Economic Analysis of a Multi-Emissions Strategy

Prepared for:

Senators James M. Jeffords and Joseph I. Lieberman

U.S. Environmental Protection Agency
Office of Air and Radiation
Office of Atmospheric Programs

October 31, 2001


-------
Table of Contents

Executive Summary	ii

1.	Introduction	1

1.1.	Background	1

1.2.	Technology Scenarios	1

1.3.	Multi-Emission Targets	2

1.4.	Other Analytical Assumptions	3

2.	Multi-Emissions Analysis	4

2.1.	Modeling Technology Assumptions	5

2.1.1.	Reference Case Scenario	5

2.1.2.	Advanced Technology Scenario	5

2.1.3.	CEF Moderate Case Scenario	6

2.1.4.	CEF Advanced Technology Scenario	7

2.1.5.	Implementation of the Technology Assumptions	7

2.1.6.	Reasonableness of the Scenario Assumptions	9

2.2.	Results of the Scenario Analysis	11

2.2.1.	Emissions Projections	11

2.2.2.	Changes in Electric Generating Expenditures	12

2.2.3.	Marginal Costs	I I

2.2.4.	Fuel Use Impacts	16

2.2.5.	Energy Price Impacts	16

2.2.6.	Economy - wide Imp acts	17

2.3.	The Results in Context	18

3.	Conclusions	18

4.	References	20

5.	Appendices	22

5.1.	Description of the AMIGA Model	22

5.2.	Summary Tables for Study Scenarios	24

5.2.1.	Reference Case Proj ections	24

5.2.2.	Scenario A: Emission Constraints Using Reference Case Technologies	25

5.2.3.	Scenario B: Emission Constraints Using Advanced Case Technologies	26

5.2.4.	Scenario C: Emission Constraints Using the Moderate CEF Scenario Assumptions	27

5.2.5.	Scenario D: Emission Constraints Using the Advanced CEF Scenario Assumptions	28

l


-------
Executive Summary

In response to a May 17, 2001 request from Senators James M. Jeffords (VT) and Joseph I.
Lieberman (CT), this report describes the results of a modeling study done to evaluate the
potential impacts of reducing nitrogen oxides (NOx), sulfur dioxide (SO2), mercury (Hg), and
carbon dioxide (CO2) emissions from the US electric power sector. In their request, Senators
Jeffords and Lieberman asked the Environmental Protection Agency to undertake an economic
assessment of four technology-based scenarios designed to achieve the following emissions caps
in the US electric power sector by the year 2007:

•	Reduce nitrogen oxides (NOx) emissions to 75 percent below 1997 levels;

•	Reduce sulfur dioxide (SO2) emissions to 75 percent below full implementation of the
Phase II requirements under title IV;

•	Reduce mercury (Hg) emissions to 90 percent below 1999 levels; and

•	Reduce carbon dioxide (CO2) emissions to 1990 levels.

The request also specified that EPA should evaluate the cost of achieving these reductions using
four alternative technology scenarios:

•	The Energy Information Agency's Standard Technology Scenario.

•	The Energy Information Agency's High Technology Scenario, including technology
assumptions with earlier introduction, lower costs, higher maximum market potential, or
higher efficiencies than the Standard Scenario.

•	Two scenarios from Scenarios for a Clean Energy Future published by Oak Ridge
National Laboratory, National Renewable Energy Laboratory, and Lawrence Berkeley
National Laboratory, which include assumptions about changes in consumer behavior,
additional research and development, and voluntary and information programs.

Under each scenario, the costs of meeting the emission constraints are included in the price of
electricity. Such costs include the purchase and installation of emissions control equipment and
the purchase of emissions permits. Factors that mitigate projected cost increases include the
availability of more cost-effective, energy efficient technologies for both consumers and
electricity suppliers. EPA's analysis indicates that, under the conditions described above:

•	Electricity prices in 2015 would increase by about 32% to 50%, depending on the
technology scenario.

•	Coal-fired electric generation would decline by 25% to 35% by the year 2015.

•	Overall costs, measured by the decline in household consumption of goods and services,
would be between $13 and $30 billion annually or 0.1% to 0.3% of total consumption.
Under all four of the policy scenarios evaluated in this assessment, gross domestic
product (GDP) would remain relatively unchanged as sacrificed consumption permits
higher investment and government spending to reduce emissions.

•	Oil and gas-fired generation would be expected increase by about 8% under more
restrictive technology assumptions, but decrease by as much as 20% under scenarios that

ii


-------
embody more optimistic assumptions about energy-efficiency demand and supply
technologies.

The combination of increased prices and the availability of more energy-efficient equipment and
appliances are projected to reduce electricity demand by about 10%. With the combination of
higher prices and improved efficiency, total expenditures for electricity consumption in 2015 are
projected to increase by about 17% to 39%, depending on the scenario.

The increase in electricity prices and cost of the program, as well as the impact on the fuel mix,
varies considerably based the technology future that is assumed. For example, the 30%
electricity price increase, the $13 billion reduction in personal consumption, and the 25% decline
in coal use are all associated with the Clean Energy Future Advanced Scenario, which includes
the most optimistic technology assumptions. Likewise, the 50% electricity price increase, the
$30 billion reduction in personal consumption, and the 35% decline in coal usage are all
associated with EIA's Standard Technology Scenario.

EPA was not asked to evaluate the merits of the alternative technology scenarios. We note,
however, that they are the subject of considerable controversy. The Clean Energy Future
scenarios have been criticized on several grounds: assumed changes in consumer behavior that
are not consistent with historic behavior patterns, results from research and development funding
increases that have not occurred, and voluntary and information programs for which there is no
analytic basis for evaluating the impacts. On the other hand, supporters of those scenarios point
to economic analyses showing that the assumed investments can pay for themselves over time.
The range of estimates associated with the different technology scenarios highlights the
importance of the technology assumptions.

In conducting the modeling requested by Senators Jeffords and Lieberman, EPA has assumed
that the reductions would be achieved through a nationwide "cap-and-trade" system similar to
the Acid Rain program established under the 1990 Amendments to the Clean Air Act, together
with increasing penetration and performance of energy technologies. In accordance with the
Senators' request, the analysis also assumes the use of banked allowances made possible by early
emissions reductions achieved in the years 2002 through 2006. (In practice, significant
reductions beginning in 2002 would be difficult to achieve.) Because of the contribution of those
banked allowances to overall emissions reductions, the analysis shows emissions in 2007 above
the caps. Regardless, 2007 emissions are substantially reduced from current levels. At the end
of 2015 a small pool of banked allowances continues to be available for use in later years. The
analysis contained in the report covers the years 2002 through 2015.

The results provided in this analysis should not be construed as forecasts of actual scenario
outcomes. Rather, they are assessments of how the future might unfold compared to a
previously defined reference case — given the mix of technology and policy assumptions
embodied in each of the scenarios. The results also imply a national commitment that is
successful in achieving the level of emission reductions described within the report.

The economic impacts of the emissions reduction scenarios are evaluated using Argonne
National Laboratory's AMIGA model, a 200-sector computer general equilibrium model of the

111


-------
U.S. economy. The modular design and economy-wide coverage of the AMIGA model makes it
a logical choice to analyze alternative technology scenarios. Although it does employ the same
plant-level coverage of the electricity sector as the IPM and NEMS models used in other
analyses, the pollution control technology assumptions are not included at the same level of
detail as the IPM model. This may be particularly relevant for mercury controls, where the
effectiveness varies by coal type, and may be difficult to model correctly without additional
detail. In addition, we note that the AMIGA model is relatively new and has not been subject to
the same degree of peer-review and scrutiny as the older IPM and NEMS models. It would be
desirable in future work to establish the comparability of results across these models.

IV


-------
1. Introduction

1.1.	Background

Responding to an earlier Congressional request, the Energy Information Administration (EIA)
released a detailed study reviewing the effects of a so-called "three pollutant" strategy in
December 2000 (Energy Information Administration, 2000). The three emissions in the EIA
assessment included nitrogen oxides (NOx), sulfur dioxide (SO2), and carbon dioxide (CO2).
Although a coordinated climate and air quality policy appeared to lower costs compared to a
series of separate policy initiatives, the EIA assessment indicated significant costs associated
with capping emissions.

At about the same time, five of the nation's national energy laboratories released an extensive
review of some 50 different policy options that might achieve cost-effective reductions of both
air pollutants and carbon dioxide (CO2) emissions. The study, Scenarios for a Clean Energy
Future (Interlaboratory Working Group, 2000), indicated that domestic investments in energy-
efficient and clean energy supply technologies could achieve substantial reductions in both sets
of emissions at a small but net positive benefit for the economy.

On May 17, 2001, Senators James M. Jeffords (VT) and Joseph I. Lieberman (CT) sent a letter to
EIA and EPA seeking further clarity in the scenarios examined by the December EIA analysis,
stating that "the analysis appears to unnecessarily limit the market and technology opportunities
that might significantly affect the costs and benefits of emission reductions. In particular, the
potential contributions of demand-side efficiency, gas-fired cogeneration and of renewable
energy sources appear to be inadequately represented."

In responding to this request, EPA modeled the combined impacts of both the emissions caps and
the advanced technology scenarios specified by the Senators. We are aware that EIA has
modeled the combined impacts but has also modeled the effects of the emission caps and the
advanced technology scenarios separately. This approach provides perhaps a better technique
for isolating the actual costs of the emissions caps. We have reviewed the EIA analysis of these
separate effects and we believe that they offer interesting and important insights and that if we
had performed the same kind of analysis we would have seen similar results.

This report responds to the Senators' request. The results provided in this analysis should not be
construed as forecasts of actual scenario outcomes. Rather they are assessments of how the
future might unfold compared to a previously defined reference case — given a national
commitment to achieve the emission reductions, and given the mix of technology and policy
assumptions embodied in each of the scenarios.

1.2.	Technology Scenarios

In the letter to Administrator Whitman, Senators Jeffords and Lieberman asked for an analysis of
four different scenarios, requesting that EPA "analyze the cost and benefits, including all sectors

1


-------
of the economy and impacts on both the supply and demand side of the equation, of the
following multi-pollutant emission control scenarios for the nation's electricity generators.

Where feasible, this should include power plants both within the conventionally defined electric
utility sector as well as electricity generated by industrial cogenerators and other independent
power producers."

The four scenarios are identified as follows:

•	Scenario A: Standard Technology Scenario. Assume standard technology characteristics as
defined in AEO2001. Further assume a start date of 2002. By 2007 reduce NOx emissions

75 percent below 1997 levels, reduce SO2 emissions to 75 percent below full implementation
of the Phase II requirements under title IV, reduce mercury emissions 90 percent below 1999
levels, and reduce CO2 emissions to 1990 levels.

•	Scenario B: High Technology Scenario. Continue the 2002 start date, but assume the
advanced technology assumptions of both the supply and demand-side perspectives that are
referenced in AEO2001. By 2007 reduce NOx emissions 75 percent below 1997 levels,
reduce SO2 emissions to 75 percent below full implementation of the Phase II requirements
under title IV, reduce mercury emissions 90 percent below 1999 levels, and reduce CO2
emissions to 1990 levels.

•	Scenario C: Moderate Clean Energy Future Scenario. Continue the 2002 start date, but
assume the moderate supply and demand-side policy scenario of the Clean Energy Future
(CEF) study. By 2007 reduce NOx emissions 75 percent below 1997 levels, reduce SO2
emissions to 75 percent below full implementation of the Phase II requirements under title

IV, reduce mercury emissions 90 percent below 1999 levels, and reduce CO2 emissions to
1990 levels.

•	Scenario D: Advanced Clean Energy Future Scenario. Continue the 2002 start date, but
assume the advanced supply and demand-side policy scenario of the Clean Energy Future
study. By 2007 reduce NOx emissions 75 percent below 1997 levels, reduce SO2 emissions
to 75 percent below full implementation of the Phase II requirements under title IV, reduce
mercury emissions 90 percent below 1999 levels, and reduce CO2 emissions to 1990 levels.

In requesting an analysis of these four scenarios, the Senate request asked for ".. .results through
2020, in periods of five years or less, using the Annual Energy Outlook 2001 (AEO2001) as the
baseline."

1.3. Multi-Emission Targets

Table 1 identifies the 2007 emission caps used for each of the four scenarios. The emission cap
is defined by a benchmark emission level that is modified by the desired level (percentage) of
reduction. For example, the benchmark for the SO2 emissions cap is the Phase II requirements of
the Clean Air Act Amendments. That total, 8.95 million short tons, is reduced by a specific
percentage (75 percent) to reach the emissions cap of 2.24 million tons. Following a similar

2


-------
pattern, the remaining emission caps are set as 1.51 million tons for NOx emissions, 4.8 tons for
mercury emissions, and 475 million metric tons (MtC) of carbon emissions.

Table 1. Benchmark Emission Levels and Assumed Emission Caps

Pollutant (Benchmark)

Benchmark Emissions

Fraction Reduced

2007 Emission Cap

S02 (tons in Title IV)

8.95 million tons

75%

2.24 million tons

NOx(tons in 1997)

6.04 million tons

75%

1.51 million tons

Hg (tons in 1999)

48 tons

90%

4.8 tons

C (metric tons in 1990)

475 million metric tons

-

475 million metric tons

1.4. Other Analytical Assumptions

As previously noted, the letter from Senators Lieberman and Jeffords requested that EPA use
four different sets of technology and policy assumptions to meet the specified emission caps
shown in Table 1. The full set of technology and policy assumptions are described more fully in
section two of this report. All scenarios are implemented in 2002. At the same time, there are
other key assumptions that EPA adopted to facilitate the evaluation of the four scenarios.

In addition to the different technology scenarios, EPA was asked to include the assumption that
utilities would begin to make cost-effective emission reductions in the five years that precede the
2007 compliance date. These early reductions would be "banked" for use in the post-2007
period of analysis. For purposes of this simulation, the amount of allowances banked from 2002
through 2006 was calculated as the simple difference between the reference case projections and
the actual emission trajectory of each scenario. The decision to earn and hold early allowances is
based on the assumption that allowances are viewed as an asset that must earn at least an 8% real
return.1

Following the assumption used in the CEF study, all four of the policy scenarios assume
nationwide restructuring of the electric utility industry. This implies that prices are based on the
marginal rather than the regulated, cost-of-service pricing now used throughout much of the
country.

EPA employed the Argonne National Laboratory's AMIGA modeling system to evaluate the
impact of capping emissions under the four different technology scenarios. AMIGA is a 200
plus sector model of the U.S. economy that captures a wide variety of technology characteristics
and their resulting impact on key indicators such as emissions, employment and income.2 EPA

1	In practice, it is more likely that significant reductions that contribute to any kind of allowance bank would be
difficult to achieve before 2004. Assuming a delay in implementation to 2004 would raise the economic impact of
any of the scenarios.

2	AMIGA is especially suited to the task identifying and evaluating a different mix of technologies in the production
of goods and services within the United States. It is not only a 200 plus sector model of the U.S. economy, but it
also includes the Argonne Unit Planning and Compliance model and database that captures a wide variety of
technology characteristics within the electric generating sector, including industrial combined heat and power
systems and the typically available emission control technologies. When the electricity module is integrated with

3


-------
asked Argonne to benchmark AMIGA to the reference case projections of AEO2001. AMIGA
was then modified to approximate the assumptions behind each of the four scenarios.

An economic analysis of a policy compares the world with the policy (the policy scenario) to the
world absent the policy (the reference case or baseline scenario). The impacts of policies or
regulations are measured by the resulting differences between these two scenarios. In effect, any
meaningful analysis should compare the full set of benefits and costs to the extent possible.

For purposes of this exercise, there are at least seven categories of costs and four benefits that
might be reviewed. The costs include: (1) direct investment costs, (2) operating and maintenance
costs, (3) research and development and other government program costs, (4) transaction, search,
and compliance costs, (5) adjustment costs associated with large changes in specific capital
stocks, (6) lost economic flexibility created by additional emission requirements, and (7)
potential interactions with the existing tax system. At the same time, there are at least four
categories of benefits. These include: (1) direct savings from lower compliance costs, (2)
process efficiency and other productivity gains, (3) environmental and health benefits not
captured within normal market transactions, and (4) spillovers and/or learning induced by either
the technology investment, or the R&D efforts.

The costs associated with the emission limits in each scenario are computed as the increased
expenditures on pollution control, investment in more efficient equipment and appliances,
research and development, tax incentives, and additional government programs — all relative to
the reference case. The increased costs are coupled with credits for reductions in fuel use and
productivity gains from technology. The economic impact of each scenario is reported in two
ways. The first is as a change in household personal consumption, measuring the goods and
services available for consumers to enjoy after subtracting these net expenditures. The second is
as a change in economic output measured as Gross Domestic Product (GDP).

The AMIGA model reasonably captures those costs and benefits noted above that arise in market
transactions. Some, such as loss of flexibility and adjustment costs on the cost side, and health
benefits and spillovers on the benefit side, remain beyond the scope of this analysis.

2. Multi-Emissions Analysis

This section provides additional details about the technology assumptions that underpin the four
emission scenarios. It also describes the results of the scenario analysis, both in terms of the
various marginal costs associated with emission control strategies and the economy-wide impact
of each scenario. Although EPA made every effort to calibrate AMIGA to the AEO2001
reference case, AMIGA is a different modeling system than EIA's National Energy Modeling
System (NEMS). Hence, it was not possible to reproduce the exact AEO2001 reference case

the larger macroeconomic system, the model can then generate key outputs including projected electricity sales and
net generation, resulting emissions for each of the four pollutants under consideration, and the set of energy and
permit prices associated with the resulting production levels. Finally, AMIGA can provide an estimate of the
consequent impact on the economy including key indicators as consumption, investment, government spending,
GDP, and employment (Hanson, 1999). For more background on the AMIGA model, see Appendix 5.1.

4


-------
projections. Moreover, Argonne researchers recently upgraded AMIGA to incorporate SO2,
NOx, and mercury emissions. For this and other reasons, AMIGA currently reports results only
through the year 2015. Nonetheless, the differences in the resulting baseline projections are
minor for the purposes of this analysis.

2.1. Modeling Technology Assumptions

Scenarios A and B are based on the AEO2001 standard and advanced technology characteristics,
respectively. The standard technology assumptions of scenario A were used by EIA in the
development of the AEO2001 "reference case" projections. The advanced technology
assumptions of scenario B were used as a sensitivity analysis in the AEO2001. They
demonstrated the effects of earlier availability, lower costs, and/or higher efficiencies for more
advanced equipment than the reference case.3

Scenarios C and D are based on the recently published DOE-sponsored report, Scenarios for a
Clean Energy Future (Interlaboratory Working Group, 2000; see also, Brown, et al, 2001). Both
of the CEF scenarios assumed nationwide restructuring of the electric utility industry. From an
analytical perspective, this means that prices are based on the marginal costs of generation,
transmission and distribution of electricity rather than the regulated, cost-of-service pricing now
used throughout much of the country. Moreover, both scenarios reflected increased spending for
research and development and other programs designed to accelerate the development and
deployment of low-carbon, energy efficient technologies. Each of the scenario assumptions are
described more fully in the sections that follow.

2.1.1.	Reference Case Scenario

The scenario A reference case assumes a "business-as-usual" characterization of technology
development and deployment. As projected in the AEO2001 assessment, the nation's economy
is projected to grow at 2.9% per year in the period 2000 through 2020. Given anticipated energy
prices and the availability of standard technologies, the nation's primary energy use is expected
to grow 1.3% annually while electricity consumption is projected to increase by 1.8% annually.
Further details are provided in Appendix 5.2.1.

2.1.2.	Advanced Technology Scenario

Under the AEO 2001 advanced technology characterization, scenario B assumes that a large
number of technologies have earlier availability, lower costs, and/or higher efficiencies. For
example, the high efficiency air conditioners in the commercial sector are assumed to cost less
than in scenario A. This encourages a greater rate of market penetration as electricity prices rise
in response to the emissions caps. Building shell efficiencies in scenario B are assumed to
improve by about 50 percent faster than in scenario A.

3 The AEO2001 was published in December 2000 (Energy Information Administration, 2000).

5


-------
On the utility's side of the meter, the heat rates for new combined cycle power plants are
assumed to be less compared to the standard case assumptions. This means that more kilowatt-
hours of electricity are generated for every unit of energy consumed by the power plants.

Moreover, wood supply increases by about 10% and the capacity factor of wind energy systems
increases by about 15-20% compared to the reference case assumptions. In the AEO2001 report,
the combination of higher efficiencies and earlier availability of the technologies lowers the
growth in electricity use from 1.8% in the reference case to 1.6%.

2.1.3. CEF Moderate Case Scenario

The authors of the Clean Energy Future (CEF) report describe their analysis as an attempt to
"assess how energy-efficient and clean energy technologies can address key energy and
environmental challenges facing the US" (Brown, et al, 2001). In that regard, they evaluated a
set of about 50 policies to improve the technology performance and characterization of the
residential, commercial, industrial, transportation, and electricity generation sectors. The
policies include increased research and development funding, equipment standards, financial
incentives, voluntary programs, and other regulatory initiatives. These policies were assumed to
change business and consumer behavior, result in new technological improvements, and expand
the success of voluntary and information programs.

The selection of policies in the CEF study began with a sector-by-sector assessment of market
failures and institutional barriers to the market penetration of clean energy technologies in the
US. For buildings, the policies and programs include additional appliance efficiency standards;
expansion of technical assistance and technology deployment programs; and an increased
number of building codes and efficiency standards for equipment and appliances. They also
include tax incentives to accelerate the market penetration of new technologies and the
strengthening of market transformation programs such as Rebuild America and Energy Star
labeling. They further include so-called public benefits programs enhanced by electricity line
charges.

For industry, the policies include voluntary agreements with industry groups to achieve defined
energy efficiency and emissions goals, combined with a variety of government programs that
strongly support such agreements. These programs include expansion and strengthening of
existing information programs, financial incentives, and energy efficiency standards on motors
systems. Policies in the CEF analysis were assumed to encourage the diffusion and improve the
implementation of combined heat and power (CHP) in the industrial sector. For electricity, the
policies include extending the production tax credit of 1.5 cents/kWh over more years and
extending it to additional renewable technologies.

Broadly speaking, the CEF Moderate scenario can be thought of as a 50% increase in funding for
programs that promote a variety of both demand-side and supply-side technologies. For
example, the moderate scenario assumes a 50% or $1.4 billion increase in cost-shared research,
development, and demonstration of efficient and clean-energy technologies (in 1999 dollars with
half as federal appropriations and half as private-sector cost share). It further assumes a careful
targeting of funds to critical research areas and a gradual, 5-year ramp-up of funds to allow for

6


-------
careful planning, assembly of research teams, and expansion of existing teams and facilities. In
addition, the CEF moderate scenario anticipates increased program spending of $3.0 and $6.6
billion for the years 2010 and 2020, respectively. These expenditures include production
incentives and investment tax credits for renewable energy, energy efficiency and transportation
technologies. They further include increased spending for programs such as DOE's Industrial
Assessment Centers and EPA's Energy Star programs.

The combined effect of the R&D and program expenditures, together with other policies
described in the CEF report, implies a steady reduction in total energy requirements over the
period 2000 through 2020. By the year 2020, for example, primary energy consumption and
electricity sales were projected to decrease by 8% and 10%, respectively, compared to the CEF
reference case.

2.1.4.	CEF Advanced Technology Scenario

Building on the policies of the moderate scenario, the CEF advanced scenario assumes a
doubling of cost-shared R&D investments, resulting in an increased spending of $2.9 billion per
year (again, in 1999 dollars with half as federal appropriations and half as private-sector cost
share). In addition, the advanced scenario anticipates increased program spending of $9.0 and
$13.2 billion for the years 2010 and 2020, respectively. The added spending covers all sectors
including buildings, industry, transportation, and electric generation.

The combined effect of the program and R&D expenditures, together with other policies
described in the CEF report (including a $50 carbon charge applied in the CEF Advanced
Scenario), drove a steady reduction in the need for energy compared to the CEF reference case.
By 2020 total energy use fell by 19% compared to the reference case. At the same time,
electricity sales in 2020 were projected to decrease by 24% compared to the CEF reference case.

2.1.5.	Implementation of the Technology Assumptions

The assumptions embedded in each of these scenarios have the effect of progressively increasing
market penetration of higher performance energy efficiency and energy supply technologies. As
shown in Table 2, the net effect of these assumptions is to lower the expected level of electricity
consumption while continuing to meet the same level of service demanded by utility customers.
The technology assumptions also have the effect of increasing the availability of cleaner energy
supply technologies that reduce the level of emissions per kilowatt-hour of generation. The
critical assumption used in the EPA analysis is that program spending affects both supply and
demand technologies in a way that interacts with the emission caps that are to be imposed in
2007.

Benchmarked to the year 2010, Table 2 shows the percentage change of key indicators for each
scenario with respect to its respective reference case. These changes provide EPA with
approximate targets so that each of the scenarios can be mapped into the AMIGA model. As
such, the figures in Table 2 should be seen as inputs into the AMIGA model, not outputs of the
model.

7


-------
Table 2. Influence of Technology Assumptions on Key Scenario Indicators - 2010

Indicator

Scenario A
Standard
Technology Case

Scenario B
Advanced
Technology Case

Scenario C
CEF Moderate
Case

Scenario D
CEF Advanced
Case

Primary Energy

0%

-2.5%

-3.4%

-6.3%

Electricity Sales

0%

-2.4%

-5.9%

-6.8%

Carbon Emissions

0%

-5.0%

-7.4%

-10.7%

NOx Emissions

0%

-2.6%

-5.4%

-8.1%

By definition, scenario A assumes the standard technology assumptions of the AEO2001
reference case. Hence, there are no additional programs or policies that generate changes in the
reference case technologies when the emission caps are imposed by the year 2007. The level of
technology responsiveness grows for scenarios B, C, and D as a result of greater program
spending.

The CEF advanced scenario, for example, assumes a significant increase in program funds to
promote a variety of both demand-side and supply-side technologies. As a result of this greater
level of program activity, there is an accelerated penetration of energy-efficient technologies that
drives electricity sales down by 6.8 percent in 2010 (compared to the CEF reference case for that
same year). At the same time, the combination of a lower demand for electricity and an
increased investment in cleaner energy supply technologies reduces both carbon and NOx
emissions by 10.7 and 8.1 percent, respectively (again, compared to the CEF 2010 reference
case). As EPA modeled this scenario, the bundle of policies in the CEF advanced scenario
became, in effect, a complement to the emission caps imposed by 2007.

To avoid overestimating the impact of the policy scenarios in this analysis, EPA made a number
of adjustments before implementing the CEF assumptions in the four scenarios reported here.

First, the CEF analysis was benchmarked to a 1999 reference case. In the AEO2001 reference
case, however, the demand for electricity in 2020 is about 10% higher compared to the CEF
reference case. Second, the Senate request asked EPA to assume a 2002 start date in running the
technology and policy scenarios. In effect, there are fewer years in which programs can achieve
the desired level of technology improvement compared to the CEF scenarios. In addition, the
CEF analysis includes a significant review of transportation technologies and policies. EPA
chose to exclude all assumptions related to transportation, focusing only on the supply and
demand-side technologies associated with electricity and natural gas consumption.

With the adjustments described above now reflected in the current analytical framework, and
using the program cost information documented in the CEF study, Table 3 summarizes the
incremental program costs that were assumed as necessary to drive the kind of changes in
electricity consumption and emissions described in Table 2. Since transportation programs
drove a significant part of the CEF expenditures, and since there are fewer years to implement
policies, the estimated program expenditures are also smaller compared to the CEF assumptions.

8


-------
Table 3. Incremental Policy Costs of the Technology Scenarios (billion 1999 dollars)

Scenario

2002

2005

2010

2015

Scenario A

0.0

0.0

0.0

0.0

Scenario B

0.8

1.6

2.7

2.9

Scenario C

1.2

2.3

4.3

4.8

Scenario D

2.1

3.9

5.2

5.5

Because scenario A characterizes existing program and technology performance, no additional
funds are required to drive that scenario. Scenario B, on the other hand, anticipates some
changes in the technology characterization that will affect the electricity sector as shown in Table
2. While the AEO2001 analysis anticipated no program spending to drive these changes, EPA
assumed that additional spending would be required for scenario B. Calibrating to the CEF
policy scenarios, EPA estimated that program and policy spending would increase by $0.8
billion in 2002, rising steadily to $2.9 billion by 2015. For scenario C, program spending
increased by $1.2 billion starting in 2002, rising to $4.8 billion by 2015. Finally, program
spending in scenario D started at $2.1 billion in 2002 and increased to $5.5 billion by the last
year of this analysis.4

The net effect of mapping increased program spending together with adjustments needed to
update the assumptions of the CEF policy scenarios can be highlighted by reviewing the change
in electricity generation for scenario D. In the CEF Advanced Scenario (based on a 1999
reference case), for example, the level of electricity generation in 2010 was lowered by 10%
from the reference case requirements of 3,920 billion kilowatt-hours (kWh). As the CEF
technology assumptions were applied in scenario D within this analysis (updated to the
AEO2001 reference case), electricity generation was reduced by 9% from 4,253 billion (kWh).
The trend was more pronounced in 2015. Rather than a roughly 16% reduction from a
generation level of 4,200 billion kWh in the 1999 CEF Advanced Case, the scenario D
equivalent in this analysis achieved only a 12% reduction from a generation of 4,580 billion
kWh.

2.1.6. Reasonableness of the Scenario Assumptions

The results of the technology-driven scenarios should not be interpreted as an EPA endorsement
of any of the policies or technology assumptions behind each of scenarios described in this
report. On the one hand, EPA has not conducted any significant review of the EIA assumptions
that underpin the AEO2001 projections. On the other hand, some analysts do not necessarily
agree with the assumptions and projected level of impacts in the CEF assessment despite the fact
that it was peer-reviewed and its findings published this fall in an academic journal. The EIA
(2001), for example, notes that the CEF policies assume changes in consumer behavior that are
not consistent with historically observed behavior patterns. Moreover, the EIA suggests that
there is little documentation to support the assumed technological improvements generated by
the research and development (R&D) initiatives described in the report. Finally, EIA notes that

4 The program spending assumptions developed in this analysis are used only to approximate the impact of the CEF
scenarios. They do not reflect EPA endorsement of these spending levels.

9


-------
the effectiveness of voluntary or information programs may be less than assumed in the CEF
scenarios. At the same time, the lead CEF analysts have responded to the EIA assertions by
citing relevant economic literature and noting that the CEF study is one of "the most carefully
documented and complete analysis of U.S. energy futures that has ever been funded by the U.S.
government" (Koomey, et al, 2001).

Notwithstanding these concerns, EPA attempted to respond to the Senators' request by mapping
in the critical assumptions of the CEF as a range of policies that provide a set of alternative
assumptions about the future. In this regard, the scenarios are more like descriptions of
alternative future outcomes rather than predictions or recommendations about how the future
should unfold.

To provide a more complete context for understanding the magnitude of the changes in
electricity generation that are suggested by the different scenarios, the figure below illustrates
both the historical and projected trends in the nation's electricity generation. The information is
shown as the number of kWh per dollar of GDP (measured in constant 1999 dollars). The
historical data covers the period 1970 through 2000 while the projected trends are through the
year 2015. The historical period shows a moderate level of volatility. The reference case
projections suggest an annual rate of declining intensity of 1.6% per year through 2015 with a
final value 0.33 kWh/$.

Historical and Projected US Electricity Trends (kWh per 1999 $ GDP)

Historical	Reference Case	 Advanced Case

In comparison to the reference case, Scenario D (adapting the CEF Advanced Case assumptions)
reflects a national commitment to improve both electricity supply and the efficiency of demand-
side technologies. The presumption is that such a commitment would be supported by a
significant increase in R&D and program spending as described above. Under these
assumptions, the nation's electricity intensity is projected to decline at an annual rate of 2.5%,

10


-------
dropping to a final intensity of 0.28 kWh/$. This level of decline is greater than previously seen
in the recent past. In the period 1980 through 1986, for example, and again 1993 through 2000,
the annual rate of decline was only 1.7 percent. Hence, it appears that the assumptions driving
the advanced scenario are aggressive. At the same time, however, the research undertaken by the
CEF analysts indicates that the technology is available to achieve such a reduction should a
national commitment be successful in driving similar policies.

2.2. Results of the Scenario Analysis

With the model benchmarked to AEO2001, and given the different mix of scenario assumptions
previously described, AMIGA reports the results in the figures and tables that follow. More
complete data, including reference case assumptions, are available in Appendix 5.2.

2.2.1. Emission Projections

All program and policy assumptions have a start date of 2002. Moreover, the analysis
anticipates the use of banked allowances made possible by early emissions reductions achieved
in the years 2002 through 2006 (as requested in the Senate letter). Figures 1 through 4 on the
following page illustrate both the emissions projections and the impact of banking the early
reductions on all four emissions caps implemented in 2007.

Although all four categories of emissions are down substantially, they only achieve 50-75% of
the proposed cap by 2007 (shown as the dotted horizontal line in each of the above figures).

This is because of the availability of the banked allowances that can be used by sources to meet
emissions caps in 2007 and beyond. Note that costs would be noticeably higher if power plants
were required to actually hit the target in 2007. In 2015, carbon and mercury emissions continue
to be 15% or more above the target.

The reductions that generate the banked allowances are shown as the area to the left of each
vertical dotted line as the differences between the reference case and scenario emission
trajectories. The emissions above the cap are shown to the right of each vertical dotted line and
between the scenario emissions and the dotted horizontal line. Subtracting these two areas on
each graph reveals the level of the bank in 2015. Using Scenario D as an example, the remaining
allowances in 2015 are 100 million metric tons for carbon, 1.3 million tons for SO2, 0.2 million
tons for NOx and 25 tons for mercury. In the case of carbon, the bank would last another two
years at the rate of drawdown in 2015, or longer if the drawdown declined.

11


-------
Figure 1. Carbon Emissions (million metric tons)

Figure 2. S02 Emissions (million tons)

" firsnft	firsnR

2010	2015

E ffrfin E "*~Srfin [J

Figure 3. NOx Emissions (million tons)

Figure 4. Mercury Emissions (tons)

"firsnft	firfinR

2.2.2. Changes in Electric Generation Expenditures

Given the assumptions and economic drivers in each of the scenarios, the AMIGA model
calculates the capital investment, operation and maintenance, and fuel costs necessary to meet
consumer demand for electricity. The incremental expenditures required to generate electricity
under each of the four scenarios as compared to the reference case are summarized in Figure 5
(in billions of 1999 dollars). In effect, the incremental expenditures reflect the range of decisions
made by the electricity sector to comply with each of the four scenario constraints—but do not
reflect efforts made outside the electricity sector. Because these expenditures ignore spending on
energy efficiency, research and development outside the electricity sector—spending that can be
substantial—they are not measures of program costs. Note that incremental expenditures are
incurred as early as 2002 in all four scenarios to generate early reductions that can be banked for
use in 2007 and beyond.

The generation expenditures vary in each of the scenarios change for at least three reasons: (1)
the size of the allowance bank made possible by early reductions driven, in part, by program
spending prior to the introduction of the caps; (2) the varying levels of demand for electricity

12


-------
over time, resulting in changes in the overall mix of generation resources; and, (3) the gradual
reduction in the banked allowances available for withdrawal necessitating additional actions to
reduce emissions.

As expected, scenario A has the largest increase with expenditures rising by nearly $17 billion in
2015 compared to the reference case. The higher level of expenditures is driven by a 21%
increase in unit generation costs caused primarily by the emissions caps and offset only slightly
by a small decrease in electricity demand. With less energy efficiency technology penetrating
the market, a greater level of control equipment must be installed and operated which, in turn,
drives up the cost of generation. Scenario B follows a similar pattern with expenditure increases
being offset by further reductions in electricity demand as more efficient technology penetrates
the market. The expenditures for scenario C decline even further as reduced demand continues
to lower both the level generation and the unit cost of that generation compared to scenario A.
Scenario D, on the other hand, actually shows a decline in total expenditures by 2015. The
combination of a 12.5% reduction on generation load together with only an 11.9% increase in the
unit cost of generation (both with respect to the reference case) results in a $3.11 billion
reduction in total electric generation expenditures.

Figure 5. Incremental Expenditures on Electric Generation (Billions of 1999$)

Seen A	""""Seen B	~**~Scen C	~*~Scen D

13


-------
2.2.3. Marginal Costs

The marginal costs of emission reductions over the period 2005 through 2015 are shown in
Figures 6 through 9 for all four scenarios.

Figure 6. Projected Marginal Cost of Carbon Reductions ($/Metric Ton)

| —^Scen A	^KScenB	Seen C	^^ScenD |

Figure 7. Projected Marginal Cost of SO2 Reductions ($/Ton)

Figure 8. Projected Marginal Cost of NOx Reductions ($/Ton)

~^~ScenA^^_^^_^^ScenB^^_^^_*ScenC^^_^^^^^ScenD

14


-------
Figure 9. Projected Marginal Cost of Hg Reductions ($Million/Ton)

The marginal cost of carbon reductions range from $46 to $138/metric ton through 2015 with
each scenario showing successively smaller costs as technology characteristics improve and
more energy-efficient and/or low carbon technologies penetrate the market. The marginal cost of
SO2 and NOx reductions through 2015 are less than $450/ and $2,300/ton, respectively, in all
four multi-emissions reduction scenarios. The marginal cost of mercury reductions by 2015
ranges from $350 million/ton to $432 million/ton, again depending on the scenario.

It is important to note that marginal cost reflects the additional cost of one more ton of
reductions, and not the total cost associated with each pollutant. One can make a very rough
estimation of this overall cost for each pollutant, on top of the costs associated with the other
three, by multiplying half the marginal cost (to approximate average cost) by the volume of
reductions. By 2015, as an example, scenario A returns cost estimates of $15.2 billion for
carbon, $1.1 billion for SO2, $2.7 billion for NOx, and $6.4 billion for mercury. In Scenario D,
the cost estimates are $8.6 billion for carbon, $1.6 billion for SO2, $3.3 billion for NOx, and $7.8
billion for mercury. Note that these figures cannot be added together for an overall estimate
because they (a) double count the benefits of controlling multiple pollutants simultaneously, and
(b) ignore the consequences of the underlying technology policy. We discuss overall costs
below.

Surprisingly, the marginal cost of SO2. NOx, and Hg reductions increases as the marginal cost of
carbon decreases. The reason appears to be that as efficiency technology penetrates the market
and reduces carbon prices, more of a price signal is required to generate further reductions in the
three conventional pollutants. In the advanced scenarios, for example, both demand reductions
and the increased use of gas tends to reduce carbon emissions. But gas prices begin to rise which
allows coal to make a modest comeback with respect to scenario A. This is especially true as
cleaner and more efficient coal technologies begin to penetrate the market as assumed in

15


-------
scenarios B through D. In order to offset the tendency for coal-generated emissions to increase,
permit prices need to adjust upward.

2.2.4. Fuel Use Impacts

Figure 10 shows both total electricity consumption and the fossil fuel consumption used in the
generation of electricity for the year 2010. The results are in quadrillion Btu in both the
reference case and each of the four policy scenarios. As each successive scenario generates a
greater reduction in electricity demand, coal use is reduced significantly (by about 30 percent).
Gas consumption increases slightly in scenarios A and B, and decreases by a small amount in
scenarios C and D as lower electricity consumption reduces the need for new capacity.

Figure 10. Total Electricity Consumption and Fossil Fuel Generation in 2010 (Quadrillion Btu)

O

= 25 •

O 20 •

Seen A
I ~ Total

Seen C
Ppil/Gas I

2.2.5. Energy Price Impacts

The model suggests that under the conditions described above, electricity prices are expected to
increase by about 30% (under scenario D) to 50% (under scenario A) by the year 2015. This is
the logical result of increased control costs and permit prices. The combination of increased
prices and the availability of more energy-efficient equipment and appliances reduce electricity
demand by about 10%. Total electricity expenditures increase by about 15% to 30% depending
on the year and the scenario (see Table 3, below, and the tables in Appendix 5.2 for more detail
on the changing pattern of expenditures).

16


-------
2.2.6. Economy-wide Impacts

Table 3 provides a summary of key macroeconomic data for the year 2010 to compare the impact
of emissions reductions on both personal consumption and other components of gross domestic
product (GDP). The effects on personal consumption show a decline of between $13 billion and
$31, or 0.1% to 0.3%, depending on the scenario. This reflects the cost of the program in terms
of the decreased well being of households who must forego a fraction of their consumption of
goods and services in order to pay for both research and development programs, energy
efficiency improvements, and more expensive electricity production. Table 3 shows little change
in GDP under any of the policy scenarios, reflecting the fact that this foregone consumption turns
up as expenditures in other categories of GDP, namely, investment and government spending.5

Table 3. Summary of Economic Impacts by Scenario - 2010

Analytical Scenario

Electricity
End Use
Demand
(Billion
Kilowatt-
hours)

Natural Gas

Use in
Electricity
Generation
(Quads)

Coal Use in
Electricity
Generation
(Quads)

Electricity
Expenditures
(Billion 1999
Dollars)

Personal
Consumption
(Billion 1999
Dollars)

Investment
(Billion 1999
Dollars)

Gross
Domestic
Product
(Billion 1999
Dollars)

Reference

4,346

8.3

22.3

269.4

8,902.0

3,042.4

13,211.7

A. Standard Tech

4,156

9.3

14.6

353.9

8,870.9

3,067.3

13,204.3

B. High Tech

4,112

8.9

15.0

337.4

8,873.7

3,067.0

13,209.5

C. Mod CEF

4,070

8.2

15.6

323.0

8,881.7

3,066.8

13,218.9

D. Adv CEF

4,025

7.7

15.9

308.9

8,889.2

3,066.7

13,227.2

The AMIGA modeling system reports the costs and benefits of each scenario with several major
exceptions. The first omitted benefit is spillover and productivity gains beyond energy bill
savings. A number of studies suggest that energy efficiency technology investments also tend to
increase overall productivity of the economy, especially in the industrial sector. (Sullivan, et al.,
1997; Finman and Laitner, 2001; and Laitner, et al, 2001). To date, however, no systematic
effort has been undertaken to incorporate such benefits into the current generation of policy
models. Hence, this potential benefit is not reported at this time. The second missing benefit
includes gains in environmental quality, especially improved health benefits.

On the cost side, the model ignores costs associated with rapid changes in capital stocks, as well
as potential loss of flexibility and interactions with the existing tax system. For example, the
model forecasts significant changes in the level and composition of electricity generation in
2002, ignoring the difficulty of rapidly changing the capital stock by then end of 2001. Losses in
flexibility occur when pollution control activities potentially interfere with efficiency and other
operational programs at a regulated facility. Finally, there are interactions with the tax system
when, in response to a rise in the relative cost of purchased goods, people decide to enjoy more

5 A more complete assessment of each policy scenario can be made by reviewing the more detailed data contained in
the Appendix.

17


-------
leisure (which is now relatively less expensive), work less, and lower taxable income (Parry and
Oates, 2000).

2.3. The Results in Context

Recent studies suggest significant economic consequences as a result of substantial emission
reduction strategies (EPRI, 2000; and EIA, 2000). On the other hand, the presumption of a
trade-off between environmental and economic benefits may not provide an entirely appropriate
framework for analysis of such policies (DeCanio, 1997). Indeed, there are a number of studies
that show net economic benefits may be possible when a full accounting of both benefits and
costs are included within an appropriate analysis (Krause, et al, 2001; and Bailie, et al, 2001).

At the same time, understanding the proper characterization and role of technology
improvements (Edmonds, et al, 2000), and then capturing that characterization within an
appropriate model structure (Peters, et al, 2001), is a critical aspect of all such economic
assessments.

Finally, it is important to recognize that the mere existence of technologies and the potential for
positive net benefits does not assure that these technologies will be commercialized and adopted,
nor that the net benefits will be realized (Jaffe, et al, 2001). An unanswered question is whether
and how policies might encourage these activities.

This current study, while drawing on credible data sources and applying a state-of-the-art
modeling system, cannot adequately capture all such nuances associated with emission reduction
scenarios. The results of this analysis should be viewed within this larger context.

3. Conclusions

The analysis suggests that under the conditions described above, emissions through 2015 will be
significantly reduced although they won't meet the 2007 target. This is largely because of
assumptions about the banking of allowances earned prior to 2007. At the same time, coal-fired
electric generation is expected to decline by 25% to 35% by the year 2015. On the other hand,
oil and gas-fired generation is projected to increase by about 8% under more restrictive
technology assumptions, but decrease by as much as 20% under scenarios that embody more
optimistic assumptions about energy-efficiency demand and supply technologies. Electricity
prices are expected to increase by 32% to 50% in 2015, depending on the scenario.

The combination of increased prices and the availability of more energy-efficient equipment and
appliances are projected to reduce electricity demand by about 10% compared to the reference
case. With the combination of higher prices and improved efficiency, total expenditures for
electricity consumption in 2015 are projected to increase by about 17% to 39% depending on the
scenario. Interacting with other changes in consumer and business spending that is driven by
each of the scenario assumptions, the personal consumption reduced by about 0.1% to 0.3%.
This again depends on the year and the scenario.

18


-------
The results provided in this analysis should not be construed as forecasts of actual scenario
outcomes. Rather they are assessments of how the future might unfold compared to a previously
defined reference case — given the mix of technology and policy assumptions embodied in each
of the scenarios. The results from these scenarios imply a strong national commitment, one that
is successful in developing the programs and policies necessary to achieve the level of emission
reductions described within the report.

19


-------
4. References

Alison, Bailie, Stephen Bernow, William Dougherty, Michael Lazarus, and Sivan Kartha, 2001.
The American Way to the Kyoto Protocol: An Economic Analysis to Reduce Carbon Pollution,
Tellus Institute and Stockholm Environment Institute, Boston, MA, July, 2001.

Brown, Marilyn A., Mark D. Levine, Walter Short, and Jonathan G. Koomey, 2001. "Scenarios
for a clean energy future," Energy Policy Vol. 29 (November): 1179-1196, 2001.

DeCanio, Stephen J., 1997. "Economic Modeling and the False Tradeoff Between
Environmental Protection and Economic Growth," Contemporary Economic Policy, Vol. 15
(October): 10-27, 1997.

Edmonds, Jae, Joseph M. Roop, and Michael J. Scott, 2000. Technology and the economics of
climate change policy, Pew Center on Global Climate Change, Washington, DC, September
2000.

E-GRID, 2000. Emissions & Generation Resource Integrated Database, US Environmental
Protection Agency, Washington, DC, http://www.epa.gov/airmarkets/egrid/factsheet.html.

Electric Power Research Institute, 2000. Energy-Environment Policy Integration and
Coordination Study, TR-1000097, Palo Alto, CA, 2000.

Energy Information Administration, 1998. Impacts of the Kyoto Protocol on U.S. Energy
Markets and Economic Activity, SR/OIAF/98-03, Washington, DC, October 1998.

Energy Information Administration, 2000. Analysis of Strategies for Reducing Multiple
Emissions from Power Plants: Sulfur Dioxide, Nitrogen Oxides, and Carbon Dioxide,
SR/OIAF/2000-05 (Washington, DC, December 2000).

Energy Information Administration, 2001. Analysis of Strategies for Reducing Multiple
Emissions from Electric Power Plants with Advanced Technology Scenarios, SR/OIAF/2001-05
(Washington, DC, October 2001).

Finman, Hodayah, and John A. "Skip" Laitner, 2001. "Industry, Energy Efficiency and
Productivity Improvements," Proceedings of the ACEEE Industrial Summer Study, American
Council for an Energy-Efficient Economy, Washington, DC, August 2001.

Hanson, Donald A, 1999. A Framework for Economic Impact Analysis and Industry Growth
Assessment: Description of the AMIGA System, Decision and Information Sciences Division,
Argonne National Laboratory, Argonne, IL, April, 1999.

Interlaboratory Working Group, 2000. Scenarios for a Clean Energy Future, ORNL/CON-476
and LBNL-44029 Oak Ridge, TN: Oak Ridge National Laboratory; Berkeley, CA: Lawrence
Berkeley National Laboratory, November 2000.

20


-------
Jaffe, AB, RN Newell, and RN Stavins, 2001. "Energy-efficient technologies and climate change
policies: Issues and evidence." In Climate Change Economics and Policy: An RFF Anthology,
edited by MA Toman. Washington: Resources for the Future.

Jeffords, James, and Joseph Lieberman, 2001. "Letter to EPA Administrator Christine Todd
Whitman," May 17, 2001.

Koomey, Jonathan, Alan Sanstad, Marilyn Brown, Ernst Worrell, and Lynn Price, 2001.
"Assessment of EIA's statements in their multi-pollutant analysis about the Clean Energy
Futures Report's scenario assumptions," Memo to EPA's Skip Laitner, Lawrence Berkeley
National Laboratory, Berkeley, CA, October 18, 2001.

Krause, Florentin , Paul Baer, and Stephen DeCanio, 2001. Cutting Carbon Emissions at a
Profit: Opportunities for the U.S., International Project For Sustainable Energy Paths, El Cerrito,
CA, May 2001.

Laitner, John A. "Skip", Ernst Worrell, and Mchael Ruth, 2001. "Incorporating the Productivity
Benefits into the Assessment of Cost-effective Energy Savings Potential Using Conservation
Supply Curves," Proceedings of the ACEEE Industrial Summer Study, American Council for an
Energy-Efficient Economy, Washington, DC, August, 2001.

Parry, I.W.H. and W.E. Oates. "Policy Analysis in the Presence of Distorting Taxes" Journal of
Policy Analysis and Management 19(4), pp 603-613.

Peters, Irene, Stephen Bernow, Rachel Cleetus, John A. ("Skip") Laitner, Aleksandr Rudkevich,
and Mchael Ruth, 2001. "A Pragmatic CGE Model for Assessing the Influence of Model
Structure and Assumptions in Climate Change Policy Analysis," Presented at the 2nd Annual
Global Conference on Environmental Taxation Issues, Tellus Institute, Boston, MA, June 2001.

Sullivan, Gregory P., Joseph M. Roop, and Robert W. Schultz, 1997. "Quantifying the Benefits:
Energy, Cost, and Employment Impacts of Advanced Industrial Technologies," 1997 ACEEE
Summer Study Proceedings on Energy Efficiency in Industry, American Council for an Energy-
Efficient Economy, Washington, DC, 1997.

US Environmental Protection Agency, 2000b. Guidelines for Preparing Economic Analysis,
EPA-240-R-00-003, Office of the Administrator, Washington, DC, September 2000.

21


-------
5. Appendices

5.1. Description of the AMIGA Model

The All Modular Industry Growth Assessment (AMIGA) model is a general equilibrium
modeling system of the U.S. economy that covers the period from 1992 through 2030.6 It
integrates features from the following five types of economic models:

1).	Multisector - AMIGA starts by benchmarking to the 1992 Bureau of Economic Analysis
(BEA) interindustry data, which a preprocessor aggregates to approximately 300 sectors;

2).	Explicit technology representation - AMIGA reads in files with detailed lists of technologies
(currently with a focus on energy-efficient and low-carbon energy supply technologies, including
electric generating units) containing performance characteristics, availability status, costs,
anticipated learning effects, and emission rates where appropriate;

3).	Computable General Equilibrium - AMIGA computes a full-employment solution for
demands, prices, costs, and outputs of interrelated products, including induced activities such as
transportation and wholesale/retail trade;

4).	Macroeconomic - AMIGA calculates national income, Gross Domestic Product (GDP),
employment, a comprehensive list of consumption goods and services, the trade balance, and net
foreign assets and examines inflationary pressures;

5).	Economic Growth - AMIGA projects economic growth paths and long-term, dynamic effects
of alternative investments including accumulation of residential, vehicle, and producer capital
stocks.

In addition, the AMIGA system includes the Argonne Unit Planning and Compliance model that
captures a wide variety of technology characteristics within the electric generating sector. This
includes a system dispatch routine that allows the retirement and the dispatch of units on the
basis of traditional cost criteria as well as the impact of various permit prices on operating costs.
It also includes non-utility generation sources such as industrial combined heat and power
applications and renewable energy systems.

Climate change mitigation policy has been the main application of the AMIGA system to date.
But the AMIGA modeling system recently has been enhanced to include policies involving the
reduction of sulfur dioxide, nitrogen oxide, and mercury emissions. Moreover, a new
intertemporal optimization module has been added to AMIGA that allows an evaluation of early
reductions and the banking of allowances to be incorporated into policy scenarios. Hence, the
system is well suited to evaluate a variety of multi-emission strategies that are driven by price
incentives as well as R&D programs, voluntary initiatives, and cap and trade policies.

6 Because of recent upgrades and enhancements made in the model, the current reporting period is extended only
through the year 2015. We expect the full reporting period to extend back to the year 2030 in the very near future.

22


-------
The model includes a complete database of all electric utility generating units within the United
States. The cost and performance characteristics of the electricity supply technologies generally
follow those modeled within the Energy Information Administration's National Energy
Modeling System. The characteristics associated with the various emission control technologies
generally follow those modeled within the Integrated Planning Model used by the Environmental
Protection Agency.

The AMIGA modeling system is a highly organized, flexible structure that is programmed in the
C language. It includes modules for household demand, production of goods, motor vehicles,
electricity supply, and residential and commercial buildings and appliances.

The production modules contain representations of labor, capital, and energy substitutions using
a hierarchy of production functions. The adoption rates for cost-effective technologies depend
on energy prices as well as policies and programs that lower the implicit discount rates
(sometimes referred to as hurdle rates) that are used by households and businesses to evaluate
energy-efficiency and energy supply measures.7

7 For a more complete documentation of the AMIGA model, see Hanson, Donald A, 1999. A Framework for
Economic Impact Analysis and Industry Growth Assessment: Description of the AMIGA System, Decision and
Information Sciences Division, Argonne National Laboratory, Argonne, IL, April, 1999. For an example of other
policy excursions using the AMIGA model, see, Hanson, Donald A. and John A. "Skip" Laitner, 2000, "An
Economic Growth Model with Investment, Energy Savings, and CO2 Reductions," Proceedings of the Air & Waste
Management Association, Salt Lake City, June 18-22, 2000. Also see, Laitner, John A. "Skip", Kathleen Hogan,
and Donald Hanson, "Technology and Greenhouse Gas Emissions: An Integrated Analysis of Policies that Increase
Investments in Cost Effective Energy-Efficient Technologies," Proceedings of the Electric Utilities Environment
Conference, Tucson, AZ, January 1999.

23


-------
5.2. Summary Tables for Study Scenarios

5.2.1. Reference Case Projections

Table 1. Summary Data



Energy Consumption and Emissions

1998

2002

2005

2007

2010

2015

Total Primary Energy (Quadrillion Btus)

96.47

102.91

107.81

110.78

115.23

122.07

Total Electricity Use (Billion Kilowatt-hours)

3,411

3,714

3,942

4,104

4,346

4,697

Total Electricity Expenditures (Billions of 1999$)

223.8

236.7

245.9

255.3

269.4

291.3

Electric Sector Carbon (Million Metric Tons)

559

603

635

658

691

738

SO2 (Million Short Tons)

13.24

10.46

10.31

10.23

10.02

9.35

NOx (Million Short Tons)

6.01

4.47

4.49

4.54

4.56

4.51

Mercury (Tons)

47.36

48.86

48.81

48.70

48.25

46.01

Table 2. Summary Data



Electric Generation (Billion Kilowatt-hours)

1998

2002

2005

2007

2010

2015

Coal

1,829

1,961

2,055

2,100

2,157

2,189

Gas and Oil

462

584

672

788

967

1,329

Nuclear

674

712

740

732

720

639

Hydropower

325

322

323

323

323

324

Renewables

57

69

76

79

86

99

Total Generator Load

3,347

3,648

3,866

4,021

4,253

4,580

Table 3. Summary Data



Cogeneration - Independent Power Production













(Billion Kilowatt-hours)

1998

2002

2005

2007

2010

2015

Coal Cogeneration

52

52

52

52

52

52

Gas and Oil Cogeneration

220

240

255

262

274

293

Biomass Cogeneration

27

29

30

32

35

40

Municipal Solid Waste and Other Cogeneration

12

10

9

9

9

9

Other Renewables Generation

6

5

5

5

5

5

Total Independent Power Production

317

336

351

361

375

399

Amount for Own Use

158

171

181

188

199

212

Sales to grid

158

165

170

173

176

186

Table 4. Summary Data



Selected Energy Prices (1999 dollars)

1998

2002

2005

2007

2010

2015

Wellhead Gas Price ($/MCF)

2.02

2.28

2.49

2.57

2.69

2.83

Average Electricity Price (S/MWh)

68.82

66.82

65.39

65.19

64.96

64.95

Carbon Permit Price ($/metric ton)

0

0

0

0

0

0

Sulfur Dioxide Permit Price ($/ton)

0

0

0

0

0

0

Nitrogen Oxide Permit Price ($/ton)

0

0

0

0

0

0

Mercury Permit Price (million $/ton)

0

0

0

0

0

0

Table 5. Summary Data



Macroeconomic Data (Billions of 1999$)

1998

2002

2005

2007

2010

2015

Real Gross Domestic Product (GDP)

8,882.2

9,770.2

11,431.3

12,116.7

13,211.7

15,264.3

Real Investment

1,577.0

2,018.4

2,474.0

2,697.5

3,042.4

3,768.4

Real Consumption

5,933.6

6,763.9

7,681.7

8,180.2

8,902.0

10,361.2

24


-------
5.2.2. Scenario A: Emission Constraints Using Reference Case Technologies

Table 1. Summary Data

Energy Consumption and Emissions

1998

2002

2005

2007

2010

2015

Total Primary Energy (Quadrillion Btus)

96.47

100.94

104.97

107.54

111.37

117.14

Total Electricity Use (Billion Kilowatt-hours)

3,411

3,685

3,831

3,958

4,156

4,417

Total Electricity Expenditures (Billions of 1999$)

223.8

279.7

298.2

318.1

353.9

404.5

Electric Sector Carbon (Million Metric Tons)

559

507

500

499

499

518

SO2 (Million Short Tons)

13.24

8.66

7.35

6.34

4.04

2.07

NOx (Million Short Tons)

6.01

3.85

3.24

2.86

2.11

1.58

Mercury (Tons)

47.36

31.82

25.39

21.11

14.43

9.34

Table 2. Summary Data



Electric Generation (Billion Kilowatt-hours)

1998

2002

2005

2007

2010

2015

Coal

1,829

1,668

1,609

1,566

1,467

1,406

Gas and Oil

462

644

746

855

1,095

1,429

Nuclear

674

712

740

732

720

639

Hydropower

325

322

323

323

323

323

Renewables

57

190

248

308

365

405

Total Generator Load

3,347

3,536

3,667

3,784

3,970

4,202

Table 3. Summary Data



Cogeneration - Independent Power Production













(Billion Kilowatt-hours)

1998

2002

2005

2007

2010

2015

Coal Cogeneration

52

52

52

52

52

52

Gas and Oil Cogeneration

220

303

318

325

337

356

Biomass Cogeneration

27

29

30

32

35

40

Municipal Solid Waste and Other Cogeneration

12

10

9

9

9

9

Other Renewables Generation

6

19

19

19

19

19

Total Independent Power Production

317

413

428

437

452

475

Amount for Own Use

158

210

221

228

239

253

Sales to grid

158

203

207

209

212

222

Table 4. Summary Data



Energy and Permit Prices (1999 dollars)

1998

2002

2005

2007

2010

2015

Wellhead Gas Price ($/MCF)

2.02

2.77

3.33

3.45

3.63

3.53

Average Electricity Price ($/MWh)

68.82

80.52

82.59

85.29

90.36

97.15

Carbon Permit Price ($/metric ton)

0

59

75

87

110

138

Sulfur Dioxide Permit Price ($/ton)

0

113

143

166

210

308

Nitrogen Oxide Permit Price ($/ton)

0

666

839

979

1233

1812

Mercury Permit Price (million $/ton)

0

129

162

189

238

350

Table 5. Summary Data



Macroeconomic Data (Billions of 1999$)

1998

2002

2005

2007

2010

2015

Real Gross Domestic Product (GDP)

8,882.2

9,764.6

11,426.0

12,109.9

13,204.3

15,260.1

Real Investment

1,577.0

2,023.0

2,488.0

2,714.6

3,067.3

3,790.2

Real Consumption

5,933.6

6,755.1

7,663.6

8,158.3

8,870.9

10,336.3

25


-------
5.2.3. Scenario B: Emission Constraints Using Advanced Case Technologies

Table 1. Summary Data



Energy Consumption and Emissions

1998

2002

2005

2007

2010

2015

Total Primary Energy (Quadrillion Btus)

96.47

101.03

104.87

107.32

111.00

116.54

Total Electricity Use (Billion Kilowatt-hours)

3,411

3,681

3,814

3,929

4,112

4,346

Total Electricity Expenditures (Billions of 1999$)

223.8

273.9

289.1

306.7

337.4

381.2

Electric Sector Carbon (Million Metric Tons)

559

516

509

504

504

524

SO2 (Million Short Tons)

13.24

8.65

7.37

6.20

3.91

2.14

NOx (Million Short Tons)

6.01

3.86

3.26

2.82

2.09

1.58

Mercury (Tons)

47.36

31.9

25.59

20.84

14.37

9.70

Table 2. Summary Data



Electric Generation (Billion Kilowatt-hours)

1998

2002

2005

2007

2010

2015

Coal

1,829

1,692

1,649

1,586

1,501

1,476

Gas and Oil

462

652

723

841

1,048

1,318

Nuclear

674

712

740

732

720

639

Hydropower

325

322

323

323

323

323

Renewables

57

170

228

288

345

385

Total Generator Load

3,347

3,548

3,663

3,769

3,937

4,141

Table 3. Summary Data



Cogeneration - Independent Power Production













(Billion Kilowatt-hours)

1998

2002

2005

2007

2010

2015

Coal Cogeneration

52

52

52

52

52

52

Gas and Oil Cogeneration

220

290

305

313

324

343

Biomass Cogeneration

27

29

30

32

35

40

Municipal Solid Waste and Other Cogeneration

12

10

9

9

9

9

Other Renewables Generation

6

16

16

16

16

16

Total Independent Power Production

317

398

412

422

436

460

Amount for Own Use

158

203

213

220

231

245

Sales to grid

158

195

200

202

205

215

Table 4. Summary Data



Energy and Permit Prices (1999 dollars)

1998

2002

2005

2007

2010

2015

Wellhead Gas Price ($/MCF)

2.02

2.65

3.12

3.25

3.45

3.53

Average Electricity Price (S/MWh)

68.82

78.72

80.29

82.69

86.96

92.95

Carbon Permit Price ($/metric ton)

0

51

64

75

94

119

Sulfur Dioxide Permit Price ($/ton)

0

130

164

191

240

353

Nitrogen Oxide Permit Price ($/ton)

0

725

913

1065

1342

1972

Mercury Permit Price (million $/ton)

0

137

173

202

254

374

Table 5. Summary Data



Macroeconomic Data (Billions of 1999$)

1998

2002

2005

2007

2010

2015

Real Gross Domestic Product (GDP)

8,882.2

9,767.2

11,429.0

12,114.0

13,209.5

15,264.0

Real Investment

1,577.0

2,022.8

2,487.8

2,714.3

3,067.0

3,790.5

Real Consumption

5,933.6

6,757.1

7,665.2

8,160.3

8,873.7

10,337.1

26


-------
5.2.4. Scenario C: Emission Constraints Using the Moderate CEF Scenario Assumptions

Table 1. Summary Data



Energy Consumption and Emissions

1998

2002

2005

2007

2010

2015

Total Primary Energy (Quadrillion Btus)

96.47

101.08

104.81

107.25

110.76

116.21

Total Electricity Use (Billion Kilowatt-hours)

3,411

3,678

3,797

3,903

4,070

4,279

Total Electricity Expenditures (Billions of 1999$)

223.8

268.3

280.9

296.3

323.0

360.4

Electric Sector Carbon (Million Metric Tons)

559

520

515

513

512

535

SO2 (Million Short Tons)

13.24

8.50

7.33

6.24

3.93

2.17

NOx (Million Short Tons)

6.01

3.80

3.26

2.86

2.11

1.63

Mercury (Tons)

47.36

31.51

25.5

21.07

14.56

10.01

Table 2. Summary Data



Electric Generation (Billion Kilowatt-hours)

1998

2002

2005

2007

2010

2015

Coal

1,829

1,701

1,680

1,641

1,559

1,558

Gas and Oil

462

657

693

775

964

1,182

Nuclear

674

712

740

732

720

639

Hydropower

325

322

323

323

323

323

Renewables

57

159

217

277

334

374

Total Generator Load

3,347

3,552

3,653

3,749

3,900

4,077

Table 3. Summary Data



Cogeneration - Independent Power Production













(Billion Kilowatt-hours)

1998

2002

2005

2007

2010

2015

Coal Cogeneration

52

52

52

52

52

52

Gas and Oil Cogeneration

220

284

299

306

318

337

Biomass Cogeneration

27

29

30

32

35

40

Municipal Solid Waste and Other Cogeneration

12

10

9

9

9

9

Other Renewables Generation

6

15

15

15

15

15

Total Independent Power Production

317

390

405

415

429

453

Amount for Own Use

158

199

209

216

227

241

Sales to grid

158

192

196

198

202

211

Table 4. Summary Data



Energy and Permit Prices (1999 dollars)

1998

2002

2005

2007

2010

2015

Wellhead Gas Price ($/MCF)

2.02

2.54

2.93

2.99

3.09

2.98

Average Electricity Price (S/MWh)

68.82

77.12

78.29

80.36

84.06

89.25

Carbon Permit Price ($/metric ton)

0

44

55

64

81

102

Sulfur Dioxide Permit Price ($/ton)

0

148

187

218

274

403

Nitrogen Oxide Permit Price ($/ton)

0

799

1006

1173

1478

2172

Mercury Permit Price (million $/ton)

0

148

186

217

273

401

Table 5. Summary Data



Macroeconomic Data (Billions of 1999$)

1998

2002

2005

2007

2010

2015

Real Gross Domestic Product (GDP)

8,882.2

9,767.6

11,431.7

12,120.2

13,218.9

15,275.7

Real Investment

1,577.0

2,022.7

2.487.7

2,714.1

3,066.8

3,790.8

Real Consumption

5,933.6

6,757.1

7,667.3

8,165.3

8,881.7

10,346.6

27


-------
5.2.5. Scenario D: Emission Constraints Using the Advanced CEF Scenario Assumptions

Table 1. Summary Data



Energy Consumption and Emissions

1998

2002

2005

2007

2010

2015

Total Primary Energy (Quadrillion Btus)

96.47

101.12

104.76

107.08

110.44

115.66

Total Electricity Use (Billion Kilowatt-hours)

3,411

3,675

3,779

3,875

4,025

4,208

T otal Electricity Expenditures (Billions of 1999$)

223.8

263.2

273.0

286.2

308.9

340.9

Electric Sector Carbon (Million Metric Tons)

559

525

521

517

514

537

SO2 (Million Short Tons)

13.24

8.41

7.24

6.13

3.88

2.24

NOx (Million Short Tons)

6.01

3.79

3.27

2.85

2.11

1.62

Mercury (Tons)

47.36

31.09

25.3

20.87

14.41

10.12

Table 2. Summary Data



Electric Generation (Billion Kilowatt-hours)

1998

2002

2005

2007

2010

2015

Coal

1,829

1,713

1,707

1,665

1,587

1,614

Gas and Oil

462

660

664

739

904

1,069

Nuclear

674

712

740

732

720

639

Hydropower

325

322

323

323

323

323

Renewables

57

149

207

267

324

364

Total Generator Load

3,347

3,556

3,642

3,726

3,859

4,009

Table 3. Summary Data



Cogeneration - Independent Power Production













(Billion Kilowatt-hours)

1998

2002

2005

2007

2010

2015

Coal Cogeneration

52

52

52

52

52

52

Gas and Oil Cogeneration

220

278

293

300

312

331

Biomass Cogeneration

27

29

30

32

35

40

Municipal Solid Waste and Other Cogeneration

12

10

9

9

9

9

Other Renewables Generation

6

14

13

13

13

13

Total Independent Power Production

317

383

398

407

422

445

Amount for Own Use

158

195

205

212

223

237

Sales to grid

158

188

192

195

198

208

Table 4. Summary Data



Energy and Permit Prices (1999 dollars)

1998

2002

2005

2007

2010

2015

Wellhead Gas Price ($/MCF)

2.02

2.41

2.70

2.79

2.92

2.98

Average Electricity Price (S/MWh)

68.82

75.62

76.39

78.16

81.26

85.85

Carbon Permit Price ($/metric ton)

0

37

46

54

68

86

Sulfur Dioxide Permit Price ($/ton)

0

165

208

243

306

449

Nitrogen Oxide Permit Price ($/ton)

0

845

1065

1242

1564

2299

Mercury Permit Price (million $/ton)

0

159

200

233

294

432

Table 5. Summary Data



Macroeconomic Data (Billions of 1999$)

1998

2002

2005

2007

2010

2015

Real Gross Domestic Product (GDP)

8,882.2

9,768.4

11,434.3

12,125.7

13,227.2

15,285.9

Real Investment

1,577.0

2,022.6

2,487.6

2,713.9

3,066.7

3,791.0

Real Consumption

5,933.6

6,757.1

7,668.4

8,170.0

8,889.2

10,355.9

28


-------