-2020
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.
D080398B, and FD03BLW.D080398B.
Refinery revenues also follow the product consumption
and product margins losses. Total projected refinery
revenues in the 1990+24% and 1990-3% cases are 5 and
24 percent lower in 2010 than they are in the reference
case, and revenues per barrel of product supplied are 3
and 14 percent lower. Total revenue losses associated
with the projected drop in world oil prices are 4 percent
and 14 percent in the 1990+24% and 1990-3% cases,
respectively, in 2010.
The projections of lower product margins, total revenues, and revenues per barrel of product supplied indicate that the U.S. refinery industry could face severe
constraints on profits and shareholder returns. Competitive pressures could force petroleum marketers to lower
prices while maintaining or improving product quality
in order grow market share. U.S. refineries may also face
competition from refiners in foreign countries that are
not parties to the Kyoto Protocol.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Coal
Background
Coal provides the largest fuel share, nearly 31 percent, of
U.S. domestic energy production. Electric utilities and
independent power producers generate more than 55
percent of all electricity via coal-fired technology and
account for approximately 89 percent of domestic coal
consumption. Steam coal is also consumed in the industrial sector to produce process heat, steam, and synthetic
gas and to cogenerate electricity, and metallurgical coal
is used to make coke for the iron and steel industry. With
more than 90 million tons75 of steam and metallurgical
coal shipped in 1996, coal is the only net energy fuel
export for the United States. In the reference case, coal
production and domestic consumption (expressed in
tons) are projected to increase at rates of 1.1 and 0.9
percent per year, respectively, and coal exports are
projected to increase somewhat more rapidly at a rate of
1.5 percent annually through 2020, primarily reflecting
the continued growth of steam coal consumption for
electricity generation in both domestic and overseas
markets.
The proposed limitations on carbon emissions will have
a significant negative impact on the coal industry. In the
carbon reduction cases analyzed here, the advantages of
the low carbon content of natural gas and the zero net
carbon emissions that are associated with renewables
offset the relatively low fuel cost of coal for use in electricity generation. Thus, coal markets are projected to be
severely affected, in terms of both overall sales and
supply patterns, as the need to reduce carbon emissions
results in significant shifts away from coal consumption
to natural gas, renewable energy, efficiency improvements in the demand sectors, andÑin some casesÑ
nuclear energy (see Chapter 4 for a discussion of
fuel switching and changes in electricity generating
capacity).
Carbon Emission Considerations
Coal, oil, and natural gas respond differently to restrictions on carbon emissions. Of the three, coal is most
affected for reasons that relate to the nature of its markets and its chemical structure. Electricity generation
markets, by far the largest market for coal, are increasingly competitive and cost-conscious as restructuring
initiatives by States have increasing influence on fuel
purchase strategies. Fossil fuels derive their energy content primarily from oxidation of their carbon and hydrogen contents. A fee based on carbon emissions from
burning fossil fuels (i.e., a carbon price) naturally falls
most heavily on coal, because coal derives a higher
percentage of its energy content from the oxidation of
carbon than do oil and natural gas.
Coal is heterogeneous in terms of both its energy content
and carbon content. Subbituminous coal derives a
higher proportion of its energy from carbon than does
bituminous coal; thus, production in the large low-
sulfur coalfields of the Northern Great Plains (Wyoming
and Montana) would be more affected by carbon emissions restrictions than would bituminous coalfields such
as those in Colorado and Utah, the Appalachian States,
and the Interior region. Lignite, which is produced primarily in Texas, North Dakota, and Louisiana, has more
carbon content than subbituminous coal, and its production would be more severely affected than that of bituminous or subbituminous coal in the carbon reduction
cases, in the absence of any offsetting factors such as
close proximity to customers.
Other factors that would affect the regional impacts of
carbon emission restrictions on different coalfields stem
from differences in mining and transportation costs.
Subbituminous coal production in the southern Powder
River Basin of Wyoming had an average mine price of
$6.41 per ton in 1996, as compared with bituminous
mine prices of $26.68 per ton in Appalachia, $21.43 in the
Interior, and $21.61 in the western States. However,
there is only a limited market for subbituminous coal in
the regions where it is mined. This coal has achieved
national importance in the past two decades because of
its low sulfur content and mining costs, giving it the ability to bear transportation costs of $20.00 per ton or more
while retaining economic competitiveness in markets on
the Atlantic, Pacific, Great Lakes, and Gulf coasts, up to
2,000 miles from its origin. A carbon price would create a
double penalty for such coal, first by penalizing the coal
for its inherent high ratio of carbon to energy content,
second by penalizing the carbon content in the transportation fuels that are required to bring it to market. Thus,
carbon emissions restrictions would most heavily penalize those coals most dependent on transportation to
reach their markets.
Coal Production
In the reference case, U.S. coal production climbs to
1,287 million tons in 2010 and 1,376 million tons in 2020
(Figure 105). In the carbon reduction cases, U.S. coal
production begins a slow decline early in the next
decade, accelerates rapidly downward through 2010,
and then continues to drop slowly through 2020. Coal
production in the 1990+24% case is 20 percent lower by
2010, at 1,032 million tons, in the 1990+9% case is 52
percent lower than reference case levels by 2010, at 624
million tons, and 71 percent lower in the 1990-3% case at
75In this section, physical quantities of coal are expressed in short tons, a unit of weight equal to 2,000 pounds. Carbon emissions are reported in metric tons, a unit of weight equal to 2,204.6 pounds.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
19701980199020002010202002004006008001,0001,2001,400MillionTons1990+24%
1990-3%
Reference1990+9%
1990+14%
19901990-7%
HistoryProjectionsFigure 105. U.S. Coal Production, 1970-2020
Sources: History: Energy Information Administration, Annual Energy Review
1997, DOE/EIA-0384(97) (Washington, DC, July 1998). Projections: Office of
Integrated Analysis and Forecasting, National Energy Modeling System runs
KYBASE.D080398A, FD24ABV.D080398B, FD1998.D080398B, FD09ABV.
D080398B, FD1990.D080398B, FD03BLW.D080398B, and FD07BLW.
D080398B.
369 million tons. By 2020, coal production in the
1990+24% case is 805 million tons and in the 1990+9%
case is 405 million tons, and production in the 1990-3%
case drops to a mere 172 million tons.
The projected declines in coal production result
primarily from sharp cutbacks in the use of steam coal
for electricity generation. Additional declines in production occur from reductions in the use of coal for
boiler fuel within the industrial sector, as a result of fuel
switching to natural gas. In 2010, coal consumption by
electricity generators in the 1990+24% case is 20 percent
lower than in the reference case, in the 1990+9% case is
57 percent lower, and in the 1990-3% case it is 79 percent
lower. Lower consumption results from a reduction (via
retirements) of in-place coal capacity, as well as lower
dispatch rates for coal-fired generation because the coal
capacity that remains available is used less intensively.
In 2010, coal-burning capability in the electricity supply
sector drops from 308 gigawatts in the reference case to
300 gigawatts (a 3-percent decline) in the 1990+24% case,
276 gigawatts (a 10-percent decline) in the 1990+9% case,
and 266 gigawatts (a 13-percent decline) in the 1990-3%
case. Utilization of existing coal capacity drops from 77
percent in the reference case to 65 percent in the
1990+24% case, to 40 percent in the 1990+9% case, and to
22 percent in the 1990-3% case.
In 2020, coal consumption by electricity generators is
projected to be 630 million tons in the 1990+24% case,
with coal-fired generating capacity at 271 gigawatts and
utilization at 55 percent, and only 235 million tons in the
1990+9% case, with coal capacity at 198 gigawatts and
utilization at 29 percent. In the 1990-3% case, increased
retirements of coal-fired plants result in coal capacity of
100 gigawatts (approximately one-third of reference
case levels), coal consumption for electricity generation
of 33 million tons, and a very low utilization rate of 9
percent. Operating and maintenance costs per unit of
electricity generated will increase for coal plants that are
run at low utilization because of thermal fatigue and the
inefficiencies of starting and stopping units that were
designed for baseload operation.
The expected reductions in coal exports and industrial
uses in the carbon reduction cases are somewhat less
severe than those in the electricity supply sector,
because not all coal-importing countries will be subject
to strict carbon caps, and because certain industrial consumers have less flexibility (because of plant configuration or fuel availability) to switch to lower carbon-
emitting fuels. As a result, coal production from regions
such as Central Appalachia that now serve this set of
customers declines somewhat less severely than that
from regions such as the Powder River Basin that have
the heaviest dependence on electricity producers. Coal
export projections are discussed later in this section.
Regional Coal Production Patterns
Reductions in coal consumption are expected to occur in
all regions and consuming sectors, but they will be of
different magnitudes and affect different coal types. As a
result, regional production patterns in the carbon reduction cases will shift differentially across regions relative
to the reference case, rather than on a basis that is strictly
proportional to national levels of coal consumption. In
the electricity generation sector, each reduction in overall coal generation will make it easier to achieve the
Clean Air Act Amendments sulfur dioxide (SO2) target
of 9 million tons of SO2, and in the more severe carbon
reduction cases, prices for the SO2 allowances will be
driven to zero. There will be upward pressure on coal
transportation rates, as a result of higher prices (from
carbon prices) on the diesel fuel used for rail, barge, and
truck transportation. At the same time, lower quantities
of coal shipments could place downward pressure on
transportation rates. The strong shift to greater use of
low-sulfur coal, particularly that mined in the West, in
the reference case will cease and reverse in consuming
regions where local mid-and high-sulfur coal can be
delivered at a lower cost than western coal.
The slower decline in coal consumption in the industrial,
metallurgical coal, and export sectors in the carbon
reduction cases will translate into relatively less severe
production cuts in regions that currently supply these
markets than the reductions in those regions that
depend more heavily on electricity generators. Nevertheless, there will be intensified intraregional competition to serve these important, albeit declining markets,
and some interregional shifts in production occur in the
forecast as regional demands shift.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
In the reference case, the western share of total U.S. coal
production increases from 47 percent in 1996 to 57
percent in 2010, as a result of its lower cost and the
growing requirements for low-sulfur coal under the
Clean Air Act Amendments (Figure 106). In contrast, the
western share in the carbon reduction cases decreases to
54 percent in the 1990+24% case, to 39 percent in the
1990+9% case, and to 28 percent in the 1990-3% case in
2010. Approximately 75 percent of the 179 million ton
reduction in western coal production in the 1990+24%
case, 486 million ton reduction in the 1990+9% case, and
the 628 million ton reduction in the 1990-3% case is
borne by subbituminous surface mines in the Powder
River Basin. The low-sulfur coal from these surface
mines is used almost exclusively for electricity
generation and must be transported over relatively long
distances to reach many of the markets that are projected
to expand in the reference case.
As overall demand falls, eastern minemouth prices are
reduced, and there is less economic incentive to
transport western coal. Western coal becomes less
competitive in electricity generation markets as transportation fuel costs increase, and its potential to expand
into most industrial and export applications is limited
by its lower heat content and other physical characteristics, such as moisture content and handling problems.
By 2020, western coal production has dropped by an
additional 189 million tons from 2010 levels in the
1990+24% case, 115 million tons in the 1990+9% case,
and by 71 million tons in the 1990-3% case, with western
production shares reaching 45, 32, and 19 percent,
respectively. In these cases, the limited coal that is
produced in the West is generally sold in markets close
to the point of production.
1990199620002005201020152020010203040506070PercentShareReference1990+24%1990+9%1990-3%
Figure 106. Western Share of U.S. Coal
Production, 1990-2020
Sources: History: Energy Information Administration, Annual Energy Review
1997, DOE/EIA-0384(97) (Washington, DC, July 1998). Projections: Office of
Integrated Analysis and Forecasting, National Energy Modeling System runs
KYBASE.D080398A, FD24ABV.D080398B, FD09ABV.D080398B, and
FD03BLW.D080398B.
Coal Prices
Because coal is heterogeneous in terms of heat content,
sulfur level, and other physical properties, trends in
national average prices are affected substantially by the
relative shares of the various coal types produced and
sold and by the units in which prices are reported. For
example, coal from the Powder River Basin is generally
the lowest-priced coal per ton on a minemouth basis;
however, because Powder River Basin coal has roughly
two-thirds the heat content of bituminous coal, its cost
advantage is somewhat less on a Btu basis and may be
nonexistent when delivered to more distant markets.
In general, to the extent that market share shifts away
from Powder River Basin coal, which has a low mine-
mouth price, to higher-priced bituminous coal, the
national average minemouth price will increase. Similarly, the greater the share represented by metallurgical
coal and by premium grades of coal for export use, the
higher will be the share-weighted average price. This
compositional effect offsets the reduction in minemouth
prices at the regional level that is likely to occur because
of intraregional competition and the lower production
quantities that occur when carbon restrictions take
effect. The regional productivity improvements projected in the reference case are assumed to occur at the
same rates in all the carbon reduction cases given the
same rate of technological progress. However, if the
level of investment in new capital equipment is severely
constrained, there could be adverse impacts on productivity.
In 2010, real minemouth prices are projected to decrease
to $14.29 per ton in the reference case but increase to
$14.72 in the 1990+24% case, to $16.42 in the 1990+9%
case, and to $17.90 in the 1990-3% case (Figure 107).
Minemouth prices in individual regions generally
decline in all cases, but the national average minemouth
price increases in the carbon reduction cases because of
the shift in quantity shares to higher grade and higher
priced coal and away from coal with a lower minemouth
price, such as that from the Powder River Basin. In some
instances, however, even the regional weighted average
price for a given coal rank will increase relative to the
reference case, if a greater share of coal is being shipped
to export or metallurgical markets that demand
premium-grade (and therefore higher priced) coals. The
pattern of higher national average prices in the carbon
reduction cases is accentuated by the projections for
2020, when prices increase from the reference case value
of $12.53 to $14.29 in the 1990+24% case, to $16.24 in the
1990+9% case, and to $19.63 in the 1990-3% case.
Delivered prices for coal, as projected in this report,
reflect the sum of the minemouth price, transportation
cost (in dollars per ton), and the carbon price associated
with meeting a carbon reduction target. The carbon
price dominates the effects on delivered prices in the
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
197019801990200020102020010203040501996DollarsperTonReference1990+24%
1990+14%
1990+9%
19901990-3%
1990-7%
HistoryProjectionsFigure 107. Average U.S. Minemouth Coal Prices,
1970-2020
Note: Carbon prices are added to the delivered price of coal, not to
the minemouth price.
Sources: History: Energy Information Administration, Annual Energy Review
1997, DOE/EIA-0384(97) (Washington, DC, July 1998). Projections: Office of
Integrated Analysis and Forecasting, National Energy Modeling System runs
KYBASE.D080398A, FD24ABV.D080398B, FD1998.D080398B, FD09ABV.
D080398B, FD1990.D080398B, FD03BLW.D080398B, and FD07BLW.
D080398B.
carbon reduction cases. In 2010, the carbon fee adds
$1.73 per million Btu to the delivered price of coal to
electricity generators in the 1990+24% case, $4.18 per
million Btu in the 1990+9% case, and $7.51 per million
Btu in the 1990-3% case. In 2020, the carbon price component drops to $2.55, $3.62, and $6.14 per million Btu,
respectively because the carbon price for all fuels is
lower in 2020.
In 2010, the national average delivered price of coal to
electricity generators increases from $22.20 per ton in the
reference case to $57.03 in the 1990+24% case, $109.56 in
the 1990+9% case, and $185.47 in the 1990-3% case (Figure 108). In 2020, the delivered price to electricity generators rises from $19.56 in the reference case, to $71.95
in the 1990+24% case, to $95.33 in the 1990+9% case, and
to $156.60 in the 1990-3% case.
Coal Industry Employment and
Productivity
Between 1978 and 1996, the number of miners employed
in the U.S. coal industry fell by 5.8 percent a year, declining from 246,000 to 83,000. The decrease primarily
reflected strong growth in labor productivity, which
increased at an annual rate of 6.7 percent over the same
period. An additional factor was increased output from
large surface mines in the Powder River Basin, which
require much less labor per ton of output than mines
located in the Interior and Appalachian regions. The
Powder River Basin share of total U.S. coal production
increased from 13 percent in 1978 to 30 percent in 1996.
1970198019902000201020200501001502002501996DollarsperTon1990+24%
1990-3%
Reference1990+9%
1990+14%
19901990-7%
HistoryProjectionsFigure 108. Coal Prices to Electricity Generators,
1970-2020
Sources: History: Energy Information Administration, Annual Energy Review
1997, DOE/EIA-0384(97) (Washington, DC, July 1998). Projections: Office of
Integrated Analysis and Forecasting, National Energy Modeling System runs
KYBASE.D080398A, FD24ABV.D080398B, FD1998.D080398B, FD09ABV.
D080398B, FD1990.D080398B, FD03BLW.D080398B, and FD07BLW.
D080398B.
In the reference case, productivity improvements are
assumed to continue but to decline in magnitude over
the forecast period. On a national basis, labor productivity increases at an average rate of 2.3 percent a year over
the whole forecast. The annual rate of increase slows,
however, from 5.8 percent in 1996 to approximately 1.6
percent per year from 2010 to 2020. With improvements
continuing over the forecast period, further declines in
employment of 1.3 and 1.1 percent per year are projected
from 1996 through 2010 and from 2010 through 2020,
respectively. In absolute terms, coal mine employment
declines from 83,000 in 1996 to 69,000 in 2010 and to
62,000 in 2020.
Regionally, labor productivity in the carbon reduction
cases is assumed to improve at the same rates as in the
reference case.76 As a result, lower levels of production
in the carbon reduction cases in all supply regions, relative to the reference case, result in lower employment
levels in all regions. Table 23 shows projections of coal
mining jobs in 2010 by region for the reference case and
the carbon reduction cases. In the 1990+24% case, coal
mine employment declines at a rate of 2.5 percent a year
between 1996 and 2010, falling from 83,000 in 1996 to
58,000 in 2010 (Figure 109). In the 1990+9% case, employment declines at a more rapid rate of 4.6 percent a year to
2010, resulting in employment of only 43,000 miners in
2010. In the 1990-3% case, coal mine employment
declines at a rate of 7.2 percent a year between 1996 and
2010, reaching 29,000 in 2010.
Production and employment are positively correlated.
In 2010, the projected levels of coal production in the
76Higher or lower rates of productivity growth could occur in the carbon reduction cases depending on the skill level and motivation of
the labor force in a rapidly contracting job market and the rate at which new capital equipment and technology are adopted.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Table 23. Projected Number of Coal Mining Jobs by Region, 2010
Region 1996 Reference 1990+24% 1990+14% 1990+9% 1990 1990-3% 1990-7%
Appalachiaa ......... 60,001 49,477 41,617 37,340 32,386 26,034 24,307 21,654
Interiorb ............. 13,477 8,043 7,801 7,617 6,257 4,315 3,484 2,663
Powder River Basinc . . 4,159 5,013 3,827 2,490 1,829 1,015 844 673
Other Westd ......... 5,825 5,693 4,785 2,859 2,254 1,034 941 895
U.S. Total.......... 83,462 68,519 58,223 50,224 42,531 32,053 29,187 25,486
aPA, OH, MD, WV, VA, and KY (east).
bIL, IN, KY (west), IA, MO, KS, AR, OK, TX, and LA.
cWY, MT, and ND.
dCO, UT, NM, AZ, AK, and WA.
Source: History: Energy Information Administration, Coal Industry Annual 1996, DOE/EIA-584(96) (Washington, DC, November 1997). Projections: Office of Inte-
grated Analysis and Forecasting, National Energy Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD1998.D080398B, FD09ABV.D080398B,
FD1990.D080398B, FD03BLW.D080398B, FD07BLW.D080398B.
197019801990200020102020050,000100,000150,000200,000250,000NumberofMiningJobsReference1990+24%
1990+14%
1990+9%
19901990-3%
1990-7%
HistoryProjectionsFigure 109. Coal Mine Employment, 1970-2020
Sources: History: Energy Information Administration (EIA), The U.S. Coal
Industry, 1970-1990: Two Decades of Change, DOE/EIA-0559, (Washington,
DC, November 1992) and EIA, Coal Industry Annual 1996, DOE/EIA-0584(96)
(Washington, DC, November 1997). Projections: Office of Integrated Analysis
and Forecasting, National Energy Modeling System runs KYBASE.D080398A,
FD24ABV.D080398B, FD1998.D080398B, FD09ABV.D080398B, FD1990.
D080398B, FD03BLW.D080398B, and FD07BLW.D080398B.
1990+24% and 1990-3% cases are 20 percent and 75
percent lower, respectively, than in the reference case. In
comparison, employment in the 1990+24% case is only
15 percent lower in 2010 than in the reference case and
employment in the 1990-3% case in 2010 is only 57
percent below the reference case. The projected declines
in employment are smaller than the declines in
production because of the relatively greater losses in
output projected from mines in the Northern Great
Plains, which require less labor per unit of output than
mines in other coal-producing regions.
Table 24 provides an indication of the importance of coal
industry jobs in the top coal-producing States. The table
shows that the wages associated with coal mining
exceeded 2 percent of all wages paid in 1996 in West
Virginia, Kentucky, and Wyoming. In West Virginia and
Wyoming, they accounted for more than 5 percent of all
wages paid. The fact that coal mining wages are higher
than average wages in these States is shown by the fact
that coal industry jobs account for a greater share of total
wages than their share of total employment. In West
Virginia, the coal industry employs 3.2 percent of all
workers in the State but accounts for 6.5 percent of all
wages paid. In Wyoming, coal industry workers account
for only 2.2 percent of all jobs but earn 5.3 percent of all
wages. Similarly, in Kentucky, the coal industry
provides 1.2 percent of all jobs but 2.1 percent of all
wages. Table 24 also shows that while the potential for
direct losses of coal-related wages and employment is
concentrated in the 10 States listed, it is much more
strongly concentrated in West Virginia, Kentucky,
Wyoming, and perhaps Pennsylvania (depending on
whether the absolute amount of wages and employment
at stake is counted, or the relative proportion of the
StateÕs total wages and employment).
In addition to the substantial contraction of the U.S. coal
industry projected in the carbon reduction cases, the
U.S. rail industry, which derives considerable revenues
from coal shipments, also stands to be greatly affected
(see box).
U.S. Coal Exports
U.S. coal producers exported 90 million tons of coal in
1996. Of that amount, 59 percent represented shipments
of coking coal for use at integrated steel plants worldwide, and 41 percent was steam coal, used primarily for
electricity generation and for the production of process
steam and direct heat for industrial applications. In 1997,
U.S. coal exports fell by 7 million tons, reversing the
upward trend of the previous 2 years. The decline was
mostly in steam coal exports, as a result of weak international coal prices and strong competition from other
coal-exporting countries.
In the reference case, U.S. coal exports are projected
to increase from 90 million tons in 1996 to 113 million
tons in 2010. All the increase reflects expected growth
in steam coal exports, with exports of metallurgical coal
projected to decline slightly. In the reference case, world
metallurgical coal trade remains relatively constant,
although regionally there is a slight shift away
from markets in Europe and Japan to Brazil and the
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Table 24. Coal Industry Wages and Employment, 1996
State
Wages Employmenta
Million 1996 Dollars Percent of State Total Number of Jobs Percent of State Total
West Virginia ................ 1,041 6.53 21,033 3.17
Kentucky.................... 815 2.06 19,372 1.20
Wyoming.................... 258 5.29 4,706 2.20
Pennsylvania ................ 512 0.34 11,214 0.22
Illinois ...................... 347 0.20 6,136 0.11
Virginia ..................... 290 0.03 7,039 0.02
Alabama .................... 332 0.74 6,552 0.04
Ohio ....................... 172 0.12 3,889 0.01
Texas ...................... 149 0.65 2,861 <0.01
Montana .................... 49 0.66 933 0.03
Subtotal ................... 3,965 0.42 83,375 0.03
United States................ 4,691 0.17 97,649 0.08
aRelative to Form EIA-7A, ÒCoal Production Report,Ó which focuses on workers directly involved in the production and preparation of coal, the data
presented in this table include coverage of corporate officials, executives, clerical workers, and other office workers. Data from Form EIA-7A indicate
that 83,462 miners were employed in the U.S. coal industry in 1996.
Source: U.S. Department of Labor, Bureau of Labor Statistics, ES-202 Program, ÒCovered Employment and Wages.Ó
developing countries in Asia. World steam coal trade is
projected to increase by 45 percent between 1996 and
2010, rising from 305 million tons to 441 million tons.
The U.S. share of total world coal trade is projected to
remain constant at about 18 percent.
In the reference case, JapanÕs remaining two coal mines
are assumed to be closed shortly after 2000. Currently
these mines have a combined annual production capacity of about 3.5 million tons, representing less than 3 percent of JapanÕs total coal consumption. In 1996, coal
consumption in Japan amounted to 144 million tonsÑ82
million tons of steam coal (including 9.5 million tons of
coal for pulverized coal injection at blast furnaces) and
62 million tons of coking coal.
In the carbon reduction cases, two alternative coal trade
scenarios were developed. In a severe carbon reduction
case (1990-3%), carbon emissions in Western Europe
were assumed to be 8 percent below their 1990 level by
2010 consistent with the limits for the European Union
that were specified in the Kyoto Protocol. Similarly, carbon emissions in Japan were assumed to be 6 percent
below their 1990 level by 2010. Coal was assumed to play
a proportionately greater role than oil or natural gas in
meeting these emission reductions, because it has a
higher carbon content (on a Btu basis) and the opportunities to substitute for petroleum products in the transportation sector are limited. In Western Europe, both
domestic coal production and imports were assumed to
decline by approximately 50 percent, but in Japan coal
imports had to account for the total reduction in coal
consumption.
In Europe, steam coal imports from all sources are
reduced from 156 million tons in the reference case in
2010 to 47 million tons in the 1990-3% case. Only steam
coal imports to the industrialized Annex I countries in
Europe are reduced. Steam coal imports to Japan, the
only Annex I country in Asia, are reduced from 99 million tons in the reference case in 2010 to 56 million tons in
the 1990-3% case. Because other fuels are not easily substituted for coal coke at steel plants, coking coal imports
are not adjusted downward.
Steam coal imports to Japan are reduced by a relatively
smaller amount than are imports to Europe, primarily
because Japan has limited access to alternative sources
of energy such as natural gas and renewable fuels. In
addition to reduced use of coal, other strategies that
Japan may pursue to meet its carbon reduction targets
include purchasing surplus emission allowances from
other signatory countries and pursuing an accelerated
nuclear program.77
U.S. coal exports to Europe and Asia in 2010 are projected to be lower by 27 and 7 million tons, respectively,
in the 1990-3% case (and all other carbon reduction cases
where U.S. carbon emissions are held at or below the
1990 level in 2010) than in the reference case. In these
cases, U.S. coal exports are projected to decline to 76 million tons in 2010.
In the moderate cases, 1990+24% and 1990+9%, developed to evaluate the potential impacts of less severe
reductions in carbon emissions, Western European coal
consumption and imports were assumed to decline by a
smaller amount than in the severe case discussed above,
77In June 1998, a panel headed by then Prime Minister Ryutaro Hashimoto urged the government to construct an additional 20 new nuclear plants over the next 12 years, with the goal of increasing JapanÕs nuclear generation by more than 50 percent between 1997 and 2010.
EIAÕs International Energy Outlook (IEO98) high nuclear case projects an increase of 12.4 gigawatts (29 percent) in JapanÕs nuclear generating
capacity over the same period. The IEO98 reference case projects an increase of only 5.2 gigawatts (12 percent) between 1996 and 2010.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
reflecting the lower emission target. Japanese coal consumption and imports were also assumed to decline by a
smaller amount as in the severe case. In Europe, projected steam coal imports from all sources are reduced
from 156 million tons in the reference case in 2010 to 96
million tons in the 1990+9% case. Only steam coal
imports to the industrialized Annex I countries in
Europe are reduced.
U.S. coal exports to Europe and Asia in 2010 are
projected to be lower by 17 and 4 million tons,
respectively, in the 1990+24% and 1990+9% cases (and in
all other carbon reduction cases where U.S. carbon
emissions are above the 1990 level in 2010) than in the
reference case. In these cases, U.S. coal exports of 89
million tons are projected for 2010, as compared with
113 million tons in the reference case.
Impacts on the Rail Industry
In 1996, 705 million of the 1,064 million tons of coal produced in the United States (66 percent) was transported to
consumers partly or entirely by rail. Coal freight provided Class I railroads with $7.7 billion, 23 percent of all
revenue earned. Coal freight car loadings and ton-miles
tend to be dominated by a handful of railroads. For the
major coal-hauling railroads, coal represented 39 percent
of all car loadings during 1996.a Available data from the
Federal Railroad Administration that summarize railroadsÕ reported return on investment and the extent of
their dependence on coal freight revenues are shown in
the table below.
Because the carbon reduction cases analyzed here project
heavier losses in coal production for western than for
eastern coalfields, and because much of the production
from western coalfields is shipped long distances into
midwestern and eastern markets to satisfy demand for
low-sulfur fuel, it is likely that the burden of reduced coal
transportation revenues would fall most heavily on railroads in the WestÑparticularly on the Burlington-
Northern and Union Pacific systems, which now include
the St. Louis Southwestern, the Chicago & Northwestern,
the Denver & Rio Grande Western, the Southern Pacific,
and the Atchison, Topeka & Santa Fe railroads.
Progressively deregulated since the Staggers Rail Act of
1986, railroads have made substantial progress in
improving productivity and reducing real costs by investing in new and more powerful locomotives, improved
maintenance of main-line rights of way, and more efficient use of labor. A major contribution to achieving the
joint goals of lower costs and maintenance of service has
been made through a number of mergers over the past
decade. Mergers have resulted in the emergence of four
major railroad companiesÑtwo in the East (CSX and
Norfolk-Southern) and two in the West (Burlington
Northern -Santa Fe and Union Pacific -Southern Pacific).
The recent merger between Union Pacific and Southern
Pacific was followed by a period of service problems (particularly in Texas, but also affecting rail shipments
throughout the Union Pacific -Southern Pacific system)
that have not yet been entirely resolved. As a result of
these service issues, there has been controversy surrounding the policies of the Surface Transportation Board as it
has sought to balance the needs of railroad shippers and
Revenue Adequacy and Relative Dependence on Coal Revenue by Railroad, 1989-1995
Railroad
1989 1991 1993 1995
Percent of
Total
Revenue
From Coal
Rate of
Return on
Investment
Percent of
Total
Revenue
From Coal
Rate of
Return on
Investment
Percent of
Total
Revenue
From Coal
Rate of
Return on
Investment
Percent of
Total
Revenue
From Coal
Rate of
Return on
Investment
Eastern District
Conrail ................... 15.4 2.6 16.8 NM 14.2 6.5 15.9 6.8
CSX ..................... 34.4 6.1 35.3 NM 29.9 0.1 29.8 6.5
FloridaEastCoast.......... 1.1 10.3 1.0 2.2 NA NA NA NA
Grand Trunk Western........ 8.1 1.9 9.4 NM 8.2 NM 7.9 NM
Illinois Central.............. 16.0 11.2 15.2 15.2 12.7 14.7 13.9 17.2
Norfolk Southern............ 36.1 11.9 37.0 6.0 32.9 12.1 30.9 12.1
Western District
Atchison Topeka & Santa Fe . . 7.7 NM 8.9 6.5 8.7 1.9 7.3 5.3
Burlington Northern ......... 33.0 12.5 33.5 NM 31.9 9.4 32.7 6.3
Chicago & Northwestern...... 12.4 8.2 14.1 7.1 13.5 10.3 15.5 NA
Kansas City Southern........ 33.7 10.7 31.9 9.3 29.9 9.0 19.7 7.9
SooLine.................. 11.3 NM 12.8 4.0 9.2 NM 3.8 NM
SouthernPacific............ 2.2 1.8 2.4 NM 3.2 3.5 9.4
1.3
St. Louis Southwestern....... 2.6 1.8 2.2 NM 3.2 3.5 9.4
1.3
UnionPacific............... 16.1 10.4 17.5 1.7 16.8 11.1 19.0 11.7
NM = negative returns on investment are described only as Ònot meaningfulÓ in the source.
NA = not available, usually because the railroad has ceased to operate as an independent entity.
Source: Federal Railroad Administration.
(Continued on page 117)
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Impacts on the Rail Industry (Continued)
the continued profitability of the Union Pacific, and of the reduction measures could create significant financial
NationÕs major railroads in general. Even if these issues problems for those firms. Lignite production in Texas,
are successfully resolved over the next few years, the Louisiana, and North Dakota would also be severely
adoption of carbon emissions restrictions would inevita-reduced by carbon emissions restrictions, but the effect on
bly result in a reduction in domestic coal traffic handled rail revenues would be minor. Because of its inherently
by the railroads. low heat content, lignite is predominantly consumed at or
As suggested by the results of the carbon reduction cases, close to the place of mining.
the reductions in coal traffic range from moderate to Although the projected losses of coal production in the
severe, depending on the case. In all cases, western coal, individual carbon reduction cases are proportionately
particularly subbituminous coal from the Powder River and absolutely less for Appalachian coalfields than for the
Basin, would be most severely restricted, because of its Powder River Basin, the two eastern rail systems (CSX
dependence on long-distance rail transportation to reach and Norfolk Southern) are also highly dependent on coal
its markets in locations up to 2,000 miles away and its revenue. In the more severe carbon reduction cases,
high ratio of carbon to energy content. As shown in the Appalachian coal production could be reduced by one-
table, the Burlington Northern and Union Pacific systems third to one-half, with potentially serious financial consehave a fairly high dependence on coal freight revenue; quences for these carriers.
therefore, the loss of revenue associated with carbon
aAssociation of American Railroads, Freight Commodity Statistics.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
6. Assessment of Economic Impacts
Objectives of the
Macroeconomic Analysis
Because energy resources are used to produce most
goods and services, higher energy prices can affect the
economyÕs production potential. Since the energy crisis
of the 1970s, economic research has led to a better understanding of the potential adverse economic consequences of rising real energy costs, in terms of both long-
run equilibrium costs and short-run adjustment costs.
Long-run equilibrium costs are associated with reducing reliance on energy in favor of other factors of productionÑincluding labor and capital, which become
relatively cheaper as energy costs rise. Short-run adjustment costs, or business cycle costs, can arise when price
increases disrupt capital or employment markets. Long-
run costs are considered unavoidable. Short-run costs
might be avoidable if price changes can be accurately
anticipated or if appropriate compensatory monetary
and fiscal policies can be implemented.
This chapter assesses possible impacts on the economy
associated with attaining the alternative carbon mitigation targets presented earlier in this report, focusing on
three target casesÑthe 3-percent-below-1990 (1990-3%),
the 9-percent-above-1990 (1990+9%), and the 24percent-above-1990 (1990+24%) casesÑand comparing
them with a reference case that does not include the
Kyoto Protocol. In evaluating these alternative targets,
three key questions are posed:
¥ What would be the unavoidable minimum impact
on the economy?
¥ With rising energy prices and inflation, what cyclical
reactions could the economy face, and how would
the Federal Reserve Board implement accommodating monetary policy?
¥ What would be the impact of fiscal policy on economic output and inflation?
78The version of the model used is US97A95.
EIA used the Data Resources, Inc. (DRI) model of the
U.S. economy to assess these issues.78 The DRI model is
a representation of the U.S. economy with detailed
output, price, and financial sectors incorporating both
long-term and short-term properties. In the DRI model,
the concept of potential GDP reflects the trajectory of the
long-term growth potential of the economy at full
employment, while actual GDP is a measure of the transition effects as the economy adjusts to its long-run path.
Energy end-use demands and prices for fuels are the key
energy inputs to the DRI model.79 In addition, for this
analysis, assumptions were made about the domestic
flow of funds that would result from a U.S. system of
carbon permits sold by the Federal Government, and
about the international flow of funds that would result
from international trading of permits. These assumptions were based on the results of the energy market
analyses described in the preceding chapters of this
report.
This chapter first presents a discussion of the U.S. permit
system and the potential role of international trading of
permits. A summary of the macroeconomic effects is
presented next, focusing on the definition and measurement of potential GDP, actual GDP, and the value of the
purchased international permits as key elements. The chapter then discusses in detail two topics. The first
addresses the unavoidable loss to the economy that
would result from a reduction in available energy
resources. The unavoidable loss has two components:
the loss in potential GDP and the value of the purchased
international permits. The chapter concludes with a discussion of the possible transitional impacts on the aggregate economy that might occur as energy prices increase
in response to carbon emission constraints. The critical
roles of monetary and fiscal policy are highlighted. Two
fiscal policies are considered as alternative methods of
returning carbon permit revenues to the economy:
through a lump sum personal income tax rebate and
through a social security tax rebate that would pass
funds back to both employers and employees.
79This macroeconomic analysis of the costs of implementing the Kyoto Protocol is limited to the consideration of investment costs that
are comparable in magnitude to those in the reference case, as well as direct fuel costs. No consideration is given to the potential incremental
costs of investment in technology and infrastructure that would be necessary in each of the specific cases analyzed. Business investments
above reference case levels may be required to reduce energy costs in response to increasing energy prices.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
The U.S. Permit System and
International Trading of Permits
Two key features shape the discussion in this chapterÑ
first, the characterization of the carbon permit trading
system as an auction run by the Federal Government;
and second, the international trading of carbon permits.
Both of these issues have important implications for the
assessment of the potential macroeconomic impacts of
carbon mitigation policies.
The U.S. Permit System
When a system is developed for the trading of carbon
permits within the United States, a number of initial
decisions must be made: How many permits will be
available? Will they be freely allocated or sold by competitive auction? If they are allocated, how will the initial
allocations be made? If they are sold, what will be done
with the revenues? How many permits will be bought in
international markets? If the permits are traded in a free
market, holders of permits who can reduce carbon emissions at a cost below the permit price will sell their permits, and those with higher costs of reduction will buy
permits, resulting in a transfer of funds between private
parties. If the permits are sold by competitive auction,
there will be a transfer of funds from emitters of carbon
to the Federal treasury.
This analysis makes the explicit assumption that carbon
permits will be sold in a competitive auction run by the
Federal Government.80 To illustrate the importance of
recycling the funds back to the economy, two fiscal policy approaches are considered: first, returning collected
revenues to consumer through personal income tax
rebates and, second, lowering the social security tax rate
as it applies to both employers and employees. The two
policies are meant only to be representative of a set of
possible fiscal policies that might accompany an initial
carbon mitigation policy.
International Trading of Permits
In the energy market assessments described earlier in
this report, the projected carbon prices reflect the price
the United States would be willing to pay to achieve a
given emissions reduction target. The more stringent the
carbon target, the higher the carbon price. The energy
market analysis in this report does not address the international implications of achieving a particular target at
the projected carbon price. In the absence of modeling
international trade of emissions permits, the energy
market assessment makes no link between the U.S. carbon price and the international market-clearing price of
permits, or the price at which other countries would be
willing to offer permits for sale in the United States.
The macroeconomic analysis in this chapter departs
from the above interpretation in order to facilitate an
evaluation of the role of the purchase of permits in an
international market. The analysis first assumes that the
U.S. State DepartmentÕs assessment of the accounting of
carbon-absorbing sinks and offsets from reductions in
other greenhouse gases will reduce the binding U.S.
emissions target to 3 percent below the 1990 level of
emissions. Then, if the United States is to meet a target
that is less stringent, the difference in emissions is
assumed to be made up through the purchase of permits
on the international market. Moreover, the United States
is assumed to purchase international permits at the marginal abatement cost in the United States. Thus, the domestic carbon price would be the same as the international
permit price under the alternative targets considered. If
unrestricted international trading among Annex I countries is allowed, the international carbon price could fall
below the levels projected here for domestic permits. If
this were to occur, to achieve equilibrium in an unconstrained market for carbon permits, the domestic carbon
price would fall to the international carbon price.
The above assumptions imply that different international supplies of permits would be available in the
alternative cases considered. This is an important
simplifying assumption, and the value placed on the
overseas transfer of funds to purchase international
permits is subject to considerable uncertainty. However,
this element must be considered a key factor in
performing any assessment of the impacts on the
economy, and therefore it is explicitly factored into the
analysis. Table 25 shows the assumed carbon
reductions, carbon prices, and number and value of
carbon emission permits purchased on the international
market in the 1990-3%, 1990+9%, and 1990+24% cases.
Summary of
Macroeconomic Impacts
In the long run, higher energy costs would reduce the
use of energy by shifting production toward less energy-
intensive sectors, by replacing energy with labor and
80A permit auction system is identical to a carbon tax as long as the marginal abatement reduction cost is known with certainty by the
Federal Government. If the target reduction is specified, as in this analysis, then there is one true price, which represents the marginal cost of
abatement, and this also becomes the appropriate tax rate. In the face of uncertainty, however, the actual tax rate applied may over-or undershoot the carbon reduction target. Auctioning of the permits by the Federal Government is evaluated in this report. The costs of administering the program are not considered. To investigate a system of allocated permits would require an energy and macroeconomic modeling
structure with a highly detailed sectoral breakout beyond those represented in the NEMS and DRI models. For a comparison of emissions
taxes and marketable permit systems, see R. Perman, Y. Ma, and J. McGilvray, Natural Resources and Environmental Economics (New York,
NY: Longman Publishing, 1996), pp. 231-233.)
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Table 25. Energy Market Assumptions for the Macroeconomic Analysis of Three Carbon Reduction Cases,
Average Annual Values, 2008 through 2012
Analysis Case
Binding Carbon
Emissions
Reduction Target
(Million Metric
Tons)
Average U.S.
Carbon Emissions
Reductions
(Million Metric
Tons)
U.S. Purchases
of International
Permits (Million
Metric Tons)
Carbon Price
Value of
Purchased
International
Permits (Billion
1992 Dollars)
1996 Dollars
per Metric Ton
1992 Dollars
per Metric Ton
1990-3% ....... 485 485 0 290 263 0
1990+9%....... 485 325 160 159 144
23
1990+24%...... 485 122 363 65 59
21
Source: Office of Integrated Analysis and Forecasting, National Energy Modeling System, runs FD24ABV.D080398B, FD09ABV.D080398B, and FD03BLW.
D080398B.
capital in specific production processes, and by encouraging energy conservation. Although reflecting a more
efficient use of higher-cost energy, this gradual reduction in energy use would tend to lower the productivity
of other factors in the production process. The derivation of the long-run equilibrium path of the economy can
be characterized as representing the ÒpotentialÓ output
of the economy when all resourcesÑlabor, capital, and
energyÑare fully employed. As such, potential gross
domestic product (GDP) in the DRI model is equivalent
to the full employment concept calculated in a number
of other models that focus on long-run growth while
abstracting from business cycle behavior.81
The ultimate impacts of carbon mitigation policies on
the economy will be determined by complex interactions between elements of aggregate supply and
demand, in conjunction with monetary and fiscal policy
decisions. As such, cyclical impacts on the economy are
bound to be characterized by uncertainty, possibly significant. Raising energy prices and, as a result, downstream prices in the rest of the economy could introduce
cyclical behavior in the economy, resulting in employment and output losses in the short run. The measurement of losses in actual output for the economy, or
actual GDP, incorporates the transitional cost to the
aggregate economy as it adjusts to its long-run path.
Resources may be less than fully employed, and the
economy may move in a cyclical fashion as the initial
cause of the disturbanceÑthe increase in energy
pricesÑplays out over time.
The possible impacts on the economy are summarized in
Table 26, which shows average changes from the
reference case projections over the period from 2008
through 2012 in the three carbon reduction analysis
cases.82 The loss of potential GDP measures the loss in productive capacity of the economy directly attributable to
the reduction in energy resources available to the economy. It represents part of the long-run, unavoidable
impact on the economy. The macroeconomic adjustment
cost reflects frictions in the economy that may result
from the higher prices of the carbon mitigation policy. It
recognizes the possibility that cyclical adjustments may
occur in the short run. The loss in actual GDP for the economy is the sum of the loss in potential and the adjustment cost. The purchase of international permits represents
a claim on the productive capacity of domestic U.S.
resources. Essentially, as funds flow abroad, other countries have an increased claim on U.S. goods and services.
The total cost to the economy is represented by the loss in
actual GDP plus the purchase of international permits
(Figure 110). These costs need to be put in perspective
relative to the size of the economy, which is projected to
average $9,425 billion between 2008 and 2012 in the
reference case.
Another way to view the macroeconomic effects is by
looking at the effects of the carbon reduction cases on the
growth rate of the economy, both during the period of
implementation and during the early part of the commitment period, from 2005 through 2010, and then over
the entire period from 2005 through 2020 (Figures 111
and 112). In all instances, the economy continues to
grow, but growth is slower than projected in the reference case. In the reference case, potential and actual
GDP grow at 2.0 percent per year from 2005 through
2010. In the 1990+9% case, the growth rate in potential
GDP slows to 1.9 percent per year, and the growth rate
in actual GDP slows to 1.6 percent per year when the
81In the DRI model, the aggregate production function (the potential GDP equation) uses the following concepts as important variables:
energy, labor, capital stocks of equipment and structures, and research and development expenditures. The aggregate supply is estimated
by a Cobb-Douglas production function that combines factor input growth and improvements in total factor productivity. Factor input
equals a weighted average of energy, labor, fixed capital (outside the energy-producing sector), and public infrastructure. Factor supplies
for the non-energy sector are defined by estimates of the full-employment labor force, the full-employment capital net of pollution abatement equipment, domestic energy consumption, and the stock of infrastructure. Total factor productivity depends on the stock of research
and development capital and a technological change trend.
82The output measures presented in this chapter are expressed in constant 1992 chain-weighted dollars. The DRI macroeconomic model
uses National Income and Products Accounts (NIPA) as an estimating framework. Expressing these output measures in 1992 dollars maintains consistency with the NIPA framework and facilitates comparison with results from other macroeconomic models. For the purposes of
recycling the funds, collections and rebates are expressed in nominal dollars, to be consistent with the Federal GovernmentÕs tax accounting
system.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Table 26. Macroeconomic Impacts in Three Carbon Reduction Cases, Average Annual Values, 2008-2012
(Billion 1992 Dollars)
Analysis Case
Loss in
Potential GDP
Macroeconomic
Adjustment Cost
Loss in
Actual GDP
Purchases of
International
Permits
Total Cost
to the Economy
1990-3%
Personal Income Tax Rebate....... 58 225 283 0 283
Social Security Tax Rebate ........ 58 70 128 0 128
1990+9%
Personal Income Tax Rebate....... 32 137 169 23 192
Social Security Tax Rebate ........ 32 59 91 23 114
1990+24%
Personal Income Tax Rebate....... 12 76 88 21 109
Social Security Tax Rebate ........ 12 44 56 21 77
Note: Loss in potential GDP plus the macroeconomic adjustment costs equals the loss in actual GDP. The actual GDP loss plus purchases of inter-
national permits equals the total cost to the economy.
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model of the U.S. Economy.
050100150200250300Billion1992DollarsMacroeconomicAdjustmentCostLossinPotentialGDPPurchaseofInternationalPermitsAssumingPersonalncomeTaxRebate1990+9%1990-3%1990+24%
AssumingSocialSecurityTaxRebateFigure 110. Projected Annual Costs of Carbon
Reductions to the U.S. Economy,
2008-2012
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model
of the U.S. Economy.
PotentialGDPActualGDP,
AssumingPersonalIncomeTaxRebateActualGDP,
AssumingSocialSecurityTaxRebate0.00.51.01.52.02.5PercentperYearReference1990+24%1990+9%1990-3%
Figure 111. Projected Annual Growth Rates in
Potential and Actual GDP, 2005-2010
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model
of the U.S. Economy.
PotentialGDPActualGDP,
AssumingPersonalIncomeTaxRebateActualGDP,
AssumingSocialSecurityTaxRebate0.00.51.01.52.02.5PercentperYearReference1990+24%1990+9%1990-3%
Figure 112. Projected Annual Growth Rates in
Potential and Actual GDP, 2005-2020
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model
of the U.S. Economy.
personal income tax rebate is assumed or 1.8 percent per
year when the social security tax rebate is assumed.
However, through 2020, with the economy rebounding
back to the reference case path, there is no appreciable
change in the projected long-term growth rate. The
results for the 1990+24% and 1990-3% cases are similar.
Aggregate impacts on the economy, as measured by
actual GDP, are shown in Table 27 in terms of losses in
actual GDP per capita. In the 1990+9% case, the loss in
potential GDP per capita is $106; however, the loss in
actual GDP for in the 1990+9% case is $567 assuming the
personal income tax rebate and $305 assuming the social
security tax rebate. Again, the lower value (loss in
potential GDP) represents part of the unavoidable loss
per person, and the higher values (loss in actual GDP)
reflect the highly uncertain, but significant, impacts that
individuals could experience as the result of frictions
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Table 27. Projected Losses in Potential and Actual GDP per Capita, Average Annual Values, 2008-2012
(1992 Dollars per Person)
Analysis Case
Loss in Potential GDP
per Capita
Loss in Actual GDP per Capita,
Personal Income Tax Rebate
Loss in Actual GDP per Capita,
Social Security Tax Rebate
1990-3% .................... 193 947 428
1990+9% ................... 106 567 305
1990+24 .................... 40 294 187
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model of the U.S. Economy.
within the economy. Again, to provide scale, actual GDP price changes. As shown in Table 25, the three cases
per capita averages $31,528 in the reference case from considered in this chapter reduce U.S. carbon emissions
2008 through 2012. by 122, 325, and 485 million metric tons a year on
average between 2008 and 2012. Figure 114 shows the
relationship between the projections of carbon emission
Estimating The Unavoidable Impact reductions and carbon prices. When the carbon
on the Economy reduction target is more stringent, the carbon price is
higher; and for the most stringent targets, the projected
Figure 113 shows the losses in the potential economic carbon prices are disproportionately higher than those
output, as measured by potential GDP, for the three in the less stringent cases (i.e., the relationship is
carbon reduction cases. The shapes of the three nonlinear). This curve can be used to measure losses to
trajectories mirror the carbon price trajectories. In the the aggregate economy by calculating the integral under
1990-3% case, potential GDP declines relative to the the curve up to the level of the specified target case.
reference case from 2005 through 2008, reaching a Results for the 1990-3%, 1990+9%, and 1990+24% cases
maximum loss of $64 billion (in 1992 dollars) in 2012 and are shown in Table 28.
then leveling off at just under $60 billion a year through
2020. In the 1990+9% case, the loss in potential GDP The 1990+9% case results in an average reduction in
declines to $35 billion by 2011 and reaches $39 billion in carbon emissions of 325 million metric tons per year
2020. In the 1990+24% case, with steadily increasing during the period from 2008 to 2012. The average carbon
carbon prices, potential GDP declines relative to the price projected for the same period is $144 per metric ton
reference case projections throughout the period and is (in 1992 dollars) (Table 25). The triangular area under
$26 billion lower than the reference case levels in 2020. the curve in Figure 114, labeled A, represents the value
of the carbon reduction to the economyÑi.e., the value
These three potential GDP trajectories represent a of reduction in economic output that would result from
valuation of the possible loss in output in the economy higher energy prices. In the 1990+9% case, the economic
in the absence of any cyclical influences brought on by loss projected by the NEMS model totals $25 billion
(Table 28). In comparison, the loss in potential GDP
199520002005201020152020010-10-20-30-40-50-60-70Billion1992Dollars1990+9%
1990-3%
1990+24%
ReferenceCasePotentialGDP:
1996=$6,9302010=$9,4822020=$10,994Figure 113. Projected Dollar Losses in Potential
GDP Relative to the Reference Case,
1998-2020
Note: Carbon permit revenues are assumed to be returned to
households through personal income tax rebates.
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model
of the U.S. Economy.
0100200300400500600AverageCarbonReductions(MillionMetricTons)
050100150200250300350AverageCarbonPrice(1992DollarsperMetricTon)
1990+9%
1990-3%
1990+24%
1990+14%
19901990-7%
ReferenceABFigure 114. Average Carbon Reductions and
Projected Carbon Prices, 2008-2012
Source: Office of Integrated Analysis and Forecasting, National Energy
Modeling System runs KYBASE.D080398A, FD24ABV.D080398B, FD1998.
D080398B, FD09ABV.D080398B, FD1990.D080398B, FD03BLW.D080398B,
and FD07BLW.D080398B.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Table 28. Average Projected Annual Losses in Economic Output, 2008-2012
Analysis Case
Value of Lost Output U.S. Purchase of
International Permits
(Billion 1992 Dollars)
NEMS Valuation
(Billion 1992 Dollars)
DRI Potential GDP Loss
(Billion 1992 Dollars)
1990-3% .................... 57 58 0
1990+9% ................... 25 32 23
1990+24% .................. 4 12 21
Sources: Office of Integrated Analysis and Forecasting, National Energy Modeling System, runs FD24ABV.D080398B, FD09ABV.D080398B, and FD03BLW.
D080398B, and simulations of the Data Resources, Inc. (DRI) Macroeconomic Model of the U.S. Economy.
0100200300400500600AverageCarbonReductions(MillionMetricTons)
01020304050607080PotentialLossFromReferenceCase(Billion1992Dollars)
1990+9%
1990-3%
1990+24%
1990+14%
19901990-7%
DRINEMSFigure 115. Comparison of Average U.S. Economic
Losses Projected by the NEMS and
DRI Models, 2008-2012
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model
of the U.S. Economy.
calculated by the DRI model over the same period is $32
billion. As a first approximation, this value closely
matches the estimate of the value of the lost output
calculated independently using the energy model
results (Figure 115).
The curve shown in Figure 114 can also be used to
estimate the international value of traded permits. The
carbon prices calculated in the NEMS model can be
characterized as the particular penalties that the United
States would be willing to pay to achieve a given carbon
mitigation target. For example, in the 1990+9% case, U.S.
carbon emission reductions average 325 million metric
tons per year during the period 2008 to 2012. The
difference between that reduction and the binding target
of 485 million metric tons under the Kyoto Protocol (as
reflected by the 1990-3% case) is assumed to be made up
through purchases of international permits abroad. The
value of those purchases is shown as the rectangle B
under the curve in Figure 114. For the 1990+9% case, this
represents a transfer of $23 billion dollars (1992 dollars)
to purchase permits abroad. For the 1990+24% case, the
transfer is $21 billion (Table 25). Even though more
permits are purchased abroad, the purchases occur
in the context of greater permit availability in the
1990+24% case, and the international price at which they
are bought is projected to be dramatically lower, as
shown in Table 25.
Focusing on the last two columns of Table 28 highlights
the role of international permit trading. Potential GDP is
a measure of the level of the output of the economy, but
as the last column indicates, there now is a cost to the
economy reflected in the transfer of funds abroad to buy
permits. Although the direct cost to the U.S. economy in
terms of lost potential GDP as a result of lower energy
consumption would be less in the 1990+24% and
1990+9% cases than in the 1990-3% case, there would be
additional losses of output available to the U.S. economy
in those cases. Funds transferred abroad for purchases
of international carbon emissions permits would, in
effect, reduce the amount of potential GDP available for
domestic use.
Energy Prices and the Role of
Monetary and Fiscal Policy
This following analysis focuses on the possible transitional impacts on the aggregate economy that would
result from efforts to reduce U.S. carbon emissions. The
measurement of actual output for the economy, or actual
GDP, is the key concept used in the examination of
changes in the aggregate economy as it adjusts to its
long-run path. In addition to internal frictions caused by
wage-price interactions and capital stock obsolescence,
losses in domestic income may occur as funds are transferred out of the United States to purchase international
carbon permits. Resources may be less than fully
employed, and the economy will move in a cyclical fashion as the initial cause of the disturbanceÑthe increase
in energy pricesÑplays out over time. Shifts in the sectoral composition of the economy would also accompany
the adjustment process.
Here, a single fiscal policy is assumed to accompany the
carbon mitigation policyÑthe revenues collected from
the domestic permit auction are returned to consumers
through personal income tax rebates. This is a stylized
analysis in that it represents only one of a wide range of
possible combinations of monetary and fiscal responses.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Impacts of Higher Energy Prices
on the Economy
As a direct consequence of the carbon price, aggregate
energy prices in the U.S. economy are expected to rise.
One way to measure this effect is to look at the percentage change in the level of prices in the economy. One
measure that can be used is the calculated wholesale
price index for fuel and power (Figure 116). In the 19903% case, aggregate energy prices are projected to double
by 2010 and then decline to 79 percent above reference
case price levels in 2020. In the 1990+9% case, energy
prices are 56 percent higher than the reference case projection in 2010 and remain more than 50 percent above
the reference case over the rest of the forecast period.
Prices in the 1990+24% case are 22 percent higher than
the reference case in 2010 and continue to rise to 33 percent in 2020.
These changes can also be expressed as rates of change.
In the reference case, overall energy prices rise by 3.9
percent per year between 2005 and 2010; however, in the
1990+9% case, aggregate energy prices rise at a rate of
13.5 percent per year, a difference of 9.6 percentage
points. The 1990-3% case shows a more dramatic rise, at
19.2 percent per year, and the 1990+24% case shows a
rise of 8.0 percent per year. Over the longer run,
measured between 2005 and 2020, the rise in energy
prices is less dramatic, with the reference case growth at
4.2 percent per year and the 1990+9% case at 7.2 percent
per year, a difference of 3.0 percentage points. For the
2005-2020 period, the 1990-3% case shows energy prices
rising by 8.3 percent and the 1990+24% case by 6.2
percent per year.
The projected energy price increases would also affect
downstream prices for all goods and services in the
economy. An intermediate measure is the producer
price index (Figure 117), which reflects price impacts on
intermediate goods and services. The projected increase
in producer prices relative to the reference case in 2010 is
16 percent in the 1990-3% case, 9 percent in the 1990+9%
case, and 4 percent in the 1990+24% case. By 2020, the
prices in the three carbon reduction case begin to
converge, as the differences in projected carbon prices
narrow.
Final prices for goods and services in 2009, as shown by
the consumer price index (CPI) series (Figure 118), are
more than 6.6 percent higher in the 1990-3% case than in
the reference case, 3.7 percent higher in the 1990+9%
case, and 1.4 percent higher in the 1990+24% case.
Again, by 2020, the differences narrow considerably. In
the reference case the CPI rises by 3.6 percent per year
between 2005 and 2010, but in the 1990+9% case, it rises
at a rate of 4.3 percent per year, a difference of 0.7
percentage points. The 1990-3% case shows a more
dramatic rise, at 4.8 percent per year, and the annual
199520002005201020152020020406080100120PercentChangeFromReferenceCase1990+9%
1990-3%
1990+24%
Figure 116. Projected Changes in Wholesale Price
Index for Fuel and Power Relative to
the Reference Case, 1998-2020
Note: Carbon permit revenues are assumed to be returned to
households through personal income tax rebates.
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model
of the U.S. Economy.
19952000200520102015202005101520PercentChangeFromReferenceCase1990+9%
1990-3%
1990+24%
Figure 117. Projected Changes in Producer Price
Index Relative to the Reference Case,
1998-2020
Note: Carbon permit revenues are assumed to be returned to
households through personal income tax rebates.
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model
of the U.S. Economy.
increase in the 1990+24% case is 3.9 percent. In the long
term, between 2005 and 2020, the increase in the
aggregate price for all goods and services is less
dramatic: 3.8 percent per year in the reference case and
3.9 percent per year in the 1990+9% case, a difference of
only 0.1 percentage points. Over the same period, the
1990-3% case projects a 4.0-percent annual increase in
the CPI and the 1990+24% case a 3.9-percent annual
increase.
One aspect of the CPI is particularly noteworthy. The
CPI measures the prices that consumers face, regardless
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
19952000200520102015202001234567PercentChangeFromReferenceCase1990+9%
1990-3%
1990+24%
Figure 118. Projected Changes in Consumer Price
Index Relative to the Reference Case,
1998-2020
Note: Carbon permit revenues are assumed to be returned to
households through personal income tax rebates.
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model
of the U.S. Economy.
of the country of origin of the product. Import prices, to
the extent that they do not rise at the rate of domestic
prices because non-Annex I countries do not face carbon
constraints, would dampen the price effects as lower
priced imports found their way into U.S. markets.
These figures suggest the following rule of thumb for the
year 2010. Each 10-percent increase in the level of aggregate prices for energy may lead to a 1.5-percent increase
in producer prices and a 0.7-percent increase in consumer prices.
Revenues Flows With International Permit
Purchases
The process of auctioning emissions permits would raise
large sums of money. If permits were purchased from
other countries, as is assumed in both the 1990+9% and
1990+24% cases, there would actually be two revenue
flowsÑdomestic and international. The carbon permit
revenues remaining within U.S. borders for each case
are calculated as the carbon permit price for that case
times the level of carbon emissions in the 1990-3% case.
Thus, the number of carbon permits purchased domestically remains constant; only the price at which they are
available varies across cases. Permits are assumed to be
purchased abroad in order for U.S. carbon emissions to
continue above the 1990-3% level. Therefore, the international revenue flow equals the difference between
actual emissions in the 1990+9% (or 1990+24%) case and
those in the 1990-3% case, times the carbon permit price
in the 1990+9% (or 1990+24%) case.
In the 1990-3% case the United States attains the binding
target level, and all the funds collected are kept within
U.S. borders. The revenue collected in 2010 is projected
to total $585 billion nominal dollars, calculated as the
level of carbon emissions (1,305 million metric tons)
times the carbon permit price ($266 in 1992 dollars),
adjusted to nominal dollars. In contrast, in the 1990+9%
case, U.S. emissions are reduced to 1,467 million metric
tons, or 162 million metric tons short of the binding
target. The domestic portion of the collected revenues is
equal to the binding target value of 1,305 million metric
tons times the new, lower carbon permit price of $148
per metric ton in 1992 dollars. The remaining 162 million
metric tons must be offset by permits purchased abroad,
again valued at $148 per metric ton. Figure 119 shows
total U.S. expenditures for carbon permits in the three
carbon reduction cases, and Figure 120 shows the
projected split between domestic and international
flows for the years 2010 and 2020.
The total projected payments for carbon permits become
substantially lower as the carbon reduction target moves
from 1990-3% to 1990+9% to 1990+24%. And, although
the flow of funds overseas represents an increasing
proportion of the total collected funds from the 1990+9%
case to the 1990+24% case, the actual level of the
transfers is relatively stable. Under the domestic-only
program of the 1990-3% case, the revenue from permits
is assumed to be returned to U.S. households through
income tax rebates. In the 1990+9% and 1990+24% cases,
only the domestic portion of the funds would be
recycled back to consumers. The international flow of
carbon permit revenue is considered an increase in the
purchase of imported services.
Dynamics of Adjustment in an Economy
With Frictions
The ultimate impacts of carbon mitigation policies
on the economy will be determined by complex
1995200020052010201520200100200300400500600700BillionNominalDollars1990+9%
1990-3%
1990+24%
Figure 119. Total Projected U.S. Payments for
Domestic and International Carbon
Emissions Permits, 1998-2020
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model
of the U.S. Economy.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
201020200100200300400500600700BillionNominalDollarsDomesticPaymentsInternationalPayments1990+9%
1990-3%
1990+24%
1990+9%
1990-3%
1990+24%
Figure 120. Projected Destinations of Funds Paid
for Carbon Emissions Permits, 2010
and 2020
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model
of the U.S. Economy.
interactions between elements of aggregate supply and
demand, in conjunction with monetary and fiscal policy
decisions. As such, any discussion of possible cyclical
impacts on the economy is bound to be characterized by
uncertainty and controversy. It should be recognized,
however, that the process of raising the price of energy
and downstream prices in the rest of the economy by the
magnitudes shown in Figure 116, 117, and 118 could
introduce cyclical behavior in the economy resulting in
employment and output losses beyond those associated
with the projected impacts on potential GDP.
The introduction of carbon emission limits would affect
both consumers and businesses. Households would be
faced with higher prices for energy and the need to
adjust spending patterns. Nominal energy expenditures
would rise, taking a larger share of the family budget for
goods and service consumption and leaving less for savings. Higher prices for energy would cause consumers
to try to reduce spending not only on energy, but on
other goods as well. Thus, changes in energy prices
would tend to disrupt both saving and spending
streams.
Energy services also represent a key input in the production of goods and services. As energy prices increase, the
costs of production rise, placing upward pressure on the
nominal prices of all intermediate goods and final goods
and services in the economy, with widespread impacts
on spending across many markets. The ultimate effect
will depend on opportunities for substitution away from
higher-cost energy to other goods and services and the
effectiveness of compensatory fiscal and monetary
policy.
The transitional adjustment of the economy can be captured by calculations of the actual GDP of the economy.
The impacts on actual GDP represent a measure of the
loss of output from the economy, recognizing that
adjustments are not frictionless and that all resources
may not be fully employed in the near term. The output
of the economy as reflected by actual GDP can cycle
around the measure of potential GDP.
The Role of Monetary Policy
Monetary policy can moderate or intensify the ultimate
impacts on the economy; however, trying to predict the
response of monetary authorities to large increases in
energy prices is a difficult task. The emphasis on controlling inflation relative to concerns about rising unemployment has changed over the past 20 years, and using
history as a guide does not remove the large amount of
uncertainty about the response of monetary authorities.
In addition, the types of financial instruments available
have become more numerous and more interdependent,
and the task of monitoring the NationÕs money supply
has become more complex.
The monetary authorities could concentrate on increased inflation resulting from higher energy prices
and choose not to increase the money supply in order to
moderate the resulting inflation. In this instance, output
and employment losses would be larger than they
would if the money supply were expanded when energy
prices increased. Another option would be to allow the
money supply to increase in order to remove the unemployment impacts while allowing substantial additional
price inflation. This analysis uses neither extreme of
these assumptions about the response of the Federal
Reserve. The discussion that follows represents a middle
path that the Federal Reserve might follow.
In the setting that has been describedÑreturning funds
in the form of personal income tax rebatesÑhigher
prices in the economy would place upward pressure on
interest rates. The Federal Reserve Board would then
seek to balance the consequences of higher energy prices
on the economy with possible adverse effects on output
and employment. The Federal Reserve would respond
to changes in inflation and unemployment brought on
by the initial carbon mitigation policy by making adjustments to influence the Federal funds rate.83 The adjustments would be designed to moderate the possible
impacts on both inflation and unemployment, and to
return the economy toward its long-run growth path.
The characterization of monetary policy reactions to
inflation and unemployment used in these simulations
is based on a DRI reaction function that has been estimated to reflect the historical relationship between the
83The Federal funds rate is the rate charged by a depository institution on an overnight sale of Federal funds to another depository institution. This rate influences the trend in behavior for other interest rates in the economy.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Federal funds rate and changes in inflation and unemployment. As such, the reaction function is a reflection of how the Federal Reserve may react to changes
in the economy caused by the carbon price, based on
past behavior.
If the rate of inflation increases, but unemployment does
not increase, the Federal Reserve may choose to let the
nominal interest rate rise in an attempt to cut the rise in
inflation. However, if this is accompanied by an increase
in the unemployment rate, the Federal Reserve may consider a cut in the rate to stimulate economic expansion
and the demand for labor. In essence, there is a balancing game between the two factorsÑinflation and unemploymentÑas the initial originating policy initiative
has uneven impacts on the two over time. Figures 121,
122, and 123 show the interrelationship between the projected inflation rate, unemployment rate, and Federal
funds rate in the 1990+24%, 1990+9%, and 1990-3%
cases. This assessment combines the monetary policy
formulation described above with a fiscal policy that
returns collected carbon permit revenues back to
consumers. An alternative combination of fiscal and
monetary policy is considered later in this section.
Focusing first on the 1990+9% case, the inflation rate
jumps from 3.3 percent per year to 5.1 percent per year, a
difference of 1.8 percentage points in 2005, the first year
of the energy price rise, and continues to remain high for
the first 4 years of the carbon reduction program. In the
same 4-year period, the unemployment rate first responds slowly and then accelerates to a peak in 2009 that
is more than a full percentage point above the reference
case unemployment rate, rising from 5.6 percent in the
reference case to 6.8 percent in the 1990+9% case. The
key point here is that the responses of inflation and
unemployment are not symmetric over time. There is a
lag between the two effects with output and employment effects lagging behind price effects. Prices rise in
the economy in response to the initial energy price
increase and then to secondary price effects as the costs
of intermediate goods and services rise. Business, in
response to rising prices and lower aggregate demand,
absorbs the near-term output loss but eventually
reduces its use of labor. The lag from initial price effects
to ultimate output and employment losses can be a year
or so.
1995200020052010201520200.00.51.01.52.02.5-0.5-1.0ChangeinPercentagePoints1990+9%
1990-3%
1990+24%
Figure 122. Projected Changes in U.S.
Unemployment Rate Relative to the
Reference Case, 1998-2020
Note: Carbon permit revenues are assumed to be returned to
households through personal income tax rebates.
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model
of the U.S. Economy.
1995200020052010201520200.00.51.01.52.02.5-0.5-1.0ChangeinPercentagePoints1990+9%
1990-3%
1990+24%
Figure 121. Projected Changes in U.S. Inflation
Rate Relative to the Reference Case,
1998-2020
Note: Carbon permit revenues are assumed to be returned to
households through personal income tax rebates.
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model
of the U.S. Economy.
19952000200520102015202000.20.40.60.8-0.2-0.4-0.6-0.8ChangeinPercentagePoints1990+9%
1990-3%
1990+24%
Figure 123. Projected Changes in U.S. Federal
Funds Rate Relative to the Reference
Case, 1998-2020
Note: Carbon permit revenues are assumed to be returned to
households through personal income tax rebates.
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model
of the U.S. Economy.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
As a result of the differential effects projected for inflation and unemployment during the years from 2005 to
2008, the Federal Reserve is assumed to allow a modest
rise in the Federal funds rate in the short term, when
concern over inflation outweighs concern over GDP
losses and unemployment. After the initial rise in energy
prices, with the carbon price actually projected to fall
after 2009, the inflation rate reverts to that projected in
the reference case; however, aggregate output is still
depressed, and unemployment in the economy remains
above the reference case value. During this period, the
Federal Reserve reacts by reducing the Federal funds
rate, in order to combat the loss in output and employment in the economy. After 10 years, by 2015, both inflation and unemployment have returned to at or about
reference case levels. The Federal Reserve again allows
interest rates to rise to bring the economy back to its
long-run growth path.
Impacts on Actual Output and
Consumption
In the 1990+9% case, potential GDP is projected to
decline smoothly over time, leveling off to a steady-state
value of approximately 0.35-percent loss in output for
the economy (Figure 124). In contrast, actual GDP is
buffeted about as the economy adjusts to the significant
price pressures brought on by higher energy prices,
losing approximately 2.5 percent in real output by 2009.
The loss in actual output can also be described in terms
of the impacts on the growth rate for actual GDP.
Between 2005 and 2010, actual GDP is projected to grow
by 2.0 percent per year in the reference case. In the
1990+9% case, the growth rate slows to 1.6 percent per
19952000200520102015202001-1-2-3-4-5PercentChangeFromReferenceCasePotentialGDPActualGDPFigure 124. Projected Changes in Potential and
Actual U.S. Gross Domestic Product in
the 1990+9% Case Relative to the
Reference Case, 1998-2020
Note: Carbon permit revenues are assumed to be returned to
households through personal income tax rebates.
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model
of the U.S. Economy.
year, reducing growth in the economy over the same
period by 0.4 percentage points.
After 2010, although the economy is still below the
reference case, actual GDP begins to cycle in response to
energy prices. The economy cycles for two fundamental
reasons. First, output effects lag price effects in the
economy as consumers and businesses adjust to the
price changes. Also, in the case considered, the rise in
energy prices levels off dramatically by 2010, and
inflation rates are actually lower than in the reference
case, as shown in Figure 121. The interesting property of
the two output concepts, actual and potential, is that
they begin to converge by 2015, 10 years after the
beginning of the initial impacts on the economy. By 2020
they have merged into a steady-state path. This suggests
that while the economy may very well be on a long-run
path that could yield a loss to the economy of about 0.3
percent if its potential output, there is the possibility that
near-term impacts may be larger as the economy adjusts
to its long-run trajectory.
The projected impacts on actual GDP in the 1990-3% case
peak at a loss of 4.1 percent in 2009, but again rebound
back toward and merge with the ultimate potential GDP
impact measure of 0.55 percent (Figure 125). The growth
rate between 2005 and 2010 slows to 1.3 percent per year,
a reduction of 0.7 percentage points from the reference
case growth rate of 2.0 percent. In the 1990+24% case,
actual GDP shows a peak loss of 1.0 percent relative to
the reference case in 2010, with no significant impact on
the growth rate, then begins to return to its long-run
potential GDP path. In this case, however, because the
carbon price is still rising, the economy continues to
show a slight divergence between actual and potential
GDP in 2020, although the gap is significantly narrowed
(Figure 126).
Beyond the aggregate impact on GDP, a significant
change in the composition of final demand is projected
in the carbon reduction cases (Table 29). In the 1990+9%
case, consumption in 2009 is projected to be 1.9 percent
lower than projected in the reference case (Figure 127).
Returning the carbon permit revenues to households
through personal income tax rebates moderates the
impacts on disposable income in the economy, which, in
turn moderates the adverse impact on purchases of consumer goods and services, and therefore the impact on
the aggregate economy measured by actual GDP.
Investment is more severely affected, with rising interest
rates and a general loss in demand in the economy projected in the years immediately after the imposition of
the carbon price (Figure 128). In 2007, investment in the
1990+9% case is projected to be 5.9 percent below the reference case projection. After 2008, with lower interest
rates, the economy begins to rebound as investment
expands rapidly. By 2013, investment is above the reference case by 3.2 percent and is leading the recovery.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
19952000200520102015202001-1-2-3-4-5PercentChangeFromReferenceCasePotentialGDPActualGDPFigure 125. Projected Changes in Potential and
Actual U.S. Gross Domestic Product in
the 1990-3% Case Relative to the
Reference Case, 1998-2020
Note: Carbon permit revenues are assumed to be returned to
households through personal income tax rebates.
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model
of the U.S. Economy.
19952000200520102015202001-1-2-3-4-5PercentChangeFromReferenceCasePotentialGDPActualGDPFigure 126. Projected Changes in Potential and
Actual U.S. Gross Domestic Product in
the 1990+24% Case Relative to the
Reference Case, 1998-2020
Note: Carbon permit revenues are assumed to be returned to
households through personal income tax rebates.
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model
of the U.S. Economy.
1995200020052010201520200246-2-4-6-8-10PercentChangeFromReferenceCase1990+9%
1990-3%
1990+24%
Figure 127. Projected Changes in Real
Consumption in the U.S. Economy
Relative to the Reference Case,
1998-2020
Note: Carbon permit revenues are assumed to be returned to
households through personal income tax rebates.
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model
of the U.S. Economy.
1995200020052010201520200246-2-4-6-8-10PercentChangeFromReferenceCase1990+9%
1990-3%
1990+24%
Figure 128. Projected Changes in Real
Investment in the U.S. Economy
Relative to the Reference Case,
1998-2020
Note: Carbon permit revenues are assumed to be returned to
households through personal income tax rebates.
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model
of the U.S. Economy.
The 1990-3% case shows a pattern of adjustment similar
to that projected in the 1990+9% case, except that the
reaction in terms of both consumption and investment is
more extreme, given the higher carbon price. Consumption reaches its lowest point in the year 2009 at 2.8
percent below the reference case. Thereafter, consumption returns to the reference case level in 2013 and by
2015 is 0.8 percent above the reference case level.
Investment is more volatile, falling to 9.1 percent below
reference case levels by 2008. Again, with interest rates
declining relative to the reference case after 2010, investment recovers rapidly and by 2013 is 5.1 percent above
the reference case.
The 1990+24% case reflects a much smoother path for
both consumption and investment. Consumption remains below the reference case throughout the period,
but with a maximum loss of only 0.8 percent in 2010. The
impact on investment, likewise, is more moderate than
in the 1990-3% and 1990+9% cases, falling to 2.2 percent
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Table 29. Projected Economic Impacts of Carbon Reduction Cases Assuming Personal Income Tax Rebate
(Changes From Reference Case)
Analysis Case 2010 2015 2020
1990-3%
Collections(BillionNominalDollars)................................ 585 633 674
Wholesale Price Index for Fuel and Power (Percent Change) ............ 99.1 89.9 78.9
ProducerPriceIndex(PercentChange)............................. 15.7 14.6 12.9
ConsumerPriceIndex(PercentChange)............................ 6.0 4.2 2.9
UnemploymentRate(DifferenceinRate)............................ 1.6 -0.4 0.1
FederalFundsRate(DifferenceinRate)............................ -0.7 0.5 0.1
PotentialGDP(PercentChange).................................. -1.2 -0.8 -0.6
RealGDP(PercentChange)..................................... -3.5 -0.1 -0.7
RealGDP(Billion1992Chain-WeightedDollars) ..................... -327 -12 -72
Consumption(PercentChange)................................... -2.3 0.8 0.4
Investment(PercentChange)..................................... -3.6 3.3 -0.0
IndustrialOutput(PercentChange)................................ -5.8 -2.5 -3.6
1990+9%
Collections(BillionNominalDollars)................................ 317 340 391
Wholesale Price Index for Fuel and Power (Percent Change) ............ 55.7 53.7 52.5
ProducerPriceIndex(PercentChange)............................. 9.0 8.8 8.7
ConsumerPriceIndex(PercentChange)............................ 3.5 2.5 2.1
UnemploymentRate(DifferenceinRate)............................ 0.9 -0.2 0.2
FederalFundsRate(DifferenceinRate)............................ -0.4 0.2 -0.1
PotentialGDP(PercentChange).................................. -0.7 -0.4 -0.4
RealGDP(PercentChange)..................................... -2.0 -0.1 -0.6
RealGDP(Billion1992Chain-WeightedDollars) ..................... -187 -15 -68
Consumption(PercentChange)................................... -1.5 0.2 -0.2
Investment(PercentChange)..................................... -1.5 1.8 -0.1
IndustrialOutput(PercentChange)................................ -3.0 -1.6 -3.1
1990+24%
Collections(BillionNominalDollars)................................ 128 206 271
Wholesale Price Index for Fuel and Power (Percent Change) ............ 21.5 29.3 32.9
ProducerPriceIndex(PercentChange)............................. 3.6 4.9 5.5
ConsumerPriceIndex(PercentChange)............................ 1.5 1.6 1.4
UnemploymentRate(DifferenceinRate)............................ 0.5 0.2 0.1
FederalFundsRate(DifferenceinRate)............................ -0.2 -0.0 -0.1
PotentialGDP(PercentChange).................................. -0.2 -0.3 -0.3
RealGDP(PercentChange)..................................... -1.0 -0.5 -0.5
RealGDP(Billion1992Chain-WeightedDollars) ..................... -96 -54 -49
Consumption(PercentChange)................................... -0.8 -0.4 -0.3
Investment(PercentChange)..................................... -1.8 0.2 0.1
IndustrialOutput(PercentChange)................................ -1.3 -1.3 -2.0
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model of the U.S. Economy.
below the reference case in 2008. Thereafter, investment
returns to the reference case level and essentially remains at that position for the remainder of the forecast
period.
Figure 129 shows the projected impacts on consumption
and investment in terms of growth rates between 2005
and 2010 and between 2005 and 2020. Between 2005 and
2010, consumption growth rates fall from 2.0 percent per
year in the reference case to 1.9 percent in the 1990+24%
case, 1.7 percent in the 1990+9% case, and 1.6 percent in
the 1990-3% case. Investment shows a similar, but more
pronounced profile, with growth declining from 2.9
percent per year in the reference case to 2.5 percent, 2.6
percent, and 2.2 percent in the respective carbon
reduction cases. Slight variations in the order of the
impactsÑthe 1990+24% case at 2.5 percent and the
1990+9% case at 2.6 percentÑcan be explained by the
highly cyclical effects on investment, as shown in Figure
128. In the long run, as indicated by the projected growth
rates between 2005 and 2020, growth in both consumption and investment returns to the reference case
rates.
These results indicate that, as a result of higher energy
prices, the economy may absorb a near-term loss in output in response to higher inflation and a rise in the
unemployment rate. However, with appropriate action
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Consumption2005-2010Investment2005-2010Consumption2005-2020Investment2005-20200.00.51.01.52.02.53.03.5PercentperYearReference1990+24%1990+9%1990-3%
Figure 129. Consumption and Investment
Growth Rates
Note: Carbon permit revenues are assumed to be returned to
households through personal income tax rebates.
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model
of the U.S. Economy.
on the part of the monetary authorities, these impacts
could be mitigated, and in the long-term the economy
could rebound.
The Role of Fiscal Policy
This analysis assumes that revenues from carbon permits would be collected by the Federal Government,
which would have a number of alternatives with regard
to their disposition. The producers of carbon-intensive
fuels could keep the permit revenues; or the Government could either use the revenues to reduce the
national debt, return them to businesses through reductions in corporate income tax rates or increased business
tax credits, return them to consumers through personal
income tax rebates, or return them to both consumers
and businesses through social security tax rebates. Each
method of using the collected permit revenue is
plausible, and each method would have a different
economic impact.
Returning the funds to consumers through personal
income tax rebates or returning them to consumers and
businesses through social security tax rebates would
work to ameliorate the short-term impacts on the economy by bolstering disposable income. Alternative fiscal
policies, such as having the Federal Government use the
funds to lower the Federal debt level, or a corporate
income tax rebate, probably would result in larger
near-term impacts, because disposable income and
therefore consumption would fall by greater amounts.
Conversely, policies that serve to shift the economy
away from consumption toward investment may have
greater long-term benefits in terms of expansion of the
aggregate capital stock.
All the projections discussed so far in this chapter have
assumed a policy of returning carbon permit revenues to
households through personal income tax rebates, using
a lump sum transfer.84 To highlight the potential significance of an alternative fiscal regime this chapter next
reviews the potential effects of a rebate of social security
taxes that passes funds back to both employees and
employers in equal amounts. The analysis of a hypothetical rebate of the social security tax is meant only to
be descriptive of a tax measure that could have the effect
of reducing price pressures in the economy by lowering
business costs, while also accomplishing a partial compensation to consumers for the higher energy bills they
would face. The two policies considered in this analysisÑthe personal income tax rebate and the social security tax rebateÑare only meant to be representative of a
set of possible fiscal policies that might accompany an
initial carbon mitigation policy.
The fundamental difference between the two policies is
in their treatment of business. On the employer side, the
reduction in employer contributions to the social security system would lower costs to the firm and, thereby,
moderate the near-term price consequences to the economy. Since it is the price effect that produces the predominately negative effect on the economy, any steps to
reduce inflationary pressures would serve to moderate
adverse impacts on the economy. The smaller impact on
aggregate prices would also moderate the monetary policy reaction, as shown in Figures 130, 131, and 132. In all
the carbon reduction cases, the reaction of the Federal
funds rate to the economic effects of higher energy
prices would be less pronounced than projected under
the assumption of a personal income tax reduction.
Similarly, the social security tax option would moderate
the potential impact on actual GDP in the carbon reduction cases (Figure 133), largely because of the cost-
cutting aspects of lowering of the employer portion of
the tax. Similar moderating effects would be seen for
consumption (Figure 134) and investment (Figure 135).
Under both policies, the economy would eventually
revert to a long-run path consistent with the path of
potential output.
84In the DRI model for personal taxes only, a lump sum transfer produces the same effects as a cut in the personal income tax rate.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
19952000200520102015202000.20.40.60.8-0.2-0.4-0.6-0.8ChangeFromReferenceCasePersonalIncomeTaxRebateSocialSecurityTaxRebateFigure 130. Projected Changes in U.S. Federal
Funds Rate in the 1990-3% Case
Relative to the Reference Case Under
Different Fiscal Policies, 1998-2020
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model
of the U.S. Economy.
19952000200520102015202000.20.40.60.8-0.2-0.4-0.6-0.8ChangeFromReferenceCasePersonalIncomeTaxRebateSocialSecurityTaxRebateFigure 132. Projected Changes in U.S. Federal
Funds Rate in the 1990+24% Case
Relative to the Reference Case Under
Different Fiscal Policies, 1998-2020
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model
of the U.S. Economy.
19952000200520102015202000.20.40.60.8-0.2-0.4-0.6-0.8ChangeFromReferenceCasePersonalIncomeTaxRebateSocialSecurityTaxRebateFigure 131. Projected Changes in U.S. Federal
Funds Rate in the 1990+9% Case
Relative to the Reference Case Under
Different Fiscal Policies, 1998-2020
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model
of the U.S. Economy.
19952000200520102015202001-1-2-3-4-5PercentChangeFromReferenceCasePotentialGDPActualGDP,
SocialSecurityTaxRebateActualGDP,
PersonalIncomeTaxRebateFigure 133. Projected Changes in Potential and
Actual U.S. Gross Domestic Product in
the 1990+9% Case Relative to the
Reference Case Under Different Fiscal
Policies, 1998-2020
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model
of the U.S. Economy.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
1995200020052010201520200246-2-4-6-8-10PercentChangeFromReferenceCase1990+9%
1990-3%
1990+24%
Figure 134. Projected Changes in Real
Consumption in the U.S. Economy
Relative to the Reference Case,
1998-2020, Assuming a Social Security
Tax Rebate
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model
of the U.S. Economy.
1995200020052010201520200246-2-4-6-8-10PercentChangeFromReferenceCase1990+9%
1990-3%
1990+24%
Figure 135. Projected Changes in Real
Investment in the U.S. Economy
Relative to the Reference Case,
1998-2020, Assuming a Social Security
Tax Rebate
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model
of the U.S. Economy.
Energy Investment
This macroeconomic analysis of the costs of the Kyoto
Protocol includes the direct fuel costs and only those
investment costs that are comparable in magnitude with
those in the reference case. Business investments above
reference case levels may be required to reduce energy
costs in response to increasing energy prices. The poten-
tial incremental costs of investment in technology and
infrastructure that may be necessary to obtain the emis-
sions reductions specified in each of the cases analyzed
are not included, either because they are not available or,
in cases where they are available, because there is no
direct mapping to the National Income and Product
Accounts.
Full investment costs would include: (1) fuel and equip-
ment costs, including the cost of capital and the cost of
premature obsolescence; (2) research and development
costs; (3) infrastructure costs, including equipment main-
tenance, supply, and distribution; (4) regulatory monitor-
ing and enforcement costs; (5) the costs for manufacturers
to retool prematurely; and (6) the costs of lost investment
opportunities. This macroeconomic analysis, like all oth-
ers, does not include all of these investment costs. The
premature obsolescence of capitalÑwhen a firm is forced
to retire equipment before the end of its physical or eco-
nomic lifeÑis typically ignored or assumed to be costless,
because estimates of the amount of capital retired early
are difficult to make. Estimates of the full cost of develop-
ing new technologies, particularly the associated research
and development costs, are generally unavailable. In
addition, certain new technologies may require a consid-
erable amount of additional investment in infrastructure
in order to be widely adopted. For example, widespread
adoption of carbon-free vehicles (such as hydrogen fuel
cell automobiles) may require substantial investment to
guarantee consumers that hydrogen refueling stations are
conveniently located and that the development of hydro-
gen stations does not present safety risks. Estimates of
these costs are difficult to obtain and are at best uncertain.
In NEMS, capital costs are included for newly constructed
technologies in the electricity generation sector, for major
appliances and technologies in the residential and com-
mercial sectors, for new vehicles in the transportation sec-
tor,a and for new natural gas pipelines and new oil
refineries. The investment costs in buildings include new
equipment costs but do not include costs attributable to
improving the energy efficiency of structures, such as
insulation and thermal windows. For generators, the
investment costs include additional expenditures on both
equipment and structures required for generation, trans-
mission, and distribution of electricity. The NEMS repre-
sentations of investment costs for generators probably are
the most detailed estimates available from any energy
modeling system; however, the financial accounting cate-
gories available from the electricity sector do not map
directly into the National Income and Product Accounts
included in macroeconomic models. The mapping
difficulties are even greater for the end-use sectors.
Reconciling and meaningfully incorporating investment
information from energy models into macroeconomic
models is a research area that still needs to be studied. As
a result, this analysis includes only the direct cost of fuels
when evaluating the macroeconomic impacts of the
Kyoto Protocol.
aWhile infrastructure costs are not directly included for the transportation model, the rate at which infrastructure can expand is
included in the adoption of new alternative-fuel vehicles.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
Sectoral Impacts
Regardless of the effects of carbon mitigation policies on
the ultimate level of the aggregate economy, there are
likely to be impacts on the configuration of the sectoral
output of the economy. This section describes one possible set of outcomes. While the results are very uncertain,
they indicate the potential for differential impacts
among industries, primarily as the result of four key factors:
¥ First, the direct impact of higher energy prices is a
reduction in energy demand, particularly for coal
with its high carbon content. The consequences are
reductions in output from the mining sector and
from all services connected to the production and
distribution of coal.
¥ Second, higher energy prices disproportionately
increase the cost of production for energy-intensive
industries. As energy price increases are passed
along by industry through higher prices for their
products, consumers will tend to substitute away
from the relatively expensive energy-intensive products to less energy-intensive products and services.
The consequences are reductions in gross output
from the energy-intensive sectors of the economy,
principally, chemicals and allied products; stone,
clay, glass, and concrete; and primary metals.
¥ Third, the changing composition of macroeconomic
final demand will alter the composition of sectoral
output. In the cases considered here, all the carbon
permit revenues are assumed to be returned to consumers through personal income tax rebates, moderating the projected impacts on disposable income.
Consequently, in percentage terms, consumer
spending falls by less than GDP, while investment
falls by more. This change in the composition of final
demand decreases the output from consumer-
related sectors, such as services and retail trade, by
less than the average drop for all economic output,
while decreasing the output from the construction
and manufacturing sectors by more than the average.
¥ Finally, because the carbon emissions restrictions are
placed only on Annex I countries, industries with
high levels of imports, particularly those with
imports from non-Annex I countries, will see larger
reductions in domestic output than industries with
low import penetration. If imports are already
competitive, increasing the cost of production for the
domestic industry and not for non-Annex I importers will tend to increase imports, leading to a drop
in domestic output. For this reason, output from
manufacturing sectors such as leather and leather
products, electronic and other electrical equipment,
and miscellaneous manufacturing will fall by more
than the output for the manufacturing sector as a
whole.
It is difficult, a priori, to predict the degree and rate of
change of such effects. Figure 136 shows the disaggregated impacts of restricting carbon emissions in
the 1990+9% case. The upper part of the graph shows the
projected growth rates for GDP, total gross output, and
sectoral gross output for the major SIC divisions
between 2005 and 2010. The GDP and total gross output
growth rates provide an economy-wide frame of reference against which the sectoral growth rates can
be compared. The lower part of the graph shows the
growth rates for total manufacturing gross output
and sectoral gross output by 2-digit SIC breakdown
between 2005 and 2010, with the growth rate for total
manufacturing gross output as a reference. Figures 137
and 138 show the results for the 1990-3% and 1990+24%
cases, respectively.
GrossDomesticProductTotalGrossOutputAgriculture,Forestry,andFishingMiningConstructionManufacturingTransportation/Communication/UtilitiesWholesaleTradeRetailTradeFinance,Insurance,andRealEstateServicesPublicAdministration02468-2-4-6Reference1990+9%
ManufacturingGrossOutputFoodandKindredProductsTobaccoProductsTextileMillProductsApparelandAlliedProductsLumberandWoodProductsFunitureandFixturesPaperandAlliedProductsPrintingandPublishingChemicalsandAlliedProductsPetroleumRefiningRubberandMiscellaneousPlsticsLeatherandLeatherProductsStone,Clay,Glass,andConcretePrimaryMetalsFabricatedMetalProductsIndustrialMachineryandComputersElectronicandElectricalEquipmentTransportationEqupmentInstrumentsMiscellaneousManufacturing02468-2-4-6PercentperYearFigure 136. Projected Sectoral Growth Rates in
Real Economic Output in the 1990+9%
Case, 2005-2010
Note: Carbon permit revenues are assumed to be returned to
households through personal income tax rebates.
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model
of the U.S. Economy.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
GrossDomesticProductTotalGrossOutputAgriculture,Forestry,andFishingMiningConstructionManufacturingTransportation/Communication/UtilitiesWholesaleTradeRetailTradeFinance,Insurance,andRealEstateServicesPublicAdministration02468-2-4-6Reference1990-3%
ManufacturingGrossOutputFoodandKindredProductsTobaccoProductsTextileMillProductsApparelandAlliedProductsLumberandWoodProductsFunitureandFixturesPaperandAlliedProductsPrintingandPublishingChemicalsandAlliedProductsPetroleumRefiningRubberandMiscellaneousPlsticsLeatherandLeatherProductsStone,Clay,Glass,andConcretePrimaryMetalsFabricatedMetalProductsIndustrialMachineryandComputersElectronicandElectricalEquipmentTransportationEqupmentInstrumentsMiscellaneousManufacturing02468-2-4-6PercentperYearFigure 137. Projected Sectoral Growth Rates in
Real Economic Output in the 1990-3%
Case, 2005-2010
Note: Carbon permit revenues are assumed to be returned to
households through personal income tax rebates.
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model
of the U.S. Economy.
GrossDomesticProductTotalGrossOutputAgriculture,Forestry,andFishingMiningConstructionManufacturingTransportation/Communication/UtilitiesWholesaleTradeRetailTradeFinance,Insurance,andRealEstateServicesPublicAdministration02468-2-4-6Reference1990+24%
ManufacturingGrossOutputFoodandKindredProductsTobaccoProductsTextileMillProductsApparelandAlliedProductsLumberandWoodProductsFunitureandFixturesPaperandAlliedProductsPrintingandPublishingChemicalsandAlliedProductsPetroleumRefiningRubberandMiscellaneousPlsticsLeatherandLeatherProductsStone,Clay,Glass,andConcretePrimaryMetalsFabricatedMetalProductsIndustrialMachineryandComputersElectronicandElectricalEquipmentTransportationEqupmentInstrumentsMiscellaneousManufacturing02468-2-4-6PercentperYearFigure 138. Projected Sectoral Growth Rates in
Real Economic Output in the
1990+24% Case, 2005-2010
Note: Carbon permit revenues are assumed to be returned to
households through personal income tax rebates.
Source: Simulations of the Data Resources, Inc. (DRI) Macroeconomic Model
of the U.S. Economy.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
7. Comparing Cost Estimates for the Kyoto Protocol
Introduction Differences in the cost estimates for meeting the Kyoto
Protocol targets can be related to important differences
This chapter provides a comparison of recent publicly in assumptions about (1) economic growth in the refer-
available estimates of the costs of achieving the Kyoto ence cases without the Kyoto Protocol, (2) the status of
Protocol carbon reduction targets in the United States the resources available (e.g., resource base, world oil
for the period 2008 to 2020. The projections are com-prices, and the slate of technologies available to the mar-
pared for the years 2010 and 2020, when the information ketplace), (3) the sensitivity of energy demand to price
is available, for the following projection sources: the changes, (4) the degree of foresight that decisionmakers
Energy Information Administration (EIA) using the have in the marketplace, (5) the structure and function of
National Energy Modeling System (NEMS), WEFA,85
the economy (e.g., how quickly the economy can shift to
Charles River Associates (CRA) using the Multi-less energy-intensive industries when the price of
Regional Trade model (MRT),86 the Pacific Northwest energy relative to capital and materials increases), (6) the
National Laboratory (PNNL) using the Second Genera-degree and speed of substitution for factors of production Model (SGM),87 the Massachusetts Institute of Tech-tion (capital, labor, energy, and materials) when their
nology (MIT) using the Emissions Prediction and Policy relative prices change, and (7) the representation of tech-
Analysis Model (EPPA),88 Electric Power Research Insti-nology (i.e., representation of vintaged energy equiptute (EPRI) using the MERGE model89 and Data ment and the penetration of new technologies).
Resources, Inc. (DRI).90 Differences between studies are
related, to the extent possible, to the features of the modeling systems used (e.g., level of aggregation, level of Summary of Comparisons
geographic coverage), important assumptions employed, and the particular points of view embodied in Because the information available varies considerably, a
the models.91 detailed comparison among the sources is virtually
impossible. Therefore, a comparison of common
Two cases were solicited for analyses from each group: a variables is provided in this section, with an explanation
7-percent-below-1990 (1990-7%) case in which the for the differences between the sources. Comparisons
United States is assumed to reduce carbon emissions to are provided for three of the cases analyzed in this
1990-7% levels for the period 2008-2020 without the report: the 1990-7% case and two casesÑ9 percent
benefit of sinks, offsets, international carbon permit above 1990 (1990+9%) and 14 percent above 1990
trading, or the Clean Development Mechanism (CDM); (1990+14%)Ñthat are comparable in some respects to
and a best estimate of the impact on U.S. energy markets the Annex I trading case. The variables compared are
if sinks, offsets, and Annex I emissions trading were carbon price, change in actual gross domestic product
allowed, but not global trading or CDM. (GDP) from the respective reference case in each study,
85 WEFA, Inc., Global Warming: The High Cost of the Kyoto Protocol, National and State Impacts (Eddystone, PA, 1998).
86 Both the CRA and WEFA studies have been supported to some extent by industry groups, including the American Petroleum Institute.
87 J.A. Edmonds et al., Modeling Future Greenhouse Gas Emissions: The Second Generation Model Description (Washington, DC: Pacific
Northwest National Laboratory, September 1992). Runs using PNNL's SGM model formed the basis for the testimony provided by Dr. Janet
Yellen, chairman of the Council of Economic Advisers, on March 4, 1998, before the House Commerce Committee, Energy and Power Subcommittee.
88 H.D. Jacoby, R. Eckhaus, A.D. Ellerman, et al. ÒCO2 Emission Limits: Economic Adjustments and the Distribution of Burdens,Ó Energy
Journal, Vol. 18, No. 3 (1997), pp. 31-58. MIT's analysis is part of a much larger integrated assessment methodology funded by the Office of
Energy Research, U.S. Department of Energy.
89 A.S. Manne and R.G. Richels, ÒOn Stabilizing CO2 ConcentrationsÑCost Effective Emissions Reduction Strategies,Ó Energy and Environmental Assessment, Vol. 2 (1997), pp. 251-265. EPRI's work is self-funded and is part of the research agenda of electric utilities.
90 Standard and Poors DRI, The Impact of Meeting the Kyoto Protocol on Energy Markets and the Economy (July 1998).
91 Information used in this chapter was contributed by Dr. Montgomery and Dr. Bernstein of Charles River Associates, Dr. Richels of the
Electric Power Research Institute, Dr. Edmonds of Pacific Northwest National Laboratory, and Professor Jacoby of MIT.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
actual and potential GDP loss, expenditures for
purchases of carbon emission permits, change in carbon
intensity from the respective reference case, and change
in fossil fuel consumption. Tables 30 and 31 provide
comparisons of the results for 2010 and 2020. Further
details are provided in Appendix C.
For the WEFA study, comparisons are provided only
with the 1990-7% case. For DRI comparisons are
provided only for a trading case (Case 2). WEFA does
not believe that sinks, offsets, or trading will be agreed
upon and implemented before the target period of 2008
to 2012, nor by 2020. As noted earlier in the report, EIA
does not have the capability to analyze international
trading and thus is unable to provide a most likely
estimate of the impacts of the international trading
provisions of the Kyoto Protocol, or of sinks and offsets,
on the level of the energy-related carbon reductions
required to meet the 1990-7% reduction in greenhouse
gases. EIA's 1990+9% and 1990+14% cases are used in
Table 31, because the carbon emissions levels of those
cases were most closely aligned with the other studies
presented.
Some of the major factors that result in differences in the
projected carbon prices and costs to achieve the 1990-7%
carbon reduction level are:
¥ Relative differences in reference GDP and carbon
emissions growth rates through 2020. For example,
if the GDP or carbon emissions growth rate in a
given reference case is lower than that in EIA's reference case, a smaller carbon reduction will be needed,
and it will generally be easier to achieve the emissions target. If the reference GDP growth or carbon
emissions growth is higher than in EIA's reference
case, the carbon price and GDP impacts relative to
those projected by EIA in this study will generally be
higher. Most of the major differences among the
analyses are attributable to differences in the reference case projections.
¥ Differences in assumptions about the potential for
economical life extension or refurbishment of
existing nuclear power plants beyond their normal
licensing period. If, for example, no existing nuclear
plants were retired by 2020, about 40 million metric
tons of carbon emissions would be avoided from the
combustion of fossil fuel used in plants to replace
them.
¥ The amount of knowledge about future events
assumed for decisionmakers. For example, models
that assume that decisionmakers have perfect
knowledge about future prices, demands, or policies
could underestimate compliance costs, because all
future events would be anticipated with certainty
and responded to at minimum cost. Analyses that
assume that all decisionmakers are myopic will tend
to overstate transition costs.
¥ The amount of lead time decisionmakers are
assumed to have to adjust to the Kyoto Protocol.
For example, if a model starts to begin the adjustment process in 1985, 1990 or 1995, it could underestimate the costs of complying with the Kyoto
Protocol, because it has more time to adjust. Models
that wait until the last moment to begin the adjustments could overstate adjustment costs.
¥ The level of aggregation in the model for technologies and goods. A model that deals only with aggregate products such as oil, gas, or coal without the
benefit of an explicit technology representation may
not capture important variables that can significantly affect energy efficiency and intensity or the
changing mix of industries that may result from
compliance efforts.
¥ The amount of focus on the transition process and
the associated costs. For example, a model that
assumes that all capital and labor can be immediately switched from one use to another cannot capture
the short-term or medium-term impacts of complying with the Kyoto Protocol, because those costs are
not reflected in the model.
¥ The assumed speed and extent of changes that consumers can make in energy consumption or
demand for energy services in response to changing prices (price elasticities of demand). Higher
assumed elasticities make it easier to achieve the carbon target through demand reductions. Lower elasticities make it more difficult.
Among the studies compared in Table 30, the projected
carbon prices in 2010 fall into three groups. MIT, EPRI,
CRA, and WEFA project prices in the range of $265
(WEFA) to $295 (CRA) per metric ton of carbon. PNNL
projects carbon prices of about $221 per metric ton. EIA
projects carbon prices of $348 per metric ton.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity
In the PNNL study, assumptions about consumer price
responsiveness (demand elasticities and capital/energy
substitution elasticities) are consistent with a long-term
time frame where everything is changeable.92,93 Applying the long-term elasticities to the short-term andmidterm period can overstate the ease and willingness with
which consumers change their equipment or reduce
their consumption in response to price increases.94
Further contributing to the low carbon price projection is
the amount of lead time consumers have to respond, as
well as differences in the reference case economic
growth rates. The PNNL and EIA reference case GDP
projections are very similar. However, PNNL's end-use
representation does not explicitly represent technologies, and PNNL's assumed consumer responsiveness to
prices (prompting lower energy service demand) and
interfuel substitution potential appear to be substantially higher in the medium term (through 2010) than the
implicit elasticities in EIA's explicit representation of
technologies and consumer choices. The PNNL model
begins solving in 1985 in 5-year increments. The PNNL
reference case is calibrated to AEO98. In the PNNL policy runs, the carbon policy was phased in over a 10-year
period beginning in 2000. Consequently, policy adjustments begin in 2001, consumers and producers begin to
anticipate the Kyoto Protocol in that year, making the
appropriate adjustments. In the PNNL analysis, electricity demand grows by 0.4 percent annually in the reference case between 2010 and 2020. This is a significant
departure from the annual growth rate of more than 2
percent in recent years. Most electricity demand projections have annual growth in excess of 0.9 percent
between 2010 and 2020, as compared with PNNL's 0.4
percent.95 Offsetting these factors are factors that tend to
overstate cost. For example, in the PNNL analysis, primary renewable use for generation changes only slightly
from the reference case in 2020, even with a carbon price
of $286 per metric ton.
The group of models projecting costs between $265 per
metric ton and $295 per metric ton in 2010 for the 19907% case include transitional processes and costsÑeither
in the macroeconomy or in the energy systemÑthrough
a detailed representation of the cost, performance, and
market adoption of technologies.96 This group includes
the CRA model. Through 2010, CRA projects that, in the
reference case, U.S. GDP will grow by $270 billion more
than projected in most of the other studies compared.
The higher growth rate of GDP normally makes the
reduction in emissions harder and more costly to the
U.S. economy.
If differences in the reference cases were the only factor
accounting for the different estimates of the costs of
complying with the Kyoto Protocol, then CRA's costs
would exceed EIA's and WEFA's in 2010; however, large
econometric models of the U.S. economy like those of
WEFA and DRI tend to focus on the transitional process,
including the method of recycling any carbon fees that
may be collected by the Federal Government, and unemployment that may be increased as a result of policy
implementation. The WEFA, EIA, and DRI analyses
assume that labor can be dislocated, whereas most other
analyses assume full employment97 despite the sudden
reduction of energy resources. More aggregated world
analyses, including the CRA, PNNL, EPRI, and MIT
studies, omit such details, because the inclusion of
global regional coverage and trade flows requires simplifications (some important) in the detail with which
each region is represented. Model aggregation tends to
underestimate the macroeconomic costs; on the other
hand, a lack of global coverage (as in the EIA, DRI, and
WEFA models) may overstate transition costs, particularly if international trading is implemented efficiently.
Also, fossil fuel consumption in 2010 in the CRA analysis is about 6 quadrillion British thermal units (Btu) less
than in the EIA reference case, with virtually identical
carbon emissions levels, suggesting an accounting difference in emissions coefficients.
92 The PNNL study uses a dynamic-recursive, computable general equilibrium (CGE) model with neoclassical elements. A model is a
Ògeneral equilibriumÓ model if it represents all parts of the economy, both energy and non-energy, and all markets clear (supply equals demand at the prices determined). The model is ÒcomputableÓ if a computer is used to solve for the equilibrium; it is ÒdynamicÓ if it keeps
track of variables over time. A model is ÒneoclassicalÓ if the model structure assumes that (1) its economic agents have perfect foresight and
knowledge of all past, present, and future events, (2) there is perfect and instantaneous ability of capital and labor to move between uses and
sectors, and (3) such transitions are costless and instantaneous.
93 The PNNL model (SGM) can be run with either perfect or imperfect foresight. Labor and new capital move freely.
94 A carbon price of $221 per metric ton in 2010 would increase the delivered electricity price by 49 to 69 percent and reduce electricity
consumption by 22 percent relative to PNNL's reference case. This implies that, on average, consumers will reduce consumption of electricity by 3.2 to 4.5 percent for every 10-percent increase in the price of electricity. In 2020, a carbon price of $286 per metric ton translates to an
electricity price increase of 59 to 66 percent, resulting in a 28-percent reduction in electricity consumption. This implies that consumption
will decline by about 4.2 to 4.7 percent for every 10-percent increase in price. (The estimated electricity price changes were derived from
comparable EIA cases.)
95 For example, WEFA's annual electricity growth rate is 1.7 percent and EIA's is 0.9 percent.
96 The WEFA, CRA, MIT, and DRI models are econometric, general equilibrium, macroeconomic models. WEFA and DRI model the
United States, CRA and MIT model the world.
97 The full employment assumption means that the unemployment rate is unchanged from reference case levels.
Energy Information Administration / Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity