Clemson University

Copyright © 2009

All rights reserved. No part of this book may be reproduced or transmitted

in any form without permission from the George Marshall Institute.

The George C. Marshall Institute, a nonprofit research group founded in 1984, is

dedicated to fostering and preserving the integrity of science in the policy process. The

Institute conducts technical assessments of scientific developments with a major impact

on public policy and communicates the results of its analyses to the press, Congress

and the public in clear, readily understandable language. The Institute differs from other

think tanks in its exclusive focus on areas of scientific importance, as well as a Board

whose composition reflects a high level of scientific credibility and technical expertise.

Its emphasis is public policy and national security issues primarily involving the physical

sciences, in particular the areas of missile defense and global climate change.

I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

II. Main Findings. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Impact on GDP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Impact on Consumption and Social Welfare . . . . . . . . . . . . . . . . . . . . . . . . 6

Prices of Carbon Allowances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

Impact on Employment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

Impact on Electricity Prices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

Impact on Gasoline Prices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

Impact on Natural Gas Prices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

III. Summaries of the estimates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

Lieberman-Warner Climate Security Act S. 2191 (S. 3036). . . . . . . . . . . . . 16

Features of the Lieberman-Warner Climate Security Act (S. 2191) . . . . . . . . 16

Seven Analyses of S.2191 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

IV. Other Proposals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

S. 1766 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

Carbon Tax Proposals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

Literature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

Summary of Assumptions of the Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

Estimating the Impact on Consumption Assumptions and Technical Details . . . . . 30

Text of the Program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

1

In this paper we summarize various estimates of the costs of mitigation of adverse

impact of the climate change via cap-and-trade. We find that the differences in the

estimated impacts on gross domestic product (GDP), consumption, employment, and

gasoline, electricity and natural gas prices are mainly driven by the following factors:

the timeframe of new technology development, growth potential of existing clean

sources of energy, availability of offsets (domestic, and international), and banking

of allowances.

However, our main finding is that even for more optimistic estimates, the mitigation

costs are likely to amount to as much as 1% drop in consumption starting today and

going into the future, which, as we argue, constitutes an enormous impact on social

welfare. Thus, it is important to carefully assess the costs of global warming to see

whether they justify such drastic measures.

Contact Author:

Sergey Mityakov

John E. Walker Department of Economics, Clemson University,

222 Sirrine Hall, Clemson, SC 29634. Email: smityak@clemson.edu

There is a growing body of literature which tries to assess the costs of climate change

and propose ways to mitigate its negative impacts.1 The Democratic and Republican

candidates in the 2008 U.S. president election both favored a comprehensive

regulatory regime to mitigate the adverse impacts of climate change. The winning

candidate, President Barack Obama, supports the implementation of a market-based

cap-and-trade system to reduce carbon emissions by the amount many scientists say is

necessary: 80 percent below 1990 levels by 2050.2

It appears likely that some form of cap-and-trade system to cut greenhouse-gas (GHG)

emissions will be enacted in the U.S. in the coming years. Thus, it is important for the

public to fully understand both the costs of climate change and the costs of avoiding its

negative impacts. While the media, policymakers, and others have given much

attention to the possible negative impacts of the climate change, comparatively little

effort has been devoted to presenting the cost estimates of differing mitigation

strategies. American households will bear large costs if any of the proposed plans to

curb GHG emissions are adopted.

In the present paper we summarize many of the available household-level mitigation

cost estimates. We compare these estimates to gauge their relative sensitivities to

differing assumptions. These assumptions include, but are not limited to, educated

guesses about the level and timing of proposed abatement efforts, the costs and

timeframe of developing new and cleaner technologies or improving existing ones, and

mitigation efforts on part of other countries.

In particular, we summarize different cost estimates generated for the Lieberman-

Warner Climate Security Act (S.2191) and discuss other legislative proposals such as

the Low Carbon Economy Act (S.1766) and carbon tax proposals proffered by

Representatives Dingell, Stark, and others. Yet our main focus is Lieberman-Warner,

since its provisions most closely resemble the vision put forth by President Obama.

Moreover, many have argued a carbon tax system is not a politically viable option for

the foreseeable future.

The rest of the paper is organized as follows: Section II compares the results and

assumptions of the seven cost-estimate analyses conducted upon Lieberman-Warner.

Section III discusses Lieberman-Warner in more depth and summarizes the individual

cost-estimate analyses we investigated. Section IV discusses estimates of other

abatement proposals. Section V concludes our analysis.

2

3

We analyzed seven analyses of Lieberman-Warner focusing on the cost aspects which

we think are of particular importance to American households: change in GDP and

resulting change in household consumption, employment changes, and increases in

gasoline, natural gas, and electricity prices. The following groups and organizations

conducted these studies:

1. MIT Joint Program on the Science and Policy of Global Change

2. The American Council for Capital Formation (ACCF) and the National

Association of Manufacturers (NAM)

3. CRA International

4. The Environmental Protection Agency

5. The Energy Information Administration (EIA)

6. The Heritage Foundation’s Center for Data Analysis (CDA)

7. The Clean Air Task Force (CATF)

Estimated GDP losses vary widely, from a 0.3%-0.5% to 3% drop in GDP below the

business-as-usual projections in 2015 and a 1% to 10% drop in 2050. The timeframes

of new technology development and growth in existing clean sources of energy,

availability of offsets (domestic, international), and banking of allowances are likely to

account for most of these differences in GDP costs estimates.

The studies listed above make different modeling assumptions about the abatement

process; hence, the resulting estimates of GDP losses vary considerably. Table 1 shows

the estimated impact on GDP from the seven studies under consideration. The MIT

group, EIA, and CATF predict comparatively lower damage to GDP (around 0.5% in

2015 and 2030 going up to 1% in 2050); the CRA and ACCF estimates are much

higher at 1% on average in 2015 up to 3% in 2030. The CDA and EPA estimates fall

somewhere in between these extremes.

A comparative analysis of both the models’ assumptions and results reveals that the following

three factors are likely to account for the differences in the estimated impact on GDP:

potential of nuclear and renewable sources of energy,

Table 2 compares the seven models’ assumptions regarding these three factors.

Summaries that include more information about the assumptions are available in the

Appendix, and Section III provides more detailed information about individual models.

4

Studies assuming a limited availability of alternative sources of energy or slower

development and adoption of carbon-free sources of energy predict higher GDP losses.

This is quite understandable, since hitting the same abatement target with “dirtier”

sources of energy requires greater cutbacks in energy consumption and results in

higher GDP loss. GDP could decrease by a factor of two to three, depending on

alternative assumptions.

For example, the ACCF/NAM analysis caps some alternative energy source development

and deployment such as wind, biomass and clean coal and natural gas carbon

capture and sequestration (CCS) technologies. In turn, the estimated costs reported by

ACCF are considerably higher than for alternative studies examining scenarios without

these alternative energy caps. On the other hand, the CATF study uses the same

NEMS model as ACCF, but without such severe constraints on new mitigation

technology development, and consequently arrives at much lower GDP loss estimates.

Many other studies include scenarios with different assumptions about the potential

growth of alternative energy sources. Different scenarios presuming strong constraints

on renewables, nuclear and other forms of cleaner energy development arrive at larger

cost estimates. For example, the EPA tested a scenario with constrained nuclear,

biomass and carbon capture and storage, which predicted GDP losses of 1.5 to 2 times

higher than the EPA scenarios lacking such technological constraints. Similar effects

are observed in other studies as well.

% Change in GDP % Change in GDP % Change in GDP

Group Model Scenario from Baseline 2015 from Baseline 2030 from Baseline 2050

MIT EPPA No Offsets, No CSS Subsidy -0.65% -0.31% -1.10%

15% Offsets -0.55% -0.54% -0.82%

CSS Subsidy -0.66% -0.26% -1.01%

15% Offsets, CSS Subsidy -0.57% -0.38% -0.75%

ACCF/NAM* NEMS Low Cost -0.80% -2.60% NA

High Cost -1.20% -2.40% NA

CRA MRN-NEEM S. 2191 -1.75% -1.00% -3.50%

CDA* GI Generous -0.14% -0.56% NA

Reasonable -1.02% -2.18% NA

EPA ADAGE S. 2191 -0.70% -0.90% -2.37%

IGEM -2.00% -3.76% -6.90%

ADAGE S. 2191- No Offsets NA NA NA

IGEM -3.30% -5.90% -10.10%

ADAGE S. 2191- Constrained Nuclear,

Biomass and CCS -1.10% -2.30% -4.40%

IGEM NA NA NA

EIA** NEMS S. 2191 Core -0.30% -0.30% NA

S. 2191 Limited Alternative/

No International Offsets -0.90% -0.80% NA

CATF NEMS S. 2191 NA -0.70% NA

* ACCF/NAM reports in the year 2014.

**EIA reports in the year 2020.

5

* ACCF/NAM reports in the year 2014.

**EIA reports in the year 2020.

Entities covered by Lieberman-Warner can satisfy their GHG reduction obligations by

either purchasing carbon allowances or engaging in other projects which will offset

their obligation. Firms can purchase offsets from international cap-and-trade programs

similar to the one envisioned by Lieberman-Warner or they can engage in emission

reduction for non-covered emission types, which lowers their obligation on the covered

emissions.

This essentially gives firms additional opportunities to satisfy emission caps, and, thus,

leads to lower costs of abatement. When no offsets are assumed, estimated costs go

up in all models by (approximately) a factor of 1.5.

The second major factor affecting mitigation cost estimates is the availability of

domestic/international GHG offsets. Entities covered by Lieberman-Warner can

satisfy part of their GHG reduction obligations by either purchasing carbon allowances

or engaging in other projects offsetting some of their contributions. Firms can purchase

offsets from international cap-and-trade programs similar to the one envisioned by

Lieberman-Warner. When no offsets are assumed, estimated costs increase in all

models. Greater availability of offsetting options reduces the economic impacts of

Lieberman-Warner. In the EPA model, the absence of international offsets increases

estimated costs by a factor of 1.5, from 2% in 2015 to 3% in 2015 (using the IGEM

model). The ACCF/NAM study assumes limited amounts of offsets (<20%) in the

“high cost scenario,” which increases the estimated cost by a factor of 1.5 compared

to the case where there is no such restriction on offsets.

Limited Availability of Banking of

Group Scenario alternatives Offsets Allowances

MIT No Offsets, No CSS Subsidy Yes No Yes

15% Offsets Yes Limited Yes

CSS Subsidy No No Yes

15% Offsets, CSS Subsidy No Limited Yes

ACCF/NAM* Low Cost Somewhat limited Somewhat limited No

High Cost Yes Limited No

CRA S. 2191 No Yes Yes

CDA* Generous Somewhat limited Yes No

Reasonable Yes Yes No

EPA S. 2191 No Yes Yes

S. 2191 - No Offsets No No Yes

Constrained Nuclear, Biomass, CCS Yes Yes Yes

EIA** S. 2191 Core No Yes Yes

S. 2191 Limited Alternative/No International Offsets Yes No Yes

CATF S. 2191 No Yes Yes

6

The third factor significantly influencing estimated costs is availability of “banking”

allowances. Banking enables firms to save unused allowances for later years, essentially

providing firms the flexibility to gradually adjust their operations to meet the targets and

lessen the overall abatement costs. When the permissibility of banking is assumed,

estimated costs fall by a factor of 2.

Finally, the ability to “bank” or store allowances also has a major impact on the estimated

costs of abatement. Allowance banking allows covered entities to save credits

they do not use or sell in a given year. Saving credits provides these entities more

flexibility when the total number of credits begins to decline in future years. If firms

are given the opportunity to store credits, then they can gradually adjust their operations

to meet targets, lessen their overall abatement costs and “cushion the blow” of

declining emissions. For example, the CRA study provides cost estimates with and

without the banking assumption. When banking is permitted, the entire costs of

programs such as those proposed by Lieberman-Warner are significantly decreased.

Studies assuming that no banking of allowances is permitted usually show higher

estimates of loss in GDP: for example, the ACCF/NAM, CRA, and CDA scenarios

which do not include banking estimate GDP losses 1.5-2 times higher than other models

which include banking of allowances, such as the EIA and MIT studies.

Consumption is predominantly affected by the same factors as GDP. As before, studies

which assume limited alternative sources of energy and/or limited offsets usually show

higher (by a factor of 2 or 3) consumption cost estimates.

GDP is not the most informative measure of a GHG mitigation plan’s household

impact. Measuring changes in consumption is a better way of determining the burdens

that individual American households will bear under cap-and-trade through examining

the welfare losses. While individual utility/welfare is not directly observable, measuring

household consumption is undoubtedly a more direct gauge of household well-being

than GDP. Table 3 presents estimated drops in consumption in response to the

mitigation path consistent with Lieberman-Warner in 2015, 2030 and 2050.

A comparison of Table 3’s results to those of Table 1, shows the expected GDP/

consumption correlation: studies that estimate higher drops in GDP are likely to have

higher estimated drops in consumption as well. Thus, the assumptions affecting GDP

loss (availability of offsets, the timeframe of low-carbon technology development, and

predictions about the growth of existing sources of clean and renewable energy) also

alter the magnitude of decreases in consumption.

Studies assuming limited alternative sources of energy and/or limited offsets usually

result in higher consumption cost estimates. The ACCF study, which puts caps on

development of nuclear and alternative energy sources, predicts a decline in consumption

that is two to three times higher than the MIT, EPA or EIA results, which do

not make such restrictions. Moreover, one ACCF scenario that both tightens caps on

renewable energy development and strictly limits the availability of offsets (the “high

cost” scenario) estimates consumption losses increase by a factor of 2.8. Similarly, one

EIA study scenario estimates consumption losses increase by a factor of 2 to 3 when

assuming limited alternatives to coal-generated energy production and no international

offsetting opportunities for U.S. firms.

Even under the most optimistic assumptions, every study we examined predicts huge

welfare costs in terms of consumption. A lower estimate involves a drop in consumption

of 0.8%-1% below the business-as-usual scenario in every year starting in 2008

and going into the future, which represents a huge decrease in social welfare.

7

% Change % Change % Change Balanced Growth

Group Scenario in 2015 in 2030 in 2050 Equivalent*** Reported Impact

MIT No Offsets, No CSS Subsidy -0.35% -1.93% -2.36% -0.96% Change in Market

15% Offsets -0.29% -1.60% -2.10% -0.81% Consumption

CSS Subsidy -0.37% -1.93% -2.26% -0.97%

15% Offsets, CSS Subsidy -0.31% -1.47% -2.01% -0.77%

ACCF/NAM* Low Cost -1.00% -2.90% NA -0.98% (-1.57%) Change in

High Cost -2.80% -4.90% NA -2.57% (-3.09%) Household Income

CRA -4.50% -3.50% -4.20% -3.17% Cost Per Household

CDA* Generous -0.60% -0.48% NA -0.41% (-0.42%) Change in Personal

Reasonable -1.35% -0.94% NA -0.89% (-0.90%) Consumption

EPA** S2191: ADAGE -0.43% -0.91% -2.10% -0.65% Change in Market

S.2191: IGEM -0.66% -1.44% -3.26% -1.02% Consumption

EIA** S. 2191 Core -0.40% -0.50% NA -0.31% (-0.36%) Change in Market

S. 2191 Limited Alternatives/ Consumption

No International Offsets -1.20% -1.10% NA -0.86% (-0.91%)

CATF S. 2191 NA -0.90% NA NA Change in Per

Capita GDP

* ACCF/NAM reports in the year 2014.

**EIA reports in the year 2020.

*** Estimates in brackets are computed for studies with "NA" in 2050 on the assumption that damages in 2050 equal to damages in 2030

8

Our examination of the costs to consumption permits us to more closely evaluate

Lieberman-Warner’s impact on overall social welfare. In particular we answer the

following questions: How large are the estimated drops in consumption? What does a

1% decrease in consumption in 2015 or a 3% drop in 2050 mean for the average

American household today when our leaders are considering making major policy

decisions?

We measure the impact on individual household well-being (following Lucas 1990) in

terms of balanced growth equivalency. In order to calculate the balanced growth

equivalent, we assume that consumption grows at a constant rate under the businessas-

usual scenario. The effects of mitigation efforts required by a cap-and-trade system

cause consumption to drop below this constant path. Table 3 presents data estimating

decreases in consumption in 2015, 2030, and 2030. When computing the balanced

growth equivalent, we assume that consumption is growing at the same rate as under

the business-as-usual scenario, but its level is permanently below the business-as-usual

path by some percentage. We choose a particular percentage so that individual wellbeing

under the balanced growth equivalent is the same as under a mitigation path. We

report this percentage in Table 3 as well.

Our calculations suggest that consumption under the constraints posed by Lieberman-

Warner’s cap-and-trade regime is equivalent to a constant (in percentage terms) consumption

decrease of around 0.8%-1% each year, starting today and continuing

to 2050.

At first glance, a consumption decrease of one percent may appear trivial. However,

as the 1% per year decreases compound, the welfare losses are substantial in the aggregate.

This phenomenon is familiar to long-term investors. Small initial investments

accruing modest returns over a long period of time result in large overall increases. It

is worth remembering what Nobel laureate Robert Lucas said concerning the estimated

the welfare gains from elimination of capital income taxation, which he calculated to

be around 1 percent of consumption:

“…I estimated the overall gain in welfare to be around 1 percent of consumption,

or perhaps slightly less. … It is about twice the welfare gain that I have elsewhere

estimated would result from eliminating 10 percent points of inflation, and

something like 20 times the gain from eliminating post-war sized business

fluctuations. It is about 10 times the gain Arnold Harbenger once estimated from

eliminating all product-market monopolies in the U.S.”4

Based on our analysis, we estimate the cap-and-trade impact will probably be quite

large as consumption constantly decreases, causing many negative economic results.

In light of this evidence, it is important to assess the costs of global warming and

determine whether they justify such large welfare losses.

9

What impact will constant, compounding consumption decreases have on individual

American households? Table 4 presents the estimated impact of a 1% decrease in consumption

for an average household of four people. Projections for business-as-usual

scenario consumption are taken from Paltsev et al. (2008)

We find that a mitigation path consistent with Lieberman-Warner’s provisions is

equivalent to a permanent tax increase for the average American household. This

increase is projected to amount to an additional $1100 in taxes in 2008. Moreover,

this cap-and-trade “tax” increases over time in real terms from about $1400 to $2000

during 2015-2030 and approximately $2000 to $3000 in 2030-2050. The de facto

tax increase becomes quite significant when one considers the average American

household spends about $2500 on food annually,5 or approximately $208 monthly.

The decrease in consumption per capita of $277 annually is equivalent to more than

one month’s food budget for the average American household, keeping other

consumption levels constant.6

Another way to gauge this cap-and-trade tax impact is comparing it to auto-loan

payments. For example, a new 2009 C-Class Mercedes can be leased for around $429

per month.7 A decrease in consumption by $1110 amounts is equivalent to 2.5

monthly payments on this luxury car. This tax amounts to aboutlmost three and a half

monthly payments in 2015 and almost seven payments in 2050.8

But the average American household usually does not buy a Mercedes. What about a

Honda Civic? A new 2009 Honda Civic LX can be bought for around $189 a

month.9 A decrease in consumption by $1100 equals to almost six monthly payments

on this car every twelve months in 2008 and fifteen monthly payments in 2050.

2008* 2015 2030 2050

Population (Million) 301 321 359 397

Consumption (billion 2005$) $8,217 $11,533 $17,761 $29,567

Consumption/Per capita (2005 $) 27,760 $35,928 $49,474 $74,476

Decrease in consumption per capita (2005 $) $277 $359 $495 $745

Decrease for a family of 4 (2005 $) $1,110 $1,437 $1,979 $2,979

*2005 data are used, 2008 are likely to be even higher.

10

Before we move on to discussing the cost estimates of other proposed legislation, we

must clarify our message to avoid confusion. We find that the costs of mitigation are

equivalent to a drop in consumption levels10 below the business-as-usual scenario by 1%.

The Natural Resources Defense Council (NRDC) recently published a critical analysis

of many of the analyses we have examined in this paper.11 The NRDC suggests that

abatement efforts, particularly cap-and-trade, will have only moderate impacts on

welfare because the new system will not stop the economic growth of the U.S., but just

make it slower.12 For example, GDP growth in, say, 2015 might be 3% under capand-

trade instead of 4% in its absence. Thus, the NRDC study criticizes some of the

cap-and-trade studies we examined for suggesting that abatement would involve

decreases in GDP rather than simply less overall growth.

We find that the costs of mitigation are equivalent to a drop in consumption levels

below the business-as-usual scenario by 1%.13 This does not mean that consumption

or GDP would actually drop in 2008 by 1%. Most negative economic costs will be

incurred in the future, when abatement targets become tighter. Our calculations

demonstrate the future costs are equivalent to a permanent drop in consumption by

1% below today’s level (continuing into the future) without the mitigation. That is,

under abatement, the consumer’s well-being will be the same as in the case when we

cut consumption under no abatement by 1% in every year starting in 2008 and going

into the future.

Of course, we could restate the same welfare costs in terms of a lower growth rate in

consumption/GDP rather than drops in consumption levels. These are just alternative

ways of measuring the welfare loss. We prefer balanced growth equivalent estimates

since they are more standard in macroeconomic calibration exercises.

The fact that GDP does not drop below its 2008 level under the abatement scenario

does not mean that the mitigation costs of mitigation will be small. The problem is that

GDP drops below its potential level, the one that would have been attained if

mitigation did not take place. Our analysis of available estimates suggests that the

welfare costs of mitigation consistent with the provisions of Lieberman-Warner are

going to be large as GDP consistently falls short of its potential.

11

Lieberman-Warner’s consequences for employment and the prices households pay for

power and fuels hinge on the estimated price of carbon allowances, also called credits.

An allowance’s estimated price depends on the permissibility of banking of permits,

offset availability, the technological development of carbon capture and sequestration

(CCS) systems, and the number of different GHGs covered in addition to carbon

dioxide).

Lieberman-Warner’s impact on employment and energy and fuel prices hinge on the

estimated price of carbon allowances. The estimated price of these allowances is

highly sensitive to several assumptions. Among these assumptions are the availability

and extent of banking of allowances, the availability of offsets, the development of CCS

systems, and the number of different GHGs covered.

Lieberman-Warner provides for some banking of allowances, but the estimates from

the CDA and ACCF/NAM do not include banking in their assumptions. Alternatively,

the CRA study includes banking and estimates the presence of banking will cause the

price of allowances to be higher prior to 2040, but considerably lower afterwards. The

CRA study illustrates the presence of banking dramatically reduces the overall

economic costs of cap-and-trade. The CRA estimates banking reduces the total

estimated cost of Lieberman-Warner by $4.7 trillion dollars.

Many analyses test cost-estimate sensitivity by altering the number of foreign and

domestic offsets available at a given time. Lieberman-Warner allows covered entities

to use domestic and international offsets to cover up to 30% of total emissions. The

analyses we examined typically restrict the number of offsets in the various cases that

they examine. Lowering the number of potentially available offsets increases the price

of permits, but it keeps total emissions closer to the mandated targets.

The studies also incorporate varying assumptions vis-à-vis the future feasibility of CCS

and the construction of new low-carbon power plants. Many authors note that nuclear

power is a low-carbon alternative to coal, but regulatory and societal objections present

enormous problems for constructing new plants. Other “clean-energy” alternatives

like wind and biomass are expected to have potential expansion issues as well. Some

of the analyses also examine costs when CCS technology is either too expensive to

be commercially viable or completely unavailable. Studies by ACCF/NAM and CDA

assume alternative power is limited, the EIA and EPA studies test various assumptions

about its CCS availability, and the MIT and CATF analyses contain no assumptions

limiting nuclear, wind or CCS expansion.

12

*

Finally, lower cost-estimates are achieved in scenarios where other GHGs in addition

to carbon-dioxide count towards emissions. Lieberman-Warner covers a list of GHGs

besides carbon dioxide; including methane besides merely carbon-dioxide, but only the

CDA study limits reductions to carbon dioxide instead of the full array of GHGs. This

partially explains why the CDA’s cost-estimate predictions are relatively high. The

CDA’s estimates are even higher than the ACCF/NAM study’s, which otherwise

makes similar assumptions.

The assumptions driving the price of carbon allowances also affect employment. A

higher predicted carbon allowance price gives producers a tighter margin and they are

forced to shed jobs to maintain profit levels. The estimates of job losses range from

hundreds of thousands to millions.

Three of the analyses model changes in employment. The ACCF/NAM, CDA and

CRA estimate the net changes in employment. They assume that jobs will be created

in new “green” industries, such as CCS power plants, replacing older and more

Carbon Allowance Carbon Allowance Carbon Allowance

Group Model Scenario Price 2015 Price 2030 Price 2050

MIT EPPA No Offsets, No CSS Subsidy $59.15 $106.53 $233.42

15% Offsets $50.72 $91.35 $200.17

CSS Subsidy $57.91 $104.30 $228.52

15% Offsets, CSS Subsidy $50.44 $90.83 $199.03

ACCF/NAM* NEMS Low Cost $36.69 $227.52 NA

High Cost $38.36 $271.27 NA

CRA MRN-NEEM Banking $50.00 $90.00 $190.00

No Banking $40.00 $80.00 $350.00

CDA* GI Generous $50.37 $69.90 NA

Reasonable $50.37 $90.46 NA

EPA ADAGE S. 2191 $30.55 $64.27 $167.53

IGEM $42.14 $87.45 $231.80

ADAGE S. 2191- No Offsets NA NA NA

IGEM $81.13 $168.58 $447.79

ADAGE S. 2191- Constrained Nuclear, $57.95 $118.01 $305.55

IGEM Biomass and CCS NA NA NA

EIA** NEMS S. 2191 Core $30.84 $62.71 NA

S. 2191 Limited Alternative/

No International Offsets $78.12 $160.36 NA

CATF NEMS S. 2191 $17.43 $49.03 NA

* ACCF/NAM reports in the year 2014.

**EIA reports in the year 2020.

Prices converted to 2007$ using CPI

13

carbon-intensive economic sectors. In each study, the changes in employment

correlate with movements in the price of carbon allowances. The ACCF/NAM study

shows carbon prices steadily rising as the number of allowances falls over time; as a

result, the net change in employment is negative and increasing. The ACCF/NAM

study predicts the loss of 850,000 to 1.86 million jobs in 2014 and up to 3.04 to 4.05

million jobs lost by 2030. Alternatively, the CDA study predicts an increase in

employment of 120,000 jobs in 2015 as people are hired in the new “green”

industries, under generous assumptions. However, the CDA predicts that more than

500,000 jobs could be lost by 2015 and approximately 430,000 to 460,000 by 2030.

The assumptions driving the price of carbon allowances also affect employment

estimates. A higher predicted price of a carbon allowance gives producers a tighter

margin and forces them to shed jobs in order to remain profitable. Both the

ACCF/NAM and CDA assume that there is no banking of allowances, while the CRA

includes banking. Banking allows entities covered by cap-and-trade to save allowances

for future use. Without banking the price of allowances will start low, but rise quickly

as the number of available permits falls. Naturally, banking drives up the price of

allowances in 2015, but it allows the price to be lower than it would be without banking

after 2040. In 2015, the CRA estimates that there will be 3.75 million jobs lost. The

CRA study models job losses of up to 2.5 and 7.10 million in 2030 and 2050,

respectively. The permissibility of allowance banking might explain why the CRA

estimates of job losses are higher than the ACCF/NAM and CDA estimates in 2015.

Cap-and-trade will impact the prices households pay for electricity. Table 7 shows

estimated changes in the electricity prices from the baseline year. Electricity prices are

predicted to increase much more than gasoline prices. Lieberman-Warner’s cap-andtrade

system is estimated to increase the price of electricity by anywhere from 5% to

15% in 2015 and anywhere from 14% in the EPA core scenario to 128% in the

ACCF/NAM’s high cost scenario in 2030. The CATF model predicts a 7% increase

from the 2005 price in 2030. The EIA, MIT, and ACCF/NAM studies predict a 10%,

Change in Employment Change in Employment Change in Employment

Group Model Scenario from Baseline 2015 from Baseline 2030 from Baseline 2050

(millions of jobs) (millions of jobs) (millions of jobs)

ACCF/NAM* NEMS Low Cost -.085 -3.04 NA

High Cost -1.86 4.0 NA

CRA MRN-NEEM -3.75 -2.5 -7.1

CDA* GI Generous 0.15 -0.46 NA

Reasonable -0.717 -0.43 NA

* ACCF/NAM reports in the year 2014.

14

37%, and 124% increase in electricity prices from their baseline scenarios to 2030,

respectively. By 2050, electricity prices will have leveled off somewhat, returning to

near 2015 levels according to the MIT and EPA estimates.

Cap-and-trade will burden households with higher gasoline prices. Table 8 shows the

percent difference between the baseline gasoline price and the cap-and-trade adjusted

price. All models and scenarios demonstrate that Lieberman-Warner will increase the

price of gasoline above the reference scenario price but with large amounts of

variation. The CRA predicts that gas prices rise 145% above the reference scenario

in 2015. Yet, prices are only 30% higher than the reference scenario in 2030 because

the higher CAFE standards are included in the 2030 baseline. The lowest estimates

are CATF’s and EPA’s core scenarios, predicting increases of 11.6% and 16.7% by

2030, respectively. Alternative scenarios using higher-cost assumptions show increases

from 41.2% to 145% by 2030.

Table 8a in the Appendix indicates the estimated change in the price of gasoline from

the 2005 level. Most models predict that gasoline prices will steadily rise through

2050. In 2015, models that have more generous technology assumptions find that

gas prices could be lower than they were in 2005; other models predict gas prices will

be up to 25% higher than they were in 2005. By 2030, there is a wide spread in

estimates. The CATF study has the lowest estimate of a 5% increase above the 2005

price. The MIT model and the strictest EIA models predict a 40% to 45% increase,

while the ACCF/NAM model predicts 66% increase in the generous scenario and a

130% increase in the reasonable scenario. The MIT study, however, estimates that gas

prices will hit their highest level in 2030 and return to 2015 levels (which are 20%

higher than the 2005 price) by 2050. However, the EPA model predicts that 2050

gas prices will be 66% higher than the 2005 price.

% Change from % Change from % Change from

Group Model Scenario Baseline 2015 Baseline 2030 Baseline 2050

ACCF/NAM* NEMS Low Cost 14.00% 101.00% NA

High Cost 15.10% 128.40% NA

CRA MRN-NEEM 15.00% 35.00% 60.00%

EIA** NEMS S.2191 Core 5.20% 14.40% NA

S.2191 Limited Alternative/No International 26.90% 67.50% NA

* ACCF/NAM reports in the year 2014.

**EIA reports in the year 2020.

15

We must note which models incorporate more stringent fuel efficiency mandates.

Lieberman-Warner requires that all transportation fuels must become 10% less carbon

intensive by 2020, similar in design to a Low-Carbon Fuel Standard. The CRA is the

only analysis that incorporates this fuel requirement. This provision causes the price of

gasoline to increase rapidly in the early part of the forecast; the addition of a severe fuelefficiency

assumption may be why the CRA’s price estimates are higher than the others.

The other assumptions underpinning these models are rarely specified in the papers,

but undoubtedly affect the gasoline price estimates. Many analyses are not clear in how

they model changes in the prices of gasoline. Will gasoline producers simply pass

along carbon permit costs to consumers? Only the CATF and EPA models explicitly

state they assume the full cost of the carbon permit is ultimately borne by consumers.

Table 9 shows the estimated increase in the price of natural gas from the baseline price.

Under lower cost assumptions, the models predict that the price of natural gas will be

from 12% to 17% higher in 2015 than the baseline cases. In cases with less generous

assumptions, natural gas prices could experience increases of 20% to 49% higher than

the baseline estimate in 2015. One thing is certain: any cap-and-trade system will

increase the use of natural gas. Natural gas is the best alternative now available to non-

CCS coal, so if we reduce coal-powered energy generation, we will probably rely

heavily on natural gas as a substitute. By 2030, the increased reliance on natural gas

will cause the estimated prices to rise 20% to 107% higher than baseline prices in lowcost

scenarios and 87% to 145% in the high cost/limited alternatives cases.

Natural gas prices are particularly sensitive to the development of other low-carbon

alternatives to existing coal-produced power. The pace and scope of CCS development

has massive implications for future natural gas demand. For example, in an ACCF/

NAM case assuming limited low-carbon alternatives to coal, natural gas prices rise

more than 200% above 2005 levels by 2030. The EIA also predicts increases in the

% Change from % Change from % Change from

Group Model Scenario Baseline 2015 Baseline 2030 Baseline 2050

ACCF/NAM* NEMS Low Cost 13.00% 77.00% NA

High Cost 50.00% 145.00% NA

CRA MRN-NEEM 145.00% 30.00% 82.00%

EIA** NEMS S.2191 Core 9.30% 16.70% NA

S.2191 Limited Alternative/No International Offsets 20.30% 41.20% NA

CATF NEMS S.2191 9.63% 0.12% NA

* ACCF/NAM reports in the year 2014.

**EIA reports in the year 2020.

price of natural gas of around 200% in its limited alternative case by 2030. Even the

EIA’s core scenario predicts natural gas will cost 118% more in 2030. The MIT study

also predicts that natural gas will be 64% higher than 2005. By 2050, MIT predicts

natural gas prices will have declined slightly, but do not return to near the 2015 levels.

Overall, our results demonstrate the implications of differing scenario assumptions and

illustrate the massive economic challenges facing households if cap-and-trade becomes

reality. For this reason, we emphasize that it is important to carefully assess the costs

of global warming to see whether they justify the pain that mitigation efforts will cause

the average American and his or her family.

We start our review with the Lieberman-Warner Climate Security Act, legislation which

received considerable attention before it was defeated on June 6, 2008 in the Senate.

tons (mmt) in 2012 and to 1732 mmt in 2050. Reduces emissions of CO2 and

four other global warming pollutants by 4% in 2012, 19% in 2030, and 71% in

2050 below 2005 levels.14 Mandates stricter targets for hydrofluocarbons (HFCs)

emissions: 15% in 2020, 45% in 2030, and 70% in 2040 below 2005 levels.15

(PFCs), SF4, and HFCs. Converts four non-CO2 GHGs such as methane into

CO2-equivalents (CO2e) using Global Warming Potential (GWP) scale. Thus, it

covers 86% of total GHG emissions.

16

% Change from % Change from % Change from

Group Model Scenario Baseline 2015 Baseline 2030 Baseline 2050

ACCF/NAM* NEMS Low Cost 17.90% 107.80% NA

High Cost 20.70% 145.70% NA

CRA MRN-NEEM 12.50% 20.00% 90.00%

EIA** NEMS S.2191 Core 14.20% 26.10% NA

S.2191 Limited Alternative/No International Offsets 49.50% 87.30% NA

* ACCF/NAM reports in the year 2014.

**EIA reports in the year 2020.

17

and PFCs (also known as F-gases) and nitrogen dioxide, and downstream facilities

using more than 5,000 tons of coal per year to participate in the cap-andtrade

system.

Scarcity of allowances provides incentive for covered entities to develop and adopt

carbon capture and storage (CCS) technologies. Remaining allowances are

auctioned and the revenues used to fund low-carbon technology research and

development.

capture and reducing non-covered GHG emissions. Domestic offsets may cover up

15% of total obligation. Permits the use of foreign allowances from comparable

cap-and-trade systems (e.g., the Kyoto system) to cover an additional 15% of

obligations.

of allowances and potentially adjust the number of allowances created on year-toyear

basis.

In this section we review and compare seven cost estimates of the Lieberman-Warner’s

abatement schedule.

1. Model: Emissions Prediction and Policy Analysis (EPPA)

Organization: Massachusetts Institute of Technology (MIT) Joint Program on the

Science and Policy of Global Change

Authors: Sergey Paltsev, John M. Reily, Henry D. Jacoby, Angelo C. Gurgel, Gilbert

E. Metcalf, Andrei P. Sokolov and Jennifer F. Holak

Paltsev et al. use MIT’s Emissions Prediction and Policy Analysis (EPPA) model to

estimate the legislation’s effects on total emissions in the United States, the price of

energy and the resulting effects on consumer welfare. The MIT study included some

particular provisions found in Lieberman-Warner: upstream implementation, inclusions

of non-CO2 gases, the crediting of allowances for reducing non-covered emissions,

banking of allowances, and the distribution of allowances as incentive to use carbon

capture and storage (CCS) technology.

The analysis tests various stringency assumptions by changing the amount of offsets

available and the effect of a government subsidy for CCS. It specifically does not

estimate how Lieberman-Warner interacts with other federal emissions mandates (for

instance, H.R. 6), nor the effects of both free distribution and auctioning of allowances.

18

The MIT study presents a baseline scenario and four other scenarios. The strictest

scenario does not allow any offsets nor does it include the federal subsidy for CCS

technology research and development. The second scenario relaxes the offset

restriction to cover up to 15% of emissions and assumes that foreign offsets are too

costly. The third scenario returns offset availability to zero and adds the CCS subsidy.

The fourth and final scenario’s assumptions are the most like Lieberman-Warner’s

provisions. This scenario allows both 15% offsets and the CCS subsidy.

Under the assumptions mentioned above, the paper finds that price of an allowance

will rise steadily over time as total emissions levels fall from 5775 mmt in 2012 to

1732 mmt in 2050. Allowance costs range from $47 to $56 under their strictest

assumptions in 2015 and rise to $188 to $221 by 2030.16 The EPPA model predicts

GDP in 2015 will be 0.65% lower than baseline GDP in the strictest scenario and

0.57% lower in the more relaxed case. By 2050, GDP is estimated to be 1.1% to

0.75% lower in the strictest and least strict cases.

The economy-wide amount of consumption spending falls due to allowance price

increases that firms pass on to consumers. The model predicts that consumption could

fall from 0.29% to 0.37% in 2015 and 2.01% to 2.36% in 2050. The authors use

equivalent variation techniques to measure the effects on consumer welfare; essentially,

equivalent variation gauges how much a person would pay to avoid an increase in

prices. The model predicts that welfare loss would be 0.7% in 2015 in both the cases

and would be 1.81% and 1.54% in 2050. Stated another way, consumers would be

willing to pay $9.7 billion to avoid the price increases that S. 2191 creates in 2015

and they would be willing to pay $554.2 billion to avoid the price increase in 2050 (in

the least strict model).

The paper also estimates that by 2050, the prices of petroleum products will increase

by around 22%, natural gas by around 82% and electricity by around 61% from

2005 levels.

2. Model: National Energy Modeling System (NEMS)

Organizations: American Council for Capital Formation (ACCF) and National

Association of Manufacturers (NAM)

Authors: Science Applications International Corporation (SAIC)

The analysis in this paper was conducted by SAIC based on the parameters determined

by the ACCF and NAM. The paper uses the National Energy Modeling System

(NEMS) to estimate Lieberman-Warner’s effects on national economic indicators and

energy production and prices. The estimate includes H.R. 617 mandates as well as

updated construction costs for power generating facilities.

19

The ACCF and NAM study models two different scenarios, which they label “low cost”

and “high cost.” Under the low-cost scenario, offsets are available to cover more than

20% of emissions; future energy prices are determined by long-term predictions

contained in the Annual Energy Outlook (AEO) 2008 report, and every type of

power generation faces long-term constraints. The high-cost scenario circumscribes the

available offsets to between 15% and 20%, uses the AEO 2007 High Profile Side case

(a worst-case scenario prediction) to determine future energy prices, and places tighter

caps on building new power plant infrastructure. Neither scenario permits the banking

of allowances.

Under these assumptions, SAIC finds that carbon allowance prices will rise from

$36.69 in 2014 to $271.27 in 2030.18, 19 GDP decreases as the allowance price rises

over the forecasted period. In 2014, GDP is 0.8% lower than the baseline in the low

cost scenario and 1.6% lower in the high cost scenario. By 2030, GDP is 2.6% and

2.7% lower than the baseline in the low cost and high cost scenarios, respectively.

Higher production costs cause net job losses to range from 0.85 million jobs to 1.86

million jobs in 2014 and 3.04 million jobs to 4.05 million jobs in 2030.

The model estimates that the loss to the average household income would be 1.0% to

2.8% in 2014 to 2.9% to 4.9% in 2030. The paper also estimates that the residential

price of electricity will rise by about 13% above the baseline in 2014 and between

101% and 129% in 2030. Natural gas is also predicted to rise from 18% to 21% in

2014 and by 108% to 146% in 2030. Total expenditures on energy due to the price

increases rise from 15.5% to 33.5% in 2014 to 78.7% to 114.5% in 2030.

3. Model: Multi-Region National (MRN-NEEM)

Organization: CRA International

Authors: W. David Montgomery, Anne E. Smith

CRA International uses MRN-NEEM, which is a “multi-region national” model

integrating a macroeconomic model of all economic sectors including consumer

income, consumption, investment, and trade with a model of the energy and nonenergy

sectors. It predicts Lieberman-Warner’s effects on total GHG emissions, price

of the carbon allowances and energy, as well as the share of total power of various

types of power generation.

The CRA model includes most of Lieberman-Warner’s provisions, specifically its low

carbon fuel standards and the CCS subsidy, as well integrating the HR 6 provisions. It

builds a baseline scenario using the Annual Energy Outlook 2008 to predict future

energy costs and the increased CAFE standards, renewable fuel standards and

appliance efficiency standards mandated by H.R. 6.

20

This paper finds carbon allowance prices start at around $50, rise to around $80 by

2030, and culminate at $190 by 2050.20 The CRA study also provides cost estimates

in the absence of allowance banking. The allowance price without banking remains

lower than under scenarios including banking until 2040. After 2040, allowance

prices rise sharply due to the high abatement costs that firms would incur within the

next twenty years; therefore, firms prefer to be net borrowers of allowances in the

short run. The CRA study estimates that allowing the banking of allowances reduces

the present discounted costs of Lieberman-Warner by $100 billion.

The model predicts Lieberman-Warner would cause GDP to fall by 1.9% in 2015. The

legislation’s effects are blunted from 2025 to 2035 because of the CAFE standards

already in place in the baseline. However, by 2050 GDP falls by nearly 3.5% because

emissions caps become increasingly strict. CRA estimates the present discounted cost

at approximately $5.3 trillion by 2050.

The cost per household is estimated to be over $2,000 (or 4.5% of household income)

in 2015, falling to just above $1,000 (2% of household income) in 2025 and rising

again to $2,000 by 2050 (assuming an average household income of $50,000). CRA

also estimates the loss in employment to be nearly 4 million jobs in 2015 and over 7

million by 2050.

The prices of motor fuel, natural gas and electricity noticeably rise. In 2015, electricity

and natural gas prices are around 15% above the baseline estimate and motor fuel

might climb over 140% higher. By 2050, motor fuel and natural gas are around 90%

higher than the baseline and electricity prices might increase by as much as 60%.

4. Model: Applied Dynamic Analysis of the Global Economy (ADAGE) and

Intertemporal General Equilibrium Model (IGEM)

Organization: Environmental Protection Agency (EPA)

In this study, the EPA estimated Lieberman-Warner’s implications for GHG emissions,

the price of energy and the resulting impacts on other economic indicators. The EPA

uses two models, ADAGE and IGEM,21 to provide three baseline scenarios and seven

alternative scenarios accounting for various technology developments, energy costs

and allowance availabilities. Specifically, the EPA tests the sensitivity of estimates by

constraining the growth of technology like nuclear, biomass and carbon capture and

storage (CCS), by assuming no new international agreements going beyond the

requirements of the Kyoto Protocol, and circumscribing offset availability. All cases

permit allowance banking and base future energy prices on the Annual Energy

Outlook 2006 predictions. However, no scenarios include other GHG reduction

measures such as H.R. 6 and its vast efficiency mandates.

21

The price of a carbon allowance under its core scenario assumptions is $29 in 2015

and increases to $159 in 2050, as estimated by ADAGE, and $40 in 2015, increasing

to $220 in 2050, as estimated by IGEM.22 The scenario in which no offsets are

available estimates the highest prices of $77 in 2015 and $425 in 2050. The “high

technology” Lieberman-Warner case estimates the lowest prices at $22 in 2015 rising

to $121 in 2050.

The increased cost of energy lowers GDP by 0.18% in 2010, 0.9% in 2030 and

2.37% in 2050 according to ADAGE predictions of the baseline versus the core

scenario. IGEM predicts more dire consequences to GDP, showing a loss of 0.94% in

2010, 3.76% in 2030 and 6.9% in 2050.23 ADAGE predicts losses to total U.S.

consumption of 0.43% in 2020 and 2.10% in 2050. IGEM predicts even larger losses

of 0.66% in 2020 and 3.26% in 2050

According to ADAGE, households are estimated to lose about $446 in consumption

or 0.43% of the baseline estimate in 2015. This number increases to $3,984 in 2050

or 3.26% less than the baseline household consumption. The price of a gallon of

gasoline in 2030 is estimated to be $3.11. These estimates of future oil prices do no

take into account interruptions in supply or temporary changes in the price and only

represent the expected cost changes due to the law. Electricity prices will rise over the

forecast period from the 2005 price by nearly 20% in 2015, 30% in 2030 and then

fall back to 20% over the 2005 level in 2050.

5. Model: National Energy Modeling System (NEMS)

Group: Energy Information Administration (EIA)

The EIA is an agency within the Department of Energy and uses the National Energy

Modeling System (NEMS) for its forecasts. The model provides a baseline which

includes H.R. 6, the Energy Independence and Security Act of 2007 (EISA), estimates

of voluntary low-carbon technology adaptation (provided by the EPA), and forecasts of

energy prices provided by the Annual Energy Outlook (AEO) 2008.

The “core” scenario models the cap-and-trade system for Group I GHGs (as defined in

the legislation’s text), the bonus credit for carbon capture and storage (CCS), and some

other features present in Lieberman-Warner. The EIA study presents five other

scenarios in which differing assumptions—no international offsets available, high costs

for electricity generating facilities, limited alternatives to coal power, and both limited

alternatives to coal and no international offsets—are made.

The EIA study predicts that the price of a carbon allowance will be $30 in 2020 and

$61 in 2030 under the core scenario assumptions. The highest estimated price is

found in the strictest case—limited alternatives and no international offsets—and is $76

in 2020 and $85 in 2030.24

22

6. Model: Global Insight

Organization: The Heritage Foundation’s Center for Data Analysis (CDA)

Authors: William W. Beach, David W. Kreutzer, Ben Lieberman, and Nicolas D. Loris

The CDA uses a model developed by Global Insight. This study baseline scenario

incorporates important elements of previously enacted energy legislation25 in addition

to some critical provisions of Lieberman-Warner. However, the study only caps

carbon-dioxide emissions, rather than all GHGs covered by the legislation.

Furthermore, in further contrast to Lieberman-Warner, it does not countenance

allowance banking in any of its scenarios.26

The CDA models two different scenarios — “generous” and “reasonable” — both

constraining nuclear power production to its current level, reflecting the difficulty in

expanding production. In the “generous” case, key technologies such as carbon

capture and storage (CCS) are ready to be deployed when it becomes cost effective to

use them. Alternatively, in the “reasonable” scenario, those key technologies do not

exist within a twenty-year forecast.

The price of a carbon allowance is $49 in both the generous and reasonable forecasts

in 2015. By 2030, the price rises to $68 in the generous model and $88 in the

reasonable model. The generous forecast predicts that GDP will be 0.55% lower than

the baseline in 2016 and 2030, while the reasonable forecast predicts GDP will be

1.41% lower in 2016 and 2.17% lower in 2030. The economy loses 166,000 jobs

on net in 2016 and 461,000 jobs in 2030 according to the generous assumptions.

The reasonable forecast predicts 855,000 fewer jobs than the baseline in 2016 and

431,000 fewer jobs in 2030. In 2016, personal consumption is predicted to fall by

0.89% under the generous assumptions and 1.61% in the reasonable forecast. By

2030, the predicted loss to personal consumption has been mitigated somewhat and

is estimated to be 0.48% under the generous assumptions and 0.93% under the

reasonable assumptions.

7. Model: NEMS

Group: Clean Air Task Force (CATF)

Author: Jonathan Banks

The CATF uses the National Energy Modeling System (NEMS) and assumes that

technology improves according to the Energy Information Administration’s “best

available technology” schedule, but that biomass energy production is severely

circumscribed. CATF includes unlimited banking of allowances and uses the revenue

from the auction of allowances to fund a CCS tax credit. The CATF did not

incorporate the new low carbon fuel efficiency standards or limits on the future sources

of power like nuclear or wind.

23

The study finds that the price of a carbon allowance starts at just over $15 in 2015

and rises to $45 in 2030 as the number of carbon allowances created falls. GDP falls

by 0.7% in 2030, placing predicted economic growth four months behind the businessas-

usual case. Per capita GDP falls by 0.9% from the reference case by 2030.27

The CATF study also claims that real spending on electricity falls from 2007 to 2030

due to improvements in end-use efficiency, though the price of electricity rises.

Similarly, the price of natural gas rises, but real yearly expenditures on natural gas

increase by only a dollar from 2007 to 2030. CATF also estimates that the cost of

carbon allowances is almost completely passed through to the consumer, raising the

price of gasoline by roughly $0.10 for every $10 per ton of CO2.

While other studies show that natural gas power generation increases until the point at

which CCS becomes economic, this study shows that the subsidies for CCS cause it to

enter earlier, and thus the price of natural gas does not have to rise as much as would

be expected to otherwise. However, the study does note that if either CCS or nuclear

power is not allowed to expand for political or technological reasons, then natural gas

will fill in the gap that coal-burning plants leave. The study does not predict that coal

without CCS will be removed from the market by 2030 and will still represent around

150 gigawatts of power supply.

The Bingaman-Specter Low Carbon Economy Act, introduced in June 2007, would

create a cap-and-trade system for greenhouse gases similar to Lieberman-Warner.

Bingaman-Specter caps total covered emissions (CO2e) at 660 mmt in 2015. The

government would lower the amount of allowances created until 2050, when the

allowance cap would reduce emissions to 60% of 1990 levels (1927 mmt of CO2-e).

As the number of allowances auctioned is lowered, the price of an allowances will rise.

Many industries fear that abatement will be very costly and so the only option will be

to emit and purchase allowances, which will cause the price of an allowance to be

very high. To allay those fears, Bingaman-Specter institutes a “safety valve” called a

Technology Accelerator Payment (TAP), which essentially is an upper limit on a price

of carbon allowances.28 If the price of an allowance ever rises above the TAP price,

then the cap-and-trade system becomes essentially a tax on carbon emissions.

Regulated entities can always meet their obligation by paying the TAP price, which is

set at $12 in 2012 and grows at 5% per year in real terms. Because of this, the price

of an auctioned allowance will never exceed the TAP price. A percentage of

allowances, declining over time, is given away free to regulated entities; there are also

allowance bonuses allotted for reducing GHGs from non-covered emissions and federal

subsidies for CCS technology.

24

The EIA used the National Energy Modeling System (NEMS) to estimate the economic

impact of Bingaman-Specter. The EIA estimates two reference cases. In both cases

the EIA uses the energy price forecasts from the AEO 2007. However, in one case

they estimate the effects of the law using more optimistic assumptions on the

availability of technology. The major features of S. 1766 that the EIA tests are the

cap-and-trade limits, the TAP price, and bonus credits for CCS and non-energy

abatement. For sensitivity, the EIA tests a scenario in which the CCS bonus is only

half of what S. 1766 uses, a scenario with optimistic technology assumptions, a

scenario with supporting environmental policies like H.R. 6, a scenario with both

optimistic technology assumptions supporting policies, and a scenario with limited

alternatives to coal.

In each scenario, the TAP program is activated by 2030 and the price of an allowance

does not rise above the TAP price for that year. Only in the “high technology” case in

2020 does the EIA predict that the price of an allowance is lower than the TAP price.

Because of the TAP, total emissions in 2020 and 2030 are expected to exceed the total

covered emissions. In all of the cases except limited alternatives to coal, in 2015 GDP

is higher with S. 1766 than the predicted baseline. By 2030, most scenarios are

nearly equivalent to the baseline, but the core S. 1766 GDP is around 0.05% below

the baseline and the limited alternative scenario predicts GDP will be 0.25% below

baseline. As higher energy costs raise prices across the economy, real consumption

falls by about 0.1% from the baseline in 2030 in the core scenario and by 0.2% in the

limited alternatives scenario.

The cost of the allowances is passed forward into higher prices for gasoline, natural gas

and electricity. In 2020, gasoline prices are predicted to be 0.06% higher than

baseline, natural gas prices are predicted to be 0.7% higher, and electricity prices are

predicted to be 0.5% higher. By 2030, gasoline and natural gas are predicted to be

0.8% higher and electricity is predicted to be 0.085% higher. The limited alternatives

scenario predicts a small (less than .01 percentage points) increase in the prices of

these goods.

Metcalf et al. (2008) employ MIT’s Emission Prediction and Policy Analysis (EPPA) to

estimate the effects of the carbon taxes on CO2 emissions, welfare costs, prices of

consumer goods, tax revenues, and the effects on each income decile. Each tax

proposal varies in the level of the tax and the way in which the tax grows — or remains

constant — over time. The estimated costs to consumer welfare vary with each plan,

from a nearly 1% gain under the least stringent plan to a 2% loss under the most

stringent. Using information from the Consumer Expenditure Survey, the authors

show that the carbon tax is regressive, but a lump sum per capita return of tax revenues

is progressive. Each proposal considered taxes only on carbon, buthowever, if the

taxes are extended to cover all greenhouse gases (GHGs), there are significant

reductions in the lost consumer welfare. The authors also make comparisons between

the tax plans and comparable cap-and-trade proposals and find there is little difference.

25

The three different plans analyzed are named for their main proponents in Congress,

Dingell,29 Larson,30 and Stark-McDermott.31 The Dingell bill proposes a $13.64 tax

per ton of CO2 emitted along with a separate tax on gasoline of $0.50 per gallon;

neither tax changes over time.32 The Larson bill has an initial tax rate of $19.96 that

grows in real terms of 10% per year. The Stark bill has an initial rate of $10 that grows

in nominal terms of $10 annually. Each bill has its own plan for using the tax revenues.

Metcalf et al. (2008) predict the level of GHG emissions over time for each plan.33 The

Dingell bill is the least stringent plan and as such has the smallest effect on total

emissions. This plan keeps total emissions at current levels until 2025, when emissions

begin increasing at a rate comparable to the “business-as-usual” reference scenario. In

2050, the plan reduces emissions to 12 billion metric tons (bmt) per year from 13.5

bmt in the reference scenario. The Stark bill manages to keep total emissions constant

at today’s levels of 8 bmt per year. The Larson bill’s relatively high tax rate reduces

emissions to 4 bmt per year or roughly half of the current emission levels and 40% of

the reference emission levels.

The EPPA model predicts the welfare costs of each plan. These costs include changes

in market consumption as well as effects on leisure. The aggregate present discounted

welfare change for the Dingell plan is a 0.01% gain in welfare due to the EPPA model’s

assumption that other countries will take steps to reduce emissions that in effect lower

oil prices. The Stark plan has a slight loss to welfare of 0.03%. The Larson plan has

the largest effect of a 1.2% reduction in present discounted aggregate welfare.

The authors also model the tax plans covering non-CO2 GHGs. Since initial abatement

for any gas is easier than subsequent reductions, extending the tax plan to cover all

GHGs can result in significant decrease in the tax rate. In fact, when more GHGs are

included, the tax rate required to get to the same reduction in total emissions as under

a carbon-only tax falls under each plan. The Larson plan’s initial tax rate could be

reduced to $13.30 per metric ton of CO2 emitted, the Stark plan’s initial tax rate could

be reduced to $1.50 and the Dingell plan’s initial rate could be reduced to $12.80.

Lowering the tax rate also reduces the welfare costs of each plan. The net present

value of the aggregate welfare costs is reduced from 0.3% to 0.11%.

The tax revenue from each tax plan is substantial and can be returned to consumers in

such a way as to mitigate and even reverse the regressive nature of the carbon tax. In

2015, the potential tax revenues from the plans are $88, $69 and $126 billion per

year from the Dingell, Stark and Larson plans, respectively. These tax revenues could

account for 4% of total Federal tax revenue under the Stark plan and up to 7% under

the Larson plan. As the tax rate in the Stark and Larson plans rises over time, revenues

increase substantially; for the Larson plan in 2050 the carbon tax revenues would

account for 21% of Federal tax revenue.

Using EPPA predictions on increases in prices on electricity, gasoline and other

consumer goods of a generic $15 tax per ton of CO2 emitted and information from

the Consumer Expenditure Survey, the carbon tax is found to be regressive, but the

level of its regressivity depends on a number of key assumptions. The first assumption

is that consumers do not change behavior. If the full amount of the tax is shifted onto

the consumer, then the poorest 10% of the population faces a 3.7% reduction in

income while the richest 10% faces a 0.8% reduction in income. However, a per capita

lump sum return of the tax revenue would actually result in making the carbon tax plan

progressive. Another positive effect of a carbon tax is that the revenue can be used to

reduce taxes on labor or capital and, thus, increase overall economic efficiency.

In this paper we have provided a brief summary of estimates of the greenhouse gases

emissions’ abatement costs with particular focus on households.

GDP reduction estimates vary widely from 0.3% to 3% drop below business-as-usual in

2015 and from 1% to 10% in 2050. The timeframes of new technology development

and growth potential of existing clean sources of energy, availability of offsets (domestic

and international), and permissibility of allowance banking are likely to account for

most of these differences.

Consumption costs are affected by the same factors as GDP costs. Therefore, studies

which assume limited alternative sources of energy and/or limited offsets usually

predict smaller decreases in consumption than those which do not make such

assumptions. (Estimated costs could differ by a factor of 2-3).

Despite the differences in estimates, our analysis strongly indicates the abatement costs

could cause around a 0.8%-1% of drop in consumption below the business-as-usual

scenario. This is a conservative estimate; many studies project that costs are likely to

be even higher. Given these estimates, we can conclude that the costs of mitigation

are likely to be huge.

Our research indicates that quantifying the costs of proposed policies dealing with

climate change is a vital prerequisite to determining the appropriate course of action.

26

27

American Council for Capital Formation and the National Association of Manufacturers,

“Analysis of the Lieberman-Warner Climate Security Act (S. 2191) Using

the National Energy Modeling System (NEMS/ACCF/NAM),” 2008, at http://

www.accf.org/pdf/NAM/ fullstudy031208.pdf (5-6-2008).

William W. Beach, David W. Kreutzer, Ben Lieberman, and Nicolas D. Loris, “The

Economic Costs of the Lieberman-Warner Climate Change Legislation,” CDA08-02

May 12, 2008.

Clean Air Task Force, 2008, “The Lieberman-Warner Climate Security Act—S. 2191

A Summary of Modeling Results from the National Energy Modeling System,” http://

www.catf.us/publications/presentations/CATF_LWCSA_Short_Hill_Briefing_with_

CAFE.pdf.

Energy Information Administration, “Energy Market and Economic Impacts of S.

2191, the Lieberman-Warner Climate Security Act of 2007,” SR/OIAF/2008-01,

April 2008.

Environmental Protection Agency, “EPA Analysis of the Lieberman-Warner Climate

Security Act of 2008, S. 2191 in 110th Congress, March 14, 2008” at http://www.

epa.gov/climatechange/downloads/s2191_EPA_Analysis.pdf.

Harbenger, Arnold, (1954) “Monopoly and resource allocation,” American Economic

Review 44, 77–87.

Intergovernmental Panel on Climate Change (IPCC) Climate Change 2007, the Fourth

Assessment Report (AR4) at http://www.ipcc.ch/ipccreports/ar4-syr.htm

Lucas, Robert E. (1990), “Supply-Side Economics: An Analytical Review,” Oxford

Economic Papers, New Series, 42 (2), 293–316, (April).

Gilbert E. Metcalf, Sergey Paltsev, John M. Reilly, Henry D. Jacoby and Jennifer

Holak, “Analysis of U.S. Greenhouse Gas Tax Proposals,” NBER Working Paper

(April 2008).

Nordhaus, W.D. (2007), “A Review of the Stern Review on the Economics of Climate

Change,” Journal of Economic Literature, 45(3): 686–702.

Sergey Paltsev, John M. Reily, Henry D. Jacoby, Angelo C. Gurgel, Gilbert E. Metcalf,

Andrei P. Sokolov and Jennifer F. Holak (2008), “Appendix D: Analysis of the Cap

and Trade Features of the Lieberman-Warner Security Act (S. 2191),” 2008, MIT

Joint Program on the Science and Policy of Global Change, Report 146.

Stern, N. (2007), “The Economics of Climate Change, The Stern Review,” Cambridge

University Press.

Weitzman, M.L. (2007), “A Review of The Stern Review on the Economics of Climate

Change,” Journal of Economic Literature, 45(3): 703–724.

• Banking of allowances

• No use of foreign allowances

• Four Cases

o No Domestic Offsets; No CCS Subsidy

o CCS Subsidy

o 15% Domestic Offsets

o 15% Domestic Offsets; CCS Subsidy

• No banking of allowances

• Caps on nuclear, sequestered coal-fired (IGCC) generation, sequestered natural

gas-fired (NGCC), biomass and wind energy

• Estimated capital costs of new plant construction

• Two Cases

o Low Cost

• Greater than 20% Offsets

• AEO 2008 Oil Prices

o High Cost

• 15% to 20% Offsets

• AEO 2007 “High Profile Side Case” Oil Prices

• Tighter caps on nuclear, sequestered coal-fired (IGCC) generation,

sequestered natural gas-fired (NGCC), biomass and wind energy

• Banking of allowances and one scenario of no banking

• AEO 2008 natural gas prices, electricity demand growth, non-electric CO2

emissions

• Includes effects of H.R. 6

o CAFE standards

o Renewable Fuel Standard (RFS)

o Efficiency standards on power supplies and some appliances

28

29

• Banking of allowances

• AEO 2006

• Three baseline estimations

o Normal

o High technology

o High technology and international actions

• Seven Cases

o Encapsulate different assumptions on prices, offset availability, technology

growth, limitations on nuclear power and actions of other nations

• No banking of allowances

• Focus on CO2 only

• Two Cases

o Reasonable

o Generous

• Banking of allowances

• 6 Cases

o Baseline

o S. 2191 Core

o High Cost

o Limited Alternative to Coal

o No International Offsets

o Limited Alternatives and No International Offsets

o S. 1766

30

• S. 2191

o Banking of allowances

o 30% Offsets

o Bonuses and Subsidies for CCS

o Subsidies for geological carbon sequestration (GCS), energy efficiency

o Money to offset electric and natural gas price increases

o Constrains deployment of biomass

o Unlimited nuclear growth

o EIA’s “Best Available Technology” Case

In this section we describe how to compute the balanced growth equivalent to the

mitigation path consistent with the Lieberman-Warner bill’s. To find the balanced

growth equivalent, we calculate the fraction that consumption must decrease below the

business-as-usual model in order to provide an individual the same level of utility/wellbeing

as the abatement scenario.

In order to find that fraction we make the following five assumptions:

First, we assume a representative consumer with a constant elasticity of substitution

(CES) utility function with a risk aversion parameter of 1 or 2. Consumers typically

prefer minor changes in consumption over a longer period of time to a large one-time

change. The risk aversion parameter captures how much a consumer dislikes a volatile

consumption stream. Values of this parameter around 1 and 2 are fairly standard in

macroeconomic calibration exercises, and these figures are consistent with the

assumptions made in Stern (2007) and Lucas (1990). Estimated costs differ only in

the fourth digit when we change risk aversion parameter. Thus, we present only one

of the estimates in Table 2.

The second important assumption is that the rate of pure time preferences is about 3%-

4%. The time preference reflects the consumer’s desire, other things being equal, to

consume today rather than tomorrow. This is consistent with Lucas (1989, 1990). If

we follow Stern (2007) and assume this figure to be 0.1%, we are likely to get much

higher estimates, though many authors argue that such choice of rate of time

preference would be too low.34

31

Third, we need to account for growth in the U.S. population when computing social

welfare. We assume that population grows at 0.6% annually, following Paltsev et al.

(2008).

Fourth, we must make assumptions concerning how consumption fluctuates in the

intervening years between 2015, 2030 and 2050, since we only have cost estimates

for those specific years. We use linear interpolation between the intervals so that

decrease in consumption changes linearly between 2015 and 2030, and 2030 and

2050 to attain the estimated values presented in Table 3.

Finally, where the risk aversion parameter is 2, we assume that consumption per capita

grows at 2% annually under a business-as-usual scenario, following the Paltsev et al.

(2008) model. As we show in the technical appendix, when the risk aversion coefficient

is 1, we do not need to make any assumptions about consumption growth.35 Since the

estimates for the two risk-aversion numbers are virtually the same, this last assumption

does not make a big impact on the results.

Consider an artificial economy with a single infinitely-lived consumer who has the same

consumption stream as the aggregate consumption. We assume that the consumer

maximizes discounted sum of utilities of the form:

Here is the population in period t, where n is the rate of population

growth. Following current population growth projections (e.g., Paltsev et al., 2008),

we can assume that population grows at about 0.6% to 0.8% (Assumption 3).

is an instantaneous utility function, describing the utility derived from

consumption at a given point in time. The assumption of this particular utility form is

standard in macroeconomics and usually is assumed to be somewhere between and

4, see e.g., Lucas (1990) (Assumption 1).

is the rate of pure time preference. We assume it to be 3-4% (Assumption 2).

Also let under the business-as-usual scenario, i.e., without the costs of mitigation and

the costs of climate change, consumption grows at the constant rate g. Thus, consumption

would evolve as:

.

Following Paltsev et al. (2008), we assume that g=2% under a business-as-usual

scenario.36 Given the recent economic situation, this number probably should be

adjusted downward. (Assumption 5)

The Lieberman-Warner Act requires some abatement of GHG emissions which would

result in decreased consumption by some fraction . Table 3 provides estimates

for in 2015, 2030, and 2050. We use linear interpolation (Assumption 4) to

approximate consumption drops in other years). Thus the consumption path under

Lieberman-Warner or a similar policy becomes:

.

such that if consumption declines by fraction below the business-as-usual path, then

the consumer would get the same utility as under the LW path above: i.e., would bring

the same utility as under the LW path above:

i.e., would bring the same utility as .

This means that:

Substitution definitions of and one gets that should solve:

or equivalently:

where . Thus, we can find the necessary drop in consumption

from the following equation:

This is the equation we use to compute the estimates of consumption drops. Since

are given only at 2015, 2030 and 2050, we use linear interpolation to infer the

changes linearly between known values in years 2015, 2030, and 2050.

32

33

There is slight disadvantage to the approach above. We need to make an assumption

about the growth rate of consumption in the business-as-usual scenario. It appears that

in a particular case we can overcome this problem.

Assume that the instantaneous utility function is logarithmic. This approach has the

advantage that now we need not make specific assumptions about the path of

consumption under the business-as-usual scenario. As the derivation below shows,

under the log specification, estimated growth equivalent costs of mitigation will not

depend on the path of consumption under the business-as-usual scenario. Yet the

disadvantage is that some economists would argue that may be a bit too low.37

In this case, constant (in percentage terms) drop in consumption would solve:

Thus solves:

Using the outlined method for each of the scenarios in Table 2, we computed constantover-

time loss in consumption equivalent to estimated losses in consumption reported

by Table 3. This constant loss is to be incurred every year starting today (2008) and

going into the future up to 2050 or 2030. We stop our calculations at those time

horizons because the studies do not model impacts of abatement on consumption

beyond that timeframe.

However, most studies show that over time, consumption would drop more and more

below its no-abatement level. In this regard, our estimate provides a lower bound. Also

under a rate of time preference around 3-4%, anything happening after 2050 is

unlikely to have any sizeable impact on our figures. For the studies which stopped at

2030, we compute two estimates: one for the horizon up to 2030, the other for the

horizon up to 2050 with the assumption that damages between 2050 and 2030 are

the same as the last available estimate, the one in 2030. We see that in this case, the

estimate of the costs of mitigation will be even higher.

% This program is used to compute the impact on consumption of the

% mitigation path consistent with Lieberman Warner Climate Security Act of

% 2007 S2191

n=0.006 % population growth

g=0.02 % consumption under BAU scenario

rho=0.04 % rate of pure time preference

gamma=2 % elasticity of substitution

q=exp(n+(1-gamma)*g-rho)

t0=2008;

T=2050;

for j=1:length(data(:,1))

alpha15=data(j,1);

alpha30=data(j,2);

alpha50=data(j,3);

for t=t0:T

if t<=2015

alpha(t-t0+1)=0+alpha15*(t-t0)/(2015-t0);

else

if t<=2030

alpha(t-t0+1)=alpha15+(alpha30-alpha15)*(t-2015)/(2030-2015);

else

alpha(t-t0+1)=alpha30+(alpha50-alpha30)*(t-2030)/(2050-2030);

end

end

end

SA=0;

S=0;

for t=t0:T

SA=SA+(q^(t-t0))*((1-alpha(t-t0+1))^(1-gamma));

S=S+(q^(t-t0));

end

Percent(j)=100-(SA/S)^(1/(1-gamma))*100

%log drop

ql=exp(n-rho);

SAl=0;

Sl=0;

for t=t0:T

SAl=SAl+(q^(t-t0))*log(1-alpha(t-t0+1));

Sl=Sl+(q^(t-t0));

end

Percentl(j)=100-exp(SAl/Sl)*100

End

34

35

Group Model Scenario Index 2015 Index 2030 Index 2050

MIT EPPA No Offsets, No CSS Subsidy 1.28 1.4 1.21

15% Offsets 1.29 1.45 1.23

CSS Subsidy 1.28 1.4 1.21

15% Offsets, CSS Subsidy 1.29 1.45 0.12

ACCF/NAM* NEMS Low Cost 0.98 1.66 NA

High Cost 1.3 2.3 NA

EPA** ADAGE S. 2191 NA 1.33 1.66

IGEM NA NA NA

EIA** NEMS S. 2191 Core 1.07 1.19 NA

S. 2191 Limited Alternative/

No International Offsets 1.18 1.44 NA

CATF NEMS S. 2191 0.91 1.05 NA

* ACCF/NAM reports in the year 2014.

**EIA reports in the year 2020.

‡ Index constructed by using the EPA reported price of a gallon of gasoline in 2005 as $2.35 in 2005 dollars.

Group Model Scenario Index 2015 Index 2030 Index 2050

MIT EPPA No Offsets, No CSS Subsidy 1.61 1.81 1.61

15% Offsets 1.56 1.79 1.6

CSS Subsidy 1.6 1.57 1.61

15% Offsets, CSS Subsidy 1.55 1.57 1.61

ACCF/NAM* NEMS Low Cost 1.16 2.24 NA

High Cost 1.17 2.54 NA

EPA** ADAGE S. 2191 1.1 1.3 1.2

IGEM NA NA NA

EIA** NEMS S. 2191 Core 1.02 1.1 NA

S. 2191 Limited Alternative/

No International Offsets 1.23 1.63 NA

CATF NEMS S. 2191 NA 1.07 NA

* ACCF/NAM reports in the year 2014.

**EIA reports in the year 2020.

‡ Index constructed by using the EIA reported price of residential electricity in 2006 as 8.91 cents per kwh in 2006 dollars.

Models estimate the price changes ofn gasoline, natural gas and electricity in one of

two ways: one, the model may estimate a baseline price and the price under S. 2191,

so that a percentage change in price caused by S. 2191 can be evaluated; two, the

model may estimate a price and create an index based on a base price-typically the

2005 price. This allows readers to gauge what prices will be in the future compared

to today.

These two kinds of estimates are not readily comparable without additional

information. Since the models that present price change estimates as an index do not

report the estimate of the future baseline price, it is not possible to calculate the percent

change from the baseline caused by S. 2191. However, when studies report a

predicted future price, an index can be constructed that does allow direct comparison.

For example, the EIA reports that the price of a gallon of gas in 2030 will be $2.95

in 2006 dollars. The EPA reports that the 2005 price of gasoline was $2.34 in 2005

dollars. Adjusting the EIA predicted price for inflation using the CPI, the predicted

2030 price is $2.88. Dividing the inflation-adjusted EIA predicted price by the EPA

reported price of gasoline yields 1.23, meaning there will be a 23% increase in the

price of gasoline from 2005 to 2030 under S. 2191.

36

Group Model Scenario Index 2015 Index 2030 Index 2050

MIT EPPA No Offsets, No CSS Subsidy 1.14 1.97 1.87

15% Offsets 1.15 2.12 1.98

CSS Subsidy 1.13 1.57 1.65

15% Offsets, CSS Subsidy 1.15 1.64 1.77

ACCF/NAM* NEMS Low Cost 1.63 3.33 NA

High Cost 1.67 3.94 NA

EIA** NEMS S. 2191 Core 1.74 2.18 NA

S. 2191 Limited Alternative/

No International Offsets 2.28 3.24 NA

CATF NEMS S. 2191 NA 1.03 NA

* ACCF/NAM reports in the year 2014.

**EIA reports in the year 2020

‡ Index constructed by using the EPA reported price of a tcf of natural gas in 2005 as $7.51 in 2005 dollars.

o Index constructed by using the CATF reported price of natural gas per MMBTU in 2006 as $13.80 in 2006 dollars.

37

1. See e.g., IPCC(2007), Stern (2007).

2. http://www.barackobama.com/pdf/issues/EnvironmentFactSheet.pdf

3. Carbon capture and sequestration (CCS) in particular.

4. See Lucas (1990).

5. http://www.ers.usda.gov/Publications/EIB23/

6. Of course, in equilibrium each consumer would change its consumption bundle to

avoid being without food. This example is used to illustrate is just an illustration of

the magnitude of the impact.

7. http://www.carlton.mercedescenter.com/portal/site/DWS72100/menuitem.

2bd76a9308ae9c856a916a913aa13453/?vgnextoid=e4407aaeb9a3a110VgnV

CM10000014174335RCRD

8. Note that estimated costs in Table 4 are in real (2005 $) terms. We also make an

assumption that Mercedes does not go up in price faster than other goods.

9. http://www.piedmontcars.net/

10. We prefer to speak in terms of consumption since this measure allows us to make

welfare calculations.

11. NRDC “Forecasts of the Economic Effects of Climate Change Legislation: What

Can We Conclude?” www.nrdc.org/policy

12. “The most important finding is that, regardless of whether the study is a peerreviewed

academic or government analysis, or a non-peer reviewed industrybacked

forecast, one prediction is the same: per capita household income (as

measured by per capita gross domestic product, or GDP) will not decrease from

today’s levels. In fact, all of the projections forecast robust economic growth,

despite the limits on global warming pollution contained in the CSA. … The

studies do, however, differ in a very crucial way with respect to how they present

their results: some give the impression that average household income will

decrease from today’s level (generally, these are the industry-backed studies), while

others are careful to present their estimate more accurately as how much less a

household’s income is likely to grow as a result of the CSA.”

13. We prefer to speak in terms of consumption since this measure allows us to make

welfare calculations.

14. This amounts to reduction of emissions by 40% below its 1990 levels.

15. http://www.nrdc.org/legislation/factsheets/leg_07121101A.pdf

16. All dollar values are denominated in 2005 dollars.

17. H. R. 6 increases higher CAFE standards to 35 mpg and sets the minimum mpg

at 27.5. The bill also increases production of renewable fuels from 4 billion to 36

billion gallons and increases efficiency standards on certain household appliances,

light bulbs and electric motors.

18. All dollar amounts denominated in 2007 dollars.

19. These estimates of the price (and all resulting effects on GDP and prices of energy)

are sensitive to the assumption that there is no banking of allowances. A study

conducted by CRA International tested the effect of banking on allowance price

and found that banking increases the price in the short run and decreases the price

in the long run. CRA estimates that banking reduces the present value of the costs

of S. 2191 by $100 billion. W. David Montgomery, et al., “Economic Analysis of

the Lieberman-Warner Climate Security Act of 2007 Using CRA’s MRN-NEEM

Model,” CRA International (April 2008).

20. Denominated in 2007 dollars.

21. Applied Dynamic Analysis of the Global Economy (ADAGE) and Intertemporal

General Equilibrium Model (IGEM).

22. Denominated in 2005 dollars.

23. The elasticity of the supply labor is higher for IGEM than for ADAGE, thus the

GDP losses are larger for IGEM.

24. Denominated in 2006 dollars.

25. Such as the higher CAFE standards mandated in H.R. 6.

26. The estimates of the model are also sensitive to the number of different GHGs

covered. Metcalf et al. find that by extending policy to cover more GHGs, the

same reduction in total emissions can be achieved at a lower cost because

abatement is less expensive for small amounts of reductions for many different

gases. Metcalf, Gilbert E., et al., “Analysis of U.S. Greenhouse Gas Tax Proposals,”

NBER Working Paper (April 2008).

27. Denominated in 2004 dollars.

28. If the price of an allowance ever rises above the TAP price, then the cap-and-trade

system becomes essentially a tax on carbon emissions.

29. A summary of theDingell draft is available at proposal still in draft http://www.

house.gov/dingell/carbonTaxSummary.shtml.

30. H.R. 3416; America’s Energy Security Trust Fund Act of 2007.

31. H.R. 2069; Save Our Climate Act of 2007.

32. All dollar denominated in 2005 dollars.

33. Total GHG emissions includes the amount of CO2 emitted plus all other GHG

weighted by their potential effect on global warming.

34. See Nordhaus (2007), Weitzman (2007) for the discussion on this issue.

35. We do not even need the constant consumption growth assumption.

36. Note that we mean consumption per capita.

37. Stern (2007) assume , while Lucas (1990) sets it to 2. We consider both

values.

38

Princeton University

Solutions Consulting

Aries Analytics

Los Alamos National Laboratory

President Emeritus, Grove City College

President & CEO Emeritus

New York Academy of Sciences

CACI

University of Alabama-Huntsville

1625 K Street, NW, Suite 1050

Washington, DC 20006

202-296-9655

202-296-9714

info@marshall.org

marshall.org

March 2009