University of Southern California Dissertations and Theses

The economic and political impacts of U.S. federal carbon emissions trading policy across households, sectors and states

(USC Thesis Other)

The economic and political impacts of U.S. federal carbon emissions trading policy across households, sectors and states

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content i

THE ECONOMIC AND POLITICAL IMPACTS OF U.S. FEDERAL CARBON
EMISSIONS TRADING POLICY ACROSS HOUSEHOLDS, SECTORS AND
STATES

by

Fynnwin Prager

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PUBLIC POLICY AND MANAGEMENT)

May 2013

Copyright 2013 Fynnwin Prager

ii

DEDICATION
For Lauren, Delilah, Juliet, Andi, and Norah.

iii

ACKNOWLEDGEMENTS
Thank you to all those who have provided input and feedback during the writing of this
dissertation. Special thanks go to Professor Adam Z. Rose for his invaluable support and
guidance throughout the process. Thank you also to Professor Gary Painter and Professor Jefferey
Sellers for their insightful feedback as committee members. Thank you to Professor James
Giesecke and Professor Peter Dixon of Monash University, Australia, and to Dr. Gbadebo
Oladosu and Dr. Misak Avetisyan for their assistance on CGE modeling. I very much appreciate
the contributions of other Price School of Public Policy faculty along the way, not least Professor
Juliet Musso, Professor Detlof von Winterfeldt, Professor Dan Wei, Professor Lisa Schweitzer,
and Professor Dan Mazmanian. Thank you also to the various graduate colleagues who have
contributed through debate and feedback, not least Dr. Noah Dormady, Elena Maggioni, Jie
Zhou, Mohja Rhoads, and Pushpinder Puniha.
I feel very fortunate to have the support of a wonderful family. My parents have been an obliging
audience for updates on my dissertation progress. My wife Lauren has provided a constant
sounding board for ideas and – most importantly for me – practical matters. And my daughter
Delilah has forced me – both unknowingly and knowingly – to step back from the work and pay
her some attention.
Thank you to the Price School of Public Policy for a Graduate Assistantship to fund the first four
years of my Ph.D. program, and to the Graduate School Dissertation Completion Fellowship
program for providing a fifth year of funding. This dissertation was also supported by a grant
from the Horowitz Foundation for Social Policy, 2009.

iv

TABLE OF CONTENTS
Title Page No.
Dedication ii
Acknowledgements iii
List of Tables vii
List of Figures x
Abstract xi
Introduction
Uncertainty in Climate Change Science
Uncertainty and the Precautionary Principle
Research Design
Economy-wide Costs of U.S. Climate Policy
Climate Policy Cost Distributions across U.S. Sectors, Households, and
Regions
The Political Implications of U.S. Climate Policy Costs
Hypotheses
Aggregate Impacts
Impacts by Sector
Income Distribution Impacts
Impacts by State
Political Implications
Dissertation Summary
1
6
12
18
19
20

21
22
24
25
27
28
30
32
Chapter One: Computable General Equilibrium Modeling
A Brief History of CGE Modeling
CGE Models in Environmental and Climate Policy Analysis
USCGE Model General Characteristics
USCGE Main Model
34
36
43
46
48

v

USCGE Model Mathematical Equations 54
Chapter Two: Emissions Trading Policy Modeling
Policy Approaches for Internalizing Externalities
Translating Emissions Trading Policies from Academia to Politics
Emissions Trading Policies in Practice
Emissions Trading Policy Design Problems and Solutions
Modeling Emissions Trading Policy in the USCGE Framework
General Equilibrium Impacts of Emissions Trading Policy
Testing Assumptions
Conclusion
69
73
77
78
81
94
100
106
112
Chapter Three: Inter-Household Equity Modeling
Environmental and Climate Equity
Equity in Environmental Climate Policy
Equity Impacts of Emissions Trading Policy
Conclusion
114
115
120
128
137
Chapter Four: Inter-Regional Equity Modeling
Literature on Inter-Regional Equity
Hypotheses
Regionalization Modeling Approach
State-Level Impacts of U.S. Federal Emissions Trading Policy
140
144
146
148
152
Chapter Five: The Politics of Climate Policy
Literature on U.S. Climate Politics
Hypotheses
Political Economy Analysis
Probit Regression Results
Breaking Rank Results
Results Summary
Trends in Climate Policy Worldwide
163
167
174
175
181
183
185
186
Conclusion: Seeking Balance
Chapter Results Summary
Further Research
202
205
214

vi

Bibliography 216
Appendices:
Appendix A: List of USCGE Model Sectors, Parameters, Variables, and
Other Key Elements
Appendix B: U.S. 2007 Sector-Based Carbon Emissions Inventory
Appendix C: USCGE Model Electricity Sector Disaggregation
Appendix D: Sensitivity Tests for Equity Analysis
Appendix E: Multi-Sector Income Distribution Analysis
Appendix F: Regionalization Model
Appendix G: Senate and House Representatives who ―Break Rank‖ on
Climate Bills

237

241
257
272
275
290
301

vii

LIST OF TABLES
Table 1: Annual Fatalities per 100,000 Persons at Risk. 13
Table 2: Flow Chart; Issues, Hypotheses, Literature, Research Design and Data. 22
Table 3: U.S. top eight carbon emitting sectors, 2007. 26
Table 4: States with largest proportion of gross output from utilities, 2007. 28
Table 5: Top 5 non-Utilities carbon emitting sectors as proportion of state gross output. 30
Table 6: Emissions Trading Policies in deliberation and in operation. 79
Table 7: Summary of USCGE emissions trading model features. 95
Table 8: Emissions and output for sectors regulated in broad-based policy (narrow-based
policy sectors in bold), 2007.
99
Table 9: Comparison of broad-based and narrow-based policy results across emissions
cap levels (Case 1). Percent changes to output and emissions by sector.
101
Table 10: Comparison of broad-based and narrow-based policy results across emissions
cap levels. Emission reductions levels for groups of regulated and non-
regulated sectors.
104
Table 11: Implied net emissions trading and allowance prices ($/ton) for Broad-based and
Narrow-based policy caps.
105
Table 12: Comparison of electricity conservation case results across broad-based policy
emissions cap levels (Cases 2 and 3). Percent changes to output and emissions
by sector.
107
Table 13: Impact to output per ton of emissions reduced ($/ton CO2) for regulated sectors
and sub-groupings.
110
Table 14: Impact to output ($/ton CO2) for regulated sectors and sub-groupings. 111
Table 15: Implied net emissions trading and allowance prices ($/ton) for Case 1 scenarios. 112
Table 16: Shares of household budgets spent on commodity items, by income bracket. 123
Table 17: Summary of model features designed to capture equity impacts 126
Table 18: Labor income distribution for sectors with largest carbon emissions. 127
Table 19: Capital income distribution for sectors with largest carbon emissions. 128
Table 20: Comparison of broad-based and narrow-based policy results across emissions
cap levels. Percent changes to output and emissions by sector.
129
Table 21: Comparison of broad-based and narrow-based policy results across emissions
cap levels. Level ($Bn) and percent changes to household income distribution
by bracket and Gini coefficients.
130
Table 22: Level change in labor income distribution for broad-based policy at the 9.7%
emissions reduction cap level.
131
Table 23: Level change in capital income distribution for broad-based policy at 9.7%
emissions reductions cap level.
131
Table 24: Comparison of broad-based and narrow-based policy results across emissions
cap levels. Level ($Bn) and percent changes to household disposable income by
bracket and Gini coefficients.
133
Table 25: Percent change in household consumption of commodity groups, broad-based
policy, emissions reductions cap at 6.7%.
135
Table 26: Percent change in household consumption of commodity groups, broad-based
policy, emissions reductions cap at 12.7%.
136
Table 27: Gini coefficients for household consumption of commodity groups by income
bracket. Comparison of results for broad-based and narrow-based policy
137

viii

scenarios.
Table 28: States with largest percentage of carbon intensive sectors in total state output. 147
Table 29: Percent change impacts to output, disposable income and employment across
cap level simulations, policy regulating Kerry-Lieberman sectors.
153
Table 30: Correlations between Regional Model results, Emissions levels, and Industry
shares across all states.
157
Table 31: States with largest percentage of carbon intensive sectors in total state output. 178
Table 32: Mining, fuel and electricity sector proportions of gross output, for states with
centrist voting records, 2009.
178
Table 33: Correlation matrix for probit regression data. 180
Table 34: The influence of economic factors on Congressional roll-call climate bill
voting, 07-11.
182
Table 35: Congressional representatives ―breaking rank‖ on climate bill roll-call voting,
07-11.
183
Table 36: The influence of economic factors on Congressional representatives ―breaking
rank‖ on climate bill roll-call voting, 07-11.
184
Table 37: Major Climate Change Policy Landmarks, Worldwide 1988-2010. 189
Table 38: Climate Change Mitigation Policy in Selected OECD countries. 190
Table 39: U.S. Federal Climate Change Funding ($billions). 190
Table 40: U.S. State Climate Action Plans and Regional Climate Initiatives, 2010 197
Table A1 USCGE Code Identifiers. 237
Table A2 USCGE Sectors Code and Descriptions. 238
Table A3 Household purchases groups by commodity. 240
Table B1: US 2007 carbon emissions (Tg CO
2
) by relevant sector and NAICS code. 244
Table B2: Carbon emissions data sources and assumptions by sector. 246
Table B3: Estimation approach for ratio of fossil fuel combustion to industrial process
input energy use.
252
Table B4: Fuel combustion CO
2
emissions for industrial sectors. 253
Table B5: Industrial Emissions Coefficient calculation. 255
Table B6: Fuel Emissions Constraint Coefficients. 255
Table C1: Electricity generation energy mix by sector (without adjusting for inter-state
flows).
262
Table C2: Electricity generation energy mix by sector (adjusted for inter-state flows). 264
Table C3: Disaggregated Electricity Generation, Transmission, and Distribution columns
of the input-output table (as consuming sectors, $m).
265
Table C4: Disaggregated Electricity Generation, Transmission, and Distribution rows of
the input-output table (as producing sectors, $m).
267
Table C5: Disaggregated Electricity Generation, Transmission, and Distribution rows and
columns of the make table (transposed).
269
Table C6: Chemical manufacturing subsectors value of shipments and their relevance for
the adjustment of inputs to electricity sector generation.

269
Table C7: Proportions of chemical inputs to electricity generation by estimation category. 270
Table C8: Proportions of chemical inputs to electricity generation energy type by
estimation category.
270
Table C9: Adjustments made to fuel inputs. 271
Table C10: Water use by Thermoelectric Power Plants. 271

ix

Table D1: Comparison of broad-based and narrow-based policy results across emissions
cap levels. Level ($Bn) and percent changes to household total income by
bracket and Gini coefficients.
272

Table D2: Comparison of broad-based policy results across emissions cap levels (6-16%).
Level ($Bn) and percent changes to household total income by bracket, Gini
coefficients.
272
Table D3: Comparison of broad-based and narrow-based policy results across emissions
cap levels. Level ($Bn) and percent changes to household disposable income by
bracket and Gini coefficients.
273
Table D4: Comparison of broad-based and narrow-based policy results across emissions
cap levels. Level ($Bn) and percent changes to household disposable income by
bracket and Gini coefficients.
273
Table D5: Gini coefficients for household consumption of commodity groups by income
bracket. Scenarios across Case 2 (electricity efficiency improvements assumed
to average 3.5% per year) and Case 3 (no conservation.
274
Table D6: Gini coefficients for household consumption of commodity groups by income
bracket, across emissions reductions (6-16%).
274
Table E1: Top ten largest proportions of payments (labor and capital income) from sectors
to households.
278
Table E2: Top ten smallest proportions of payments (labor and capital income) from
sectors to households.
278
Table E3: MSIDM (labor and capital) in USCGE sector scheme, ($2007m). 279
Table E4: MSIDM (labor and capital) in USCGE sector scheme coefficients, (income
bracket proportion of total sector labor and capital income payments to
households).
280
Table E5: MSIDM (labor) in USCGE sector scheme, ($2007m). 282
Table E6: MSIDM (labor) in USCGE sector scheme coefficients, (income bracket
proportion of total sector labor income payments to households).
284
Table E7: MSIDM (capital) in USCGE sector scheme, ($2007m). 286
Table E8 MSIDM (capital) in USCGE sector scheme coefficients, (income bracket
proportion of total sector capital income payments to households).
287
Table E9: U.S. Total Income by Type of Income, 2007. 289
Table E10: U.S. Labor and Capital Income Shares by Income Bracket, 2007. 290
Table E11: U.S. Labor Income Brackets, Based on Income Shares, 2007. 290
Table F1: Regionalization Model Equations, Variable Descriptions and Data Sources. 291
Table G1: Ideology Influence on Senate Voting, Bills with Explicit Climate Change
Component.
301
Table G2: Ideology Influence on House Voting, Bills with Explicit Climate Change
Component.
301
Table G3: Ideology Influence on Senate Voting, Bills with Content Related to Climate
Change.
302
Table G4: Ideology Influence on House Voting, Bills with Content Related to Climate
Change.
302

x

LIST OF FIGURES
Figure 1: Global average temperature anomaly, annual, 1970-2011. 8
Figure 2: Global average temperature anomaly, 3-year moving average, 1972-2011. 8
Figure 3: Global average temperature, 5-year moving average, 1974-2011. 9
Figure 4: A systems-dynamic representation of relationship between GHG emission,
global warming, and climate change.
9
Figure 5: USCGE nesting structure for Emissions Trading Policy simulations. 51
Figure 6: CES Production Function Isoquants. 51
Figure 7: Economic and insured losses from great natural catastrophes worldwide,
1950-2007.
118
Figure 8: Electricity generation share of total state output. 147
Figure 9: Direct impacts to Electricity and Fuel sectors only. 148
Figure 10: Distribution of State-Level Output Results (%Δ) from Regional Model. 154
Figure 11: Change to state output, Kerry-Lieberman, 9.7% emissions reduction for
regulated sectors.
160
Figure 12: Change to disposable income, Kerry-Lieberman, 9.7% emissions reduction
simulation.
160
Figure 13: Change to state employment, Kerry-Lieberman, 9.7% emissions reduction
for simulation.
161
Figure 14: Change to state output, Kerry-Lieberman, 6.7% emissions reduction for
simulation.
161
Figure 15: Change to disposable income, Kerry-Lieberman, 6.7% emissions reduction
for simulation.
162
Figure 16: Change to state employment, Kerry-Lieberman, 6.7% emissions reduction
for simulation.
162
Figure 17: U.S. Senate Average Liberal Index across States, 2009. 173
Figure 18: U.S. House of Representatives Average Liberal Index across States, 2009. 173
Figure 19: U.S. State Climate Action Plans, 2010. 196

xi

ABSTRACT
This dissertation examines the economic and political impacts of climate policy across U.S.
households, sectors and states. This dissertation is motivated by three themes that dominate
climate policy: change, inequality, and uncertainty. The impacts of future climate changes are
uncertain. Precautionary government intervention can be justified given the potentially
catastrophic outcomes of climate change, especially for the most vulnerable communities.
However, there is concern that climate policy changes would substantially burden the economy,
and inequitably impact the poorest households and regions by reducing income and purchasing
power, and by creating more difficult transitions to new green jobs. This dissertation analyzes the
economic impacts of a U.S. federal ETP, a market-based approach used by governments
worldwide to reduce greenhouse gas emissions. Computable general equilibrium modeling is
used to estimate the distributional economic impacts across U.S. household income groups and
states. ETP can be designed to alleviate regulatory burdens on specific sectors, states, or income
brackets; this dissertation compares the economic impacts of numerous policy design options.
Uncertainty over climate change has also contributed to an increasingly contentious political
debate over policy. This dissertation examines the influence of state-level computable general
equilibrium results, as well as other state-level economic indicators, on Congressional climate
policy voting.

1

INTRODUCTION

―The climate centres around the world, which are the equivalent of the pathology
lab of a hospital, have reported the Earth‘s physical condition, and the climate
specialists see it as seriously ill, and soon to pass into a morbid fever that may
last as long as 100,000 years. I have to tell you, as members of the Earth‘s family
and an intimate part of it, that you and especially civilisation are in grave
danger.‖
James Lovelock, January 2006, London Independent

―The problem is we don‘t know what the climate is doing. We thought we knew
20 years ago. That led to some alarmist books – mine included – because it
looked clear-cut, but it hasn‘t happened… The world has not warmed up very
much since the millennium. Twelve years is a reasonable time…it [the
temperature] has stayed almost constant, whereas it should have been rising –
carbon dioxide is rising, no question about that.‖
James Lovelock, April 2012, NBC News

These quotes show that James Lovelock, a climate scientist, has moved from a position of
certainty regarding the future of the climate in 2006 to a position of greater uncertainty in 2012.
Lovelock is not arguing that the climate science is wrong, nor that policy to mitigate GHG
emissions should not be enacted. Instead, his arguments highlight the uncertainty present in
numerous areas of climate science. This Introduction chapter explores the uncertainty

2

surrounding climate change and the challenges such uncertainty poses for policy makers. This
dissertation is motivated by these policy challenges and examines questions on the economic and
political impacts of climate policy.
This is not to imply that climate change does not pose risks today or in the future.
Climate scientists and policy analysts have provided policy makers with predictions about how
climate change and policy reactions may fundamentally change societies worldwide. Global
climate change is predicted to substantially alter social life across the planet for the worse (IPCC,
2007). Increasing temperatures, rising sea levels, desertification of arable lands, and increasingly
volatile weather patterns are predicted to displace populations, damage eco-systems, and harm
global economic productivity. There is broad consensus among the climate science community
that climate change is caused by emissions of Greenhouse Gases (GHGs) from human societies
(IPCC, 2007; Oreskes, 2004). If climate models are to be believed, the worst of these changes can
be avoided by significantly modifying our lifestyles, business practices, technologies, and
industrial processes. It is argued that substantial reductions in GHG emissions are required to
avoid any further temperature increases, while continuing current emission growth rates will
cause global average warming of 0.3°C per decade. Whether or not such mitigation of GHG
emissions is undertaken, substantial adaptation of social life across the globe is likely (IPCC,
2007; Pacific Council, 2010).
The societal impact of climate change is also likely to be marked by great inequality.
Although climate change is a global phenomenon, the impacts will be localized to particular
ecologies around the world (IPCC, 2007). Poorer communities often inhabit locations vulnerable
to climate change, such as low-lying coastal areas, and have fewer resources with which to adapt
and protect themselves against localized climate changes. While developing countries are
expected to be particularly vulnerable to climate change (IPCC, 2007), poor communities within

3

developed countries such as the US are also likely to be impacted (Morello-Frosch, Pastor, Sadd,
& Shonkoff, 2009).
Some governments have acted decisively in response to the threat of climate change. The
Kyoto Protocol international agreements in 1997 appeared to be paving the way for global
commitments to reduce GHG emissions. Annex I & II countries – European countries, Russia,
Japan, Australia, New Zealand, U.S. and Canada – were required to engage in significant
domestic emissions mitigation efforts, or pay for equivalent reductions elsewhere. Most of these
countries have enacted emissions reduction policies; some have not. Australia did not ratify the
Protocol initially, yet has since committed to the emissions reduction targets. Canada took the
opposite approach; it signed and ratified the Kyoto Protocol initially, yet has recently withdrawn
from the agreement. The U.S. decision not to ratify the Kyoto Protocol was a significant setback
for the implementation and effectiveness of the agreement, as that country then contributed to
around 25 percent of global GHG emissions.
1
There have been repeated failures by the
international community to agree upon a follow-up treaty to the Kyoto Protocol.
Given the threat of climate change, why have more governments not taken more decisive
action to address climate change? There are many reasons for this, including anticipated high
costs of government intervention, the domestic politics of important countries, and perceived
imbalances between countries the requirements of the Kyoto Protocol.
2
One fundamental factor

1
U.S. state and regional level governments have attempted to step into the void left by U.S. federal policy
(Rabe, 2004). A collection of states in North-eastern US and Eastern Canada have implemented the
Regional Greenhouse Gas Initiative (RGGI) since 2008, and California adopted the Global Warming
Solutions Act in 2006, with implementation of emissions controls beginning in 2012. Yet these regions
only cover a limited proportion of U.S. total emissions.
2
In particular, the Kyoto protocol does not include binding targets on large developing countries such as
Brazil, India, and China. Climate policy opponents in Annex I & II countries argue that their domestic
industries will be placed at an unfair advantage with respect to imports from countries such as Brazil, India,
and China that were not required to reduce emissions under the Kyoto Protocol. On the other hand,
representatives of these developing countries argue that they should be allowed to develop without

4

is the uncertainty surrounding climate change science and policy. Uncertainty permeates climate
science and the policy debates it influences. Climate change is likely, and most climate scientists
believe it is human-generated (IPCC, 2007). However, the IPCC also reports that the magnitude,
location and timing of climate changes are less clear. Less clear still is the extent to which
reduction of current GHG can halt the progression of climate change. There is also uncertainty
about the optimal strategies and resource allocations that governments should use to respond to
climate change.
This uncertainty raises important questions about government climate policy. Should
governments follow the precautionary principle and reduce GHG emissions in the short term and
hence attempt to avoid the most catastrophic harms? Or should they develop resilient strategies
that respond and adapt only to known climate changes on a case-by-case basis? These questions
lie at the foundation of this dissertation. It is widely agreed that a precautionary principle should
be adopted when the potential outcomes are potentially catastrophic (Sunstein, 2005), and many
argue it should be applied when the poorest and most vulnerable communities are impacted
(Morello-Frosch, Pastor, & Sadd, 2002). However, there is the danger that a precautionary
approach would result in greater costs to society as a whole, and even to the most vulnerable
communities (Wildavsky, 1988; Sunstein 2005).
This dissertation informs these fundamental questions by focusing on U.S. domestic
climate policy. The U.S. is chosen as a case for study because of the critical role the country plays
in global climate policy regimes. As the second largest emitter of GHGs worldwide, a large-scale
U.S. emissions reduction policy would bring the vast majority of developed-world – and indeed
the majority of global – emissions under policy controls. While the influence of China – now the

constraints, as developed countries did. They also argue that they should not be held accountable for the
historical emissions of developed countries.

5

world‘s largest GHG emitter – on global climate policy agreements should not be ignored, the
lack of a U.S. emissions reduction policy continues to hamper a global collective action on
emissions reduction.
The U.S. is also a fascinating case because of the numerous economic and political
interests which influence climate policy decision making. Ideology dominates public opinion and
political decision making, with liberals supporting climate policy and conservatives opposing it.
Yet economic interests contribute to such perspectives. Liberals often point to dependence on
fossil fuels and the need to transition to a renewable-energy economy, and the economic threats
of climate change to particular regions. Liberals are often also concerned that the costs of climate
policy would be disproportionately borne by poorer households or industries vulnerable to global
competition. Conservatives often highlight the economic costs of emissions reductions programs
to the economy as a whole and fossil fuel sectors in particular. Put another way, alongside
―diffuse‖ costs – i.e. those spread across the nation – are ―concentrated‖ costs, which affect
particular interest groups (Arnold, 1992).
This dissertation examines whether U.S. climate policy would be costly to society as a
whole, as well as to vulnerable economic sectors, household income groups, and regions. The
economy-wide changes required under emissions trading policy (ETP) – a common mitigation
policy tool – are examined using Computable General Equilibrium (CGE) modeling. These
changes include an emissions reduction constraint, allowance trading between regulated sectors,
technological changes to greenhouse-gas intensive sectors and electricity consumption, as well as
the ripple effects caused by changes in prices for their suppliers and changes to demand from
their customers. Distributional economic impacts are explored with respect to income and
consumption effects for household income groups, as well as output, employment, and disposable
income effects across 50 US states (including Washington D.C.). In addition, this dissertation

6

makes connections between these economic impacts of climate policy and the voting records of
political representatives in Congress.
UNCERTAINTY IN CLIMATE CHANGE SCIENCE
The political disagreements over U.S. climate policy go deeper than economic concerns.
Climate change is a ―wicked problem‖ (Rittel and Webber, 1974), as there is no consensus among
U.S. policy makers on the problem, let alone the solution. It has been suggested that so-called
―climate skeptics‖ deny there is a problem because they dislike the proposed solutions. There are
reports of anonymous conservative donors – alongside visible actors such as the industrialist
Koch Brothers and the oil company Exxon Mobil – supporting climate-skeptic research and
advocacy in both the climate science and policy analysis fields (Goldenberg, 2013). This would
suggest that those with an interest in blocking climate policy are also engaged in supporting
research which undermines climate science. However, there is an important distinction to be
made between an outright denial – and indeed active undermining – of climate change science,
and a science-based skepticism that highlights the limits of our knowledge.
As James Lovelock has highlighted in the quotes at the top of the chapter, uncertainty
permeates both climate change science and the policy debates it influences. This uncertainty
contributes to the ―wicked‖ nature of climate policy, yet also raises fundamental questions about
whether precautionary government intervention should be undertaken. While there is broad
consensus among the climate science community that global warming is caused by human
pollution, the magnitude and timing of future changes are less clear (Pacific Council, 2010; IPCC.
2007). In particular, the relationship between GHG concentrations and global warming is
obscured by complex and interacting climate features, as well as fundamental difficulties in
predicting future events.

7

The current trend of global warming is uncertain. While the 2000s decade was the
warmest decade on record, there was not a ―record‖ year of temperatures since 1998 according to
the datasets produced by the University of East Anglia‘s (UK) Climate Research Unit
(HadCRUT3, 2012)
3
. This raises the possibility that the global warming trend has flattened over
the last decade. As shown in Figure 1 below, there has been a clear trend of increasing average
global temperatures between the 1970s and 2000s. However, predicting a value for future average
global temperatures would depend on the data used (the variation between the three main datasets
is evident in Figure 1), the years included in the analysis, and the particular calculation approach.
For example, the 3-year moving average (Figure 2) calculation approach suggests that there was a
dip in global average temperature for the years 2006-2008, but that the trend has been increasing
since then (for all three datasets). In contrast, the 5-year moving average (Figure 3) suggests a
declining trend since 2005.
This data may suggest that, counter to climate theory, annual average temperatures are
not influenced by increasing GHG concentrations. The fundamental theory of climate change
appears to be simple: increased GHGs create a ―greenhouse effect‖ that traps heat in the earth‘s
atmosphere; increased GHG concentrations cause increased temperatures. However, as shown in
Figure 4, the relationship between human-generated emissions and climate change is a complex
one. Temperature changes (depicted in purple and red in Figure 4) interact with numerous factors,
or ―feedback systems,‖ that can both enhance and lessen the influence of GHG concentrations on
global temperatures.

3
2012 was the hottest year on record for the U.S., but estimated to be the fifth hottest globally. As shown in
Figure 1, higher values than 1998 have been observed in the two other major datasets, the US NASA
Goddard Institute for Space Studies (NASA, 2012a) and the US National Oceanic and Atmospheric
Administration (NOAA, 2012), for the years 2005 and 2010, though the margins are small.

8

Figure 1: Global average temperature anomaly, annual, 1970-2011

Figure 2: Global average temperature anomaly, 3-year moving average, 1972-2011.
-0.3
-0.2
-0.1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
Global Averager Annual Temperature Anomoly
Year
HadCrut3 NasaGiss NOAA
-0.3
-0.2
-0.1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
Global Averager Temperature Anomaly
Year
HadCrut3 NasaGiss NOAA

9

Figure 3: Global average temperature, 5-year moving average, 1974-2011

Figure 4: A systems-dynamic representation of relationship between GHG emission, global
warming, and climate change (Fiddaman, 2010).
-0.2
-0.1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
Global Averager Annual Temperature Anomoly
Year
HadCrut3 NasaGiss NOAA

10

Clouds prevent substantial amounts of sunlight from reaching the Earth‘s surface and can
hence influence global warming significantly. However, there is uncertainty about the precise
impact clouds have on the climate system (NASA, 2012b). Carbon dioxide, the most common
human-generated GHG, is constantly fluctuating. Natural processes in the carbon cycle remove
about half of the human-generated carbon dioxide emissions each year. However, there is
uncertainty as to the precise source of this reduction (NASA, 2012b). Other forms of pollution,
namely aerosols, dust, smoke, and soot, also produce uncertain impacts on global temperatures.
Sulfate aerosols, emitted by volcanoes and during coal and biomass combustion, are known to
decrease global temperatures. Yet there is great uncertainty regarding the impact of other non-
GHG particles (NASA, 2012b).
It is commonly asserted that reducing GHG concentrations would reduce global
temperatures. However, the timing and extent of the changes is highly unpredictable. Reducing
GHG concentrations would mean reversing a trend that began around 200 years ago with the birth
of the industrial revolution. This trend was not matched with an immediate and correlated
increase in global average temperatures. Instead, global temperatures began to increase in the
early 20
th
century. Adding further complexity to the relationship, there was also a sustained
cooling period between the 1940s and 1970s, which some have argued was caused by high levels
of non-GHGs such as aerosols, dust, smoke, and soot in the atmosphere (NASA, 2012b).
However, perhaps most importantly, GHG emissions intensities have only been observed to have
increased over the past 200 years. Hence there is no direct empirical basis for judgments as to
how long reductions in – or stabilizations of – GHG concentrations would take to influence
global average temperatures.
The impact of currently observed global warming on the climate is also uncertain. Ocean
circulation plays an important role in regional climate systems, and may serve to reduce regional

11

temperatures despite global temperatures increasing. For example, climate change is predicted to
disrupt the North Atlantic drift section of the Gulf Stream, which currently warms Northern and
Western Europe. Such a change could significantly reduce average temperatures in Europe.
However, there is uncertainty about the implications of ocean circulation worldwide as global
ocean data sets were present from only the 1990s onwards. Changes to precipitation worldwide
are another uncertainty; there are broad variations between climate model predictions for specific
regions. Finally, there is notable uncertainty about the magnitude of sea level rises. The IPCC
(2007) was unable to ―provide a best estimate or an upper bound for sea level rise‖ over the
coming century because of uncertainties surrounding arctic ice sheets. The Greenland ice sheet is
currently declining and contains the equivalent of 5-6 meters of sea level. The Antarctic ice sheets
are much larger, but are not currently declining at the same rates (NASA, 2012b).
Each of these uncertainties about climate change highlight more fundamental questions
about the extent to which we are able to predict the future. In the real world, there are no crystal
balls able to predict future outcomes; there are simply predictions with differing degrees of
uncertainty. These predictions are dependent on the data observed and our understanding of that
data. There are numerous approaches to analyzing historical data when projecting future events,
including moving averages and autoregressive sequences (among others) of single variables, as
well as more complex models based on multiple variables. Each of these approaches could render
different projections, and each set of results would contain some level of error, or uncertainty.
The question then becomes, which one would be correct? Even the model with the highest
projection accuracy for past events could be wrong. And that error could change over time – what
appeared to be a small inaccuracy previously could in the future expand to be a significant error
that in fact reflects an omitted variable or variables. Compounding the problem of changes over
time is the potential for new ―Black Swan‖ events to appear; just because an event has not

12

happened before does not mean that it will not happen in the future. Hence, basing projections of
the future on past events alone can be undermined by ―unknown unknowns.‖
UNCERTAINTY AND THE PRECAUTIONARY PRINCIPLE
The limitations of predicting the future make decision making difficult in many policy
areas, yet the high degree of uncertainty in climate policy is particularly problematic. Even if we
assume that there is agreement about the problem of climate change, the presence of uncertainty
limits policy decision making. This problem is reflected the broader policy debate regarding the
precautionary principle. The precautionary principle states that if the future outcomes of an
activity are possibly harmful, there is justification to control that activity, despite any uncertainty.
With respect to climate change, the precautionary principle would advise for governments to
reduce GHG emissions now to avoid the potentially catastrophic outcomes, despite the
uncertainty over climate science and the limits on economic growth that climate policy could
create. The precautionary principle is arguably a central pillar of many environmental campaign
organizations, and it gained particular traction in the field of environmental law and policy during
the 1980s and 1990s. For example, Principle 15 of the 1992 Rio Declaration – a seminal
international environmental agreement – states: ―Where there are threats of serious or irreversible
damage, lack of full scientific certainty shall not be used as a reason for postponing cost-effective
measures to prevent environmental degradation‖ (Percival, 2006: 21).
The precautionary principle subsequently came under criticism from legal and policy
scholars (Sunstein, 2005; Wildavsky, 1988). Critics argue that the precautionary principle
approach is incoherent. Numerous facets of daily life are risky and are subject to little or no
regulation, yet the precautionary principle is applied to risks that are of relatively limited risk to
human health. For example, the risks presented in Table 1 are subject to differing levels of

13

regulation, yet only cell phone use while driving has been subjected to an outright ban. In
contrast, according to Wildavsky and Wildavsky (2008), ―the U.S. Environmental Protection
Agency sets a safety threshold [for regulation of potentially harmful substances] of one additional
death in a million given a lifetime of exposure.‖
Table 1: Annual Fatalities per 100,000 Persons at Risk
Activity/Event Death Rate
Texas death row inmate 5,000
Space shuttle astronaut—shuttle accidents only 2,000
U.S. population—all causes (2003) 840
Smoking one pack of cigarettes per day (all causes) 300
Sport parachuting 200
Smoking one pack of cigarettes per day (cancer) 120
Lumberjack 118
Motorcycling 65
Sky diving 58
Farming 28
Mining 27
Police officer 20
Motor vehicle accidents 15
Firefighter 10
Boating 5
One drink with alcohol per day—cancer and other adverse
consequences only
5
Injury at work—all jobs 4
Manufacturing 3
Hunting 3
Fires 1
4 tbsp. peanut butter per day (aflatoxin) 0.4
Cell phone use while driving 0.15
Floods 0.05
Large meteorite impact 0.04
Lightning 0.02
Chlorinated tap water 0.01
Accidental reentry of a deep space probe 0.00000003
Source: Wildavsky and Wildavsky (2008).
Looking solely at the level of risk, the regulation of risks may appear to be incoherent.
Yet numerous other factors contribute to regulatory decisions, including the role of individual
choice (the risks of sporting activities are self-imposed while environmental harms can influence

14

individuals and communities that do not cause or benefit), the cost of regulation,
4
and political
representation.
5
For example, most of the examples provided in Table 1 are individual behavioral
choices (whether "leisure" such as smoking, motorcycling, or skydiving, or compensated labor
such as lumber-jacking, mining, or police officer) as opposed to the externally imposed harms
that are possible from environmental pollution, for which the individual may have no choice over.
The ability of individuals to make informed choices to avoid risks is an important legal principle
that is not sufficiently addressed by those opposing the precautionary principle. As such,
comparisons across risks are less helpful than case-by-case analyses of the risks and economic
costs, the stakeholders, and the political contexts. For climate change, the average risks in terms
of fatalities are projected to be relatively low when compared with other risks. However impacts
to specific areas and vulnerable communities are potentially high and the economic impacts of
climate change are potentially very serious.
6

The critics of the precautionary principle also emphasize the dangers of over-regulation.
The potentially harmful impacts of climate change, they would argue, should not lead us to
regulate polluting industries to the extent that the costs of policy outweigh the benefits. Taking a
broader perspective, Wildavsky and Wildavsky (2008) argue that the economic benefits brought
about by technologies such as industrial chemicals and nuclear power are worth the associated
risks. These economic benefits contribute to safer, healthier lives, while the technological risks

4
For example, tobacco consumption is very risky and is costly to society. However, governments are able
to leverage substantial revenues by taxing the sale of tobacco products. This is partly because the addictive
quality of tobacco makes demand for tobacco highly inelastic.
5
For example, vehicle fuel arguably faces similar demand elasticity schedules to tobacco; prices must rise
substantially before drivers will alter their fuel consumption. In Europe, policy makers appear to have
followed the same logic as with tobacco and leveraged relatively high fuel taxes. However, in the US
vehicle fuel tax rates are substantially lower. This difference can largely be explained by the different
political climates in the two regions. In the U.S., the opposition to vehicle fuel taxes is partly driven by
business interests. However, as political scientist Mark Smith (2000) observes, business interests require
popular support to be effective in influencing policy making.
6
The complexity of climate change means that risk levels are dependent on numerous assumptions,
including which climate-related events cause or influence fatalities.

15

are relatively small. Interestingly, in the climate policy field, this argument has more relevance to
nuclear technologies (which are ―clean‖ with respect to GHGs) and renewable technologies such
as wind farms and hydro power, which have the potential to impact local environments and
ecologies. While climate proponents may favor these technologies, there are often local
oppositions to the placement of these technologies because of the impacts they have on the local
environment and communities.
Moreover, critics argue that the economic costs of over-regulation are often borne by the
poor and vulnerable communities that pre-cautionary approaches aim to protect. This is
particularly salient for climate policy, as it could inequitably impact the poorest households
through reductions to income, purchasing power and limited opportunities in the new green
economy that would likely develop. Recent research on this issue has shown that climate policy
design – especially the redistribution of policy-generated revenues – can offset harms to the
poorest groups, though this may come at a cost to overall economic efficiency (Burtraw et al.,
2009; Hassett et al, 2009; Roland-Holst 2010; Rose et al, 2011).
In any case, while the broader perspective proffered by the critics of the precautionary
principle is important for determining the broader direction of risk regulation, it is injudicious to
use this as a determinant for specific policy decisions. In other words, the dangers of over-
regulation should be heeded, but the presence of possible dangers does not mean that every risk
should not be regulated.
7
So what criteria can be applied to assess whether a precautionary
approach is appropriate for a given uncertain problem such as climate change? Beyond the level

7
For example, legal scholar Robert Percival argues that ―[t]he precautionary principle does not require that
innovation come to a halt whenever any risks may be conjured. Properly understood, the precautionary
principle is neither incoherent, paralyzing, nor a prescription for overregulation. Rather, it cautions that
regulatory policy should be pro-active in ferreting out potentially serious threats to human health and the
environment, as confirmed by the history of human exposure to substances such as lead and asbestos.‖
(2006: 22)

16

of uncertainty surrounding the problem, three questions are critical. First, are the outcomes of
adopting the precautionary approach minimal? Second, are the outcomes of doing nothing
potentially catastrophic? Third, are poor and vulnerable communities likely to bear the burden of
either doing nothing or policy intervention? These questions are listed in this order for a reason. If
adopting the precautionary approach has a minimal impact, then government intervention is
advisable, even if there is great uncertainty surrounding the problem.
8

If an uncertain problem is potentially catastrophic, there is also a justification for
government intervention, even if the costs are more than minimal. If, for example, a meteor may
be on course to impact the Earth, causing catastrophic destruction, there is good reason for
governments worldwide to attempt to do something about it, even if the costs are large. This is
the argument used in favor of counter-terrorism measures on transportation systems. Even though
the additional security checks are costly in terms of government resources and business travelers‘
time – especially given the very low likelihood of an attack occurring in any particular location –
governments aim to avoid the potentially catastrophic costs of a terrorist attack, as well as the
reduce the fear-induced business losses. This same argument could be made with respect to the
catastrophic potential of climate change.
There is a further justification for government intervention if an uncertain problem has
the potential to cause substantial harm to poor and vulnerable communities. Even the
precautionary principle critic Cass Sunstein suggests that this justification is ―among the strongest
points in favor of aggressive regulation of GHGs‖ (2005: 50). Primarily this justification is based
on the social equity principles that are fundamental to democratic political systems. Building
additional levees in New Orleans would have been costly and the outcomes of Hurricane Katrina

8
What constitutes ―minimal‖ is obviously an important issue here – in a more recent book, ―Nudge:
Improving Decisions About Health, Wealth, and Happiness,‖ Cass Sunstein and Richard Thaler (2008)
outline a broad agenda for what could be described as minimal government intervention.

17

were not catastrophic for the U.S. society as a whole. But the fact that so many poor and
vulnerable individuals died as a result of Hurricane Katrina is deeply disturbing from the point of
view of social equity. In addition, because democratic systems have tended to develop structures
that respond to the plight of poor and vulnerable communities – through systems such as
unemployment benefits, health care for the poor, food stamps, etc. – there are more practical cost
saving rationales for government intervention, even when the problem is uncertain. By protecting
against the potential for dramatically costly events, governments acting under uncertainty are
nonetheless ruling out outcomes that would create dire budgetary outcomes.
It is important to note that scholars such as Aaron Wildavsky (1988) and Cass Sunstein
(2005) argue that an alternative approach is desirable. Instead of the risk-averse precautionary
approach that prevents harm from occurring, they argue that a resilience-based approach of
responding to adversity with resource capacities that allow societies and individuals to respond to
specific problems with effective solutions in a piecemeal and incremental manner. They argue
that successful solutions are more likely to appear from a system of decentralized trial-and-error,
rather than a precautionary approach whereby error is precluded. This approach is also more
pragmatic with respect to political feasibility and the many ―problems‖ which publics and policy
makers must prioritize on a daily basis. Numerous policy issues are competing for attention as
they are debated and implemented across numerous venues – i.e. research institutions, mass-
media, inter-personal interactions, government bureaucracies, legislative offices, and the courts.
This reality suggests that more focused, cheaper, and reactive policy programs are more likely to
be enacted, and more likely to be implemented effectively.
For climate change, the resilience approach would align with the perspective of scholars
such as James Lovelock (2006), who originally argued that the Earth‘s climate has moved past
the tipping point for inevitable and dramatic climate change, and hence strategies of adaptation

18

and resilience are the only options available to policy makers. Adaptation is an important element
of the climate change debate. Adaptation and mitigation are often cast as contradictory
approaches, yet in many cases they are complementary. However, addressing both approaches
simultaneously is beyond the scope of this dissertation. Instead, by focusing on mitigation, this
dissertation aims to first establish whether mitigation policies are significantly economically
harmful, both on aggregate and to income groups and regions of the economic. Nonetheless, the
analytical methods used in this dissertation can also inform adaptation policy questions.
RESEARCH DESIGN
So what does this all mean for climate policy? Climate policy is clearly operating under
high levels of uncertainty. This uncertainty raises many fundamental questions. Should
governments do nothing or respond to the possible harms of climate change? Is the precautionary
principle justified given the potentially catastrophic outcomes of climate change, especially for
the poor and vulnerable? Or is wholesale government intervention sufficiently costly that a
piecemeal approach of adapting to known problems is more desirable? If they intervene, should
governments focus on mitigation or adaptation, and which policy tools should be implemented?
As identified above, there are three critical questions when determining the whether the
precautionary principle should be adopted for a policy area:
1. Are the outcomes of adopting the precautionary approach minimal?
2. Are the consequences of doing nothing potentially catastrophic?
3. Are poor and vulnerable communities likely to bear the burden of either doing
nothing or policy intervention?
As discussed above, the consequences of doing nothing are potentially catastrophic, especially for
poor and vulnerable communities across the world. However, the uncertainty surrounding

19

outcomes in this respect encourages further examination of the policy cost side of the questions.
This dissertation aims to contribute to this debate by focusing on two key policy impacts:
1. The aggregate, economy-wide, costs of climate policy in the U.S.;
2. The distribution of policy costs across sectors, households, and regions in the U.S.
ECONOMY-WIDE COSTS OF U.S. CLIMATE POLICY
This dissertation explores whether government intervention is cost-prohibitive by
exploring a common mitigation policy approach – ETP. ETPs establish an emissions cap for each
industry.
9
The cap determines the allowances granted to industrial entities. At the end of each
phase, allowances equivalent to actual emissions must be returned to the administrator. If the
regulated industrial facility is not able to reduce emissions down to their allowance allocation
(cap level), instead of being fined – as would be the case in a command and control approach –
that facility could purchase emissions allowances from other regulated industries which have
reduced emissions to below their cap level. ETPs therefore place a regulatory burden on polluting
sectors, especially those for which emissions reduction is relatively more costly, i.e. above the
market price for emissions allowances. Conversely, some regulated entities may benefit if they
can invest in emissions reductions which are below market price. While ETPs have been
implemented in the European Union and in sub-regions of the U.S. (ten North-eastern States, and
California), there have been numerous failed attempts to enact federal level ETPs in recent years.
The economic impacts of ETP are explored through CGE modeling of a U.S. federal
ETP. CGE models are one of the most commonly used analytical approaches in the
environmental policy field; a detailed discussion of CGE modeling is provided in Chapter Two.
This dissertation refines the USCGE model, which has been developed by Gbadebo Oladosu and

9
These caps can be politically or administratively (e.g., an algorithm based on historical emissions over a
given period) determined during the policy design phase, and can be iteratively updated over time.

20

Adam Rose over the past two decades (Oladosu, 2000; Rose & Oladosu, 2002; Oladosu & Rose,
2007; Rose, Oladosu, Lee, & Beeler Asay, 2009).
CLIMATE POLICY COST DISTRIBUTION ACROSS U.S. SECTORS, HOUSEHOLDS,
AND REGIONS
The USCGE model is geared to explore the three aspects of the policy cost distribution:
the policy costs across sectors, household income brackets, and regions of the U.S. The
identification of economic impacts across sectors is a central element of CGE analysis. Chapter
Three presents the aggregate and sector impacts of ETP. This dissertation places particular
emphasis on disaggregation of the electricity generation sector into numerous fossil and non-
fossil fuel sectors; electricity generation is often included as a single sector in government data
sets and CGE models.
Chapter Four presents the inter-household impacts of ETP. Analysis of distribution across
households income brackets is examined in terms of both income and consumption effects.
Central to the income distribution analysis element of the USCGE model is a Multi-Sector
Income Distribution Matrix (MSIDM) that identifies the spread of labor and capital incomes
across each sector of the economy, primarily using occupation and income data from U.S.
government sources (BLS, 2010; IRS, 2010). Consumption across household income brackets
respond to a combination of income changes, and the price changes and substitution effects for
consumer goods.
Chapter Five presents the distribution of policy impacts across US states, which is
determined on the basis of an established regionalization approach (Dixon and Rimmer, 2004;
Dixon, Rimmer, and Tsigas, 2007). This approach combines sector-level results and various

21

national level results with state-level economic indicators including state-by-sector output
matrices and inter-state trade data.
THE POLITICAL IMPLICATIONS OF U.S. CLIMATE POLICY COSTS
Chapter Six explores the politics of U.S. climate policy. Climate change uncertainties
have also contributed to an increasingly contentious public debate on climate science and policy.
Such conflicts raise numerous important questions. How are particular policy choices likely to
impact particular constituencies and interest groups? How popular are these policy choices likely
to be among the electorate and influential interest groups? Assuming economic interests can be
known, are voters likely to support policies that represent their economic interests, or do other
factors such as ideology or cultural cognition dominate? If economic interests are relevant to
voters, are some economic group interests (e.g. class, ethnicity, sector, region) more important
than others? What role does the knowledge generated in the climate change economic analysis
literature play in the policy process?
This dissertation explores to what extent the economic interests of citizens and powerful
interest-groups fuel climate policy disagreements, as opposed to other factors such as ideology.
CGE results are also used to explore the influence of economic impacts on federal climate policy
making.
10
Chapter Six runs probit regressions to identify the impact of ideology, environmental
policy, state-level economic factors, and state level emissions consumption on Congressional roll-
call voting on climate policy bills between the years 2007 and 2011. While previous analyses by
Anderson (2011) and Chupp (2011) have provided insightful analyses of Congressional roll-call

10
The data points here include the presence and activity of state-level EPAs or equivalent agencies, the
presence and extent of state-level climate action plans, and the voting records or debating positions of
federal representatives in recent climate change bills, including Leiberman-McCain, (2005), Kerry-Boxer
(2009), Waxman-Markey (2009), Kerry-Lieberman (2010), as well as the Upton and Inhofe bills of 2011.

22

voting for environmental policy as a whole, there have been no analyses of climate policy
specifically.
HYPOTHESES
An overview of hypotheses with corresponding literature, research design and data are
provided in Table 2 below. These hypotheses relate to five areas of the dissertation,: 1) Aggregate
Economic Impacts; 2) Impacts by Sector; 3) Income Distribution Impacts; 4) Impacts by State;
and 5) Political Implications. These five areas of inquiry (or the five hypotheses) are discussed in
this section, and operationalized in the research design, methods and data section below.
Table 2: Issues, Hypotheses, Research Design and Data
Issues H: Hypotheses
R. Research Design
D. Data
Aggregate
Economic
Impacts
H1: Aggregate results will be less than negative 1% impact on output (Rose &
Dormady, 2011; Rose, Wei & Prager, 2012).
R. Policy shock to the CGE model (examples of CGE: Conrad & Schmidt,
1998; Jorgenson & Wilcoxen, 1990; Oladosu & Rose, 2007).
D. U.S. Emissions Inventory Data (EPA, 2007); U.S. Social Accounting
Matrix (IMPLAN, 2007); Further sector details (U.S. BEA, 2010; U.S.
Census, 2010).
H2: Coase Theorem: Aggregate results will be largely unaffected by the
allocation schedule (Coase, 1960).
R. Comparison of policy shocks to the CGE model.
D. As above, with addition of data and modeling assumptions from
emissions trading literature and proposed and enacted policies.
Impacts by
Sector
H3a: GHG-intensive sectors are likely to bear the greatest regulatory burden.
H3b: Downstream and upstream sectors would absorb some of this burden.
R. Policy shock to the CGE model.
D. As above.
H4a: Green technology and government sectors are likely to benefit relatively
more from ETP.
H4b: Sectors importing and exporting green technology and GHG-intensive
substitutes are likely to benefit relatively from ETP.
R. Policy shock to the CGE model based on evidence from the literature as
to the economic impact of ETP or similar policy.
D. Emissions trading literature.

23

Income
Distribution
Impacts
H5a: Coase Theorem: Allocation schedule will influence distributional impacts
(Coase, 1960).
H5b: Income measures of distribution will be more regressive than consumption
measures (Parry & Williams, 2010; Hassett, Mathur, & Metcalf, 2009).
H5c: Relative single-factor measures of wealth distribution (e.g. Gini coefficient)
will highlight different concerns to measures which examine changes to single
brackets only (e.g. analysis of the relative gains to the lowest or highest bracket)
(Rose, Wei & Prager, 2012).
R. MSIDM and Consumption Distribution Matrix (CDM)
D. Occupations and employment by sector, wages by sector, consumer
expenditure by sector (BLS, 2010); Income and tax (IRS, 2010);
Households per income bracket (U.S. Census, 2010); Sectoral output
(BEA, 2010; IMPLAN, 2007).
Impacts by
State
H6: States with GHG-intensive sectors are likely to bear a relatively greater
regulatory burden (Burtraw et al., 2009; Hassett et al, 2009; Pizer, Sanchirico, &
Batz, 2010).
R. Multi-Sector State Matrix (MSSM) with additional metric of density of
influential sectors within state.
D. Sectoral data across states (BEA, 2010; BLS, 2010; Census, 2010).
Political
Implications
H7a: States with relatively greater economic harms from climate policy are less
likely to have federal representatives supporting climate policy and ETP (Pizer et
al., 2010; Shipan & Lowry, 2001; Zahran et al., 2008).
H7b: States with relative economic harms from climate change are more likely to
have federal representatives supporting climate policy and ETP (Goulder, 2003;
Victor, 2003).
H7c: States with relatively large harms to influential GHG-intensive sectors are
less likely to have federal representatives supporting climate policy and ETP.
H7d: Direct regulatory impacts are likely to have greater influence than overall
(direct plus indirect) economic impacts on state level policy making and federal
level representation of state interests.
R. CGE model plus MSSM related to political hypotheses and political
indicators
D. As above, plus environmental policy and political science literatures,
along with data from state-level EPA websites, state level climate action
plans (CCS, 2011), and congressional voting records (League of
Conservation Voters, 2011).

24

AGGREGATE IMPACTS
H1: Aggregate results will be less than negative 1% impact on output (Rose & Dormady, 2011;
Rose, Wei & Prager, 2012).
Aggregate economic impacts of climate change policy have varied widely across recent
studies (Rose & Dormady, 2011). Much of this variance can be explained by the assumptions
adopted within the economic impact modeling. Those studies that render the largest negative
economic impacts assume the highest price per ton of emissions allowances or emissions tax rate,
as well as the largest reductions in emissions required by the policy (Rose and Dormady, 2011).
Less dramatic assumptions produce milder economic impact results, with the majority of recent
analyses suggesting that slightly positive or slightly negative results depending on the particular
context. For example, recent analyses of the California Global Warming Solutions Act (AB32)
suggested mild economic benefits from the policy (Rose, Wei & Prager, 2012). However, a
region with higher dependence on fossil fuels and less flexibility in transition to GHG intensive
would likely experience mild negative impacts to the economy from an ETP.
It is therefore hypothesized that a U.S. federal ETP would induce a mild negative
economic impact. Some states would benefit while others would be harmed, with the latter
outweighing the former. Some states have already pursued climate change policies – including
regional ETP (the RGGI is under operation, while the Western Climate Initiative and Midwestern
Governors Greenhouse Gas Accord have both stalled) and state level climate action plans – while
other states have actively rejected such proposals.
H2: Coase Theorem: Aggregate results will be largely unaffected by the allocation schedule
(Coase, 1960).

25

ETPs are popular in part because of their flexibility. Numerous design mechanisms allow
for particular interest groups, sectors, income brackets or regions to be relatively favored by the
policy, and thus offset harms caused prior to the policy, or as a result of the policy itself. The
Coase Theorem (1960) is relevant here, which states that the aggregate economic impacts of
externality rights allocations should be equivalent regardless of which groups within society (e.g.
sectors, households income bracket, or regions) receive them. This condition certainly holds in
the partial equilibrium analytical setting. However, recent studies have observed that the
allocation schedule can also have small but significant aggregate impacts (Rose, Wei, & Prager,
2012). Allowance allocation is a particularly powerful design mechanism because the sale of
allowances to regulated industries (as opposed to free-granting) can generate revenue for the
government to redistribute according to policy goals. For example, the additional cost of
allowances would increase production costs for industry, and hence further impact gross output.
There is hence a likely trade-off between redistributive policy goals (such as distributional equity
and investment in research and development for energy efficient technology) and economic
efficiency when comparing allowance allocation design mechanisms. The literature suggests that
auction-generated revenue recycling is less efficient than free allocation (Burtraw et al., 2009;
Oladosu & Rose, 2007; Parry & Williams, 2010).
IMPACTS BY SECTOR
H3a: GHG-intensive sectors are likely to bear the greatest regulatory burden.
H3b: Downstream and upstream sectors would absorb some of this burden.
Hypotheses 3a states the basic premise of any climate change policy impact model, while
hypothesis 3b states the purpose of using CGE analysis in this field. Table 3 provides the support
for hypothesis 3a, with the Coal-power Electricity Generation, Transmissions and Distribution

26

sector likely to bear the greatest regulatory burden of ETP.
11
Petroleum and Coal Products
Manufacturing, and Truck transportation are also likely to bear a sizable share of the regulatory
burden. However, as stated with hypothesis 3b, some of these burdens could also be absorbed by
downstream and upstream sectors. GHG intensive sectors would be able pass some of their costs
onto downstream customers, who might in turn substitute away to less GHG intensive products.
Similarly, upstream producers would experience reductions in demand for their products, and
may in turn be able to attract new custom with lower prices.
Table 3: U.S. top eight carbon emitting sectors, 2007
12

Sector
Tg CO
2

emissions
% of total U.S. CO
2

emissions*
Electric Power Generation, Transmission and Distribution (Coal) 1561.1 34.4%
Government Electric Utilities 495.3 10.9%
Petroleum Refining 414.5 9.1%
Truck Transportation 412.9 9.1%
Electric Power Generation, Transmissions and Distribution (Gas) 296.6 6.5%
Chemical Manufacturing 176.1 3.9%
Air Transportation 157.8 3.5%
Construction and Engineering 125.9 2.8%
Source: U.S Environmental Protection Agency, 2007 Emissions Inventory. *Without households, forestry and agro
sinks.
H4a: Green technology and government sectors are likely to benefit relatively more from ETP.
H4b: Sectors importing and exporting green technology and GHG-intensive substitutes are likely
to benefit relatively from ETP.
Sectors producing energy efficient technologies are likely to benefit relatively from ETP.
Energy efficient technologies would experience increased demand as a result of the policy, as
GHG intensive sectors would seek to make more efficient their own productive practices, and
downstream customers would seek out cheaper substitutes. In lieu of U.S. companies being able
to meet the increased demand for energy efficient technologies, these products could be imported

11
Recent federal level policy proposals would have limited this burden as the coal industry stood to gain a
significant number of additional allowances during the first years of the programs.
12
Note that this Table only includes carbon emissions, around 80 percent of greenhouse gas emissions.

27

from overseas, thus also benefitting those importing sectors. Over time, the increased capacity
and competitiveness of energy efficient technology industries within the U.S. could also lead to
the export of efficient technologies abroad.
13
Government sectors would also likely benefit from
ETP as an administration would be required to monitor and enforce the policy.
INCOME DISTRIBUTION IMPACTS
H5a: Coase Theorem: Allocation schedule will influence distributional impacts (Coase, 1960).
H5b: Income measures of distribution will be more regressive than consumption measures (Parry
& Williams, 2010; Hassett et al, 2009).
H5c: Relative single-factor measures of income distribution (e.g. Gini coefficient) will highlight
different concerns to measures which examination changes to single brackets only (e.g. analysis
of the relative gains to the lowest or highest bracket) (Rose, Wei & Prager, 2012).
The issue of distributional impacts has featured prominently in the ETP literature. The
Coase Theorem is again relevant here. While the aggregate effects of externality rights allocations
are expected to be equivalent regardless of who receives the rights, the distributional impacts are
clearly highly dependent on the allocation schedule. By implication, this condition provides an
opportunity for policy makers to achieve equity goals while also implementing a relatively
economically efficient emissions control program. Allowances may be distributed towards
vulnerable groups – such as lower-income households, or companies vulnerable to international
trade competition – without impacting aggregate economic efficiency. However, this money
would have to come from somewhere, and in these cases it would effectively come from those
companies which would have received the rights otherwise. Another design element available to

13
Importing energy efficient technologies would be necessary in the short run, as the U.S. has less current
capacity for energy efficient technology production than other OECD nations, especially Germany and
Japan.

28

policy makers is to auction-off emissions allowances, rather than to freely grant them. In this
scenario, revenues generated form allowance sales could be redistributed towards particular
groups.
A second key debate is the measurement of distributional impacts. Income-based
measures have been found to be more regressive than consumption-based measures, all else equal
(Parry & Williams, 2010; Hassett et al, 2009). It is expected that this study will reveal the same
results. Another key debate is the measurement used for comparing changes across the income
spectrum. Some measures, such as the Gini coefficient, aim to capture the spread of income
across brackets in a single factor. The Gini coefficient, for example, is produced from a
mathematical function and is expressed as a value between 0 and 1, with 0 being a perfectly fair
distribution of wealth across households, and 1 representing a scenario where all the wealth was
concentrated in the very wealthiest households. The advantage of these single-factor measures is
that the entire distributional impact of a policy is captured in comparing measures for before and
after the policy. However, single-factor measures are unable to capture the subtle changes to
specific income brackets. Moreover, single-factor measures are relative rather than absolute, and
hence misleading outcomes can occur. For example, in the situation where every income bracket
experienced an equal dollar increase to their income, the resulting Gini coefficient would appear
less fair even though the policy impact is perfectly fair. Closer inspection of changes to specific
income brackets can offer more useful analysis, especially when policy makers are interested in
achieving particular policy goals, such as mitigation of harms to the poorest households. Thus the
measurement approach itself will highlight different distributional concerns for policy makers.

29

IMPACTS BY STATE
H6a: States with GHG-intensive sectors are likely to bear a relatively greater regulatory burden
(Burtraw et al, 2009; Hassett et al, 2009; Pizer et al, 2010).
H6b: States with the largest density of GHG-intensive industries are likely to be those states with
the largest, likely negative, economic impacts.
States with the largest proportion of GHG intensive sectors are likely to bear the largest
regulatory burden, especially in terms of the direct regulatory impacts. At the national level, the
Electricity Generation, Transmission and Distribution sector emits over half of all U.S. industrial
Table 4: States with largest proportion of gross output from
utilities
14
, 2007 ($m)
State
Proportion of gross output
from utilities
West Virginia 8.11%
Alabama 7.67%
Mississippi 7.44%
North Dakota 7.12%
South Carolina 7.04%
Montana 7.03%
Oklahoma 6.85%
Arkansas 6.54%
Vermont 6.40%
Hawaii 6.17%
Maryland 6.09%
National Average 4.98%

carbon emissions, and hence states with this sector (especially Coal-power) contributing a large
proportion of gross output are likely to bear a larger regulatory burden. Following this logic, West
Virginia (8.11%) and Alabama (7.67%) would experience the largest direct impacts, with
Mississippi (7.44%) and North Dakota (7.12%) close behind (Tables 4 and 5).

14
As with Table 2 above, Tables 3 and 4 only refers to carbon emissions. ―Utilities‖ includes both private
and public utilities.

30

Table 5: Top 5 non-Utilities carbon emitting sectors as proportion of state gross output.
State
Sectoral Output ($m)
State Gross
Output
High Carbon %
of State Output Constr.
Petroleum
and Coal
Products
Chem.
Mfg
Air
Trans.
Truck
Trans.
High
Carbon
Total
IN 11,204 3,351 15,569 511 4,495 35,130 262,288 13.4%
LA 10,239 23,588 10,016 347 1,611 45.801 204,733 22.4%
MS 5,012 2,837 1,214 37 1,538 10,638 91,608 11.6%
MT 2,397 1,464 79 70 474 4,484 35,112 12.8%
TX 59,419 42,869 31,988 7,636 10,769 152,681 1,144,930 13.3%
WY 1,871 1,318 303 42 421 3,955 33,506 11.8%

POLITICAL IMPLICATIONS
H7a: States with relatively greater economic harms from climate policy are less likely to have
federal representatives supporting climate policy and ETP (Pizer et al., 2010; Shipan & Lowry,
2001; Zahran et al., 2008).
H7b: States with relative economic harms from climate change are more likely to have federal
representatives supporting climate policy and ETP (Goulder, 2003; Victor, 2003).
H7c: States with relatively large harms to influential GHG-intensive sectors are less likely to
have federal representatives supporting climate policy and ETP.
H7d: Direct regulatory impacts are likely to have greater influence than overall (direct plus
indirect) economic impacts on state level policy making and federal level representation of state
interests.
Preliminary analysis of the direct economic impacts of ETP (Tables 4 and 5 above)
suggests that there would be significant differences between states. Comprehensive analyses of
energy use across regions further supports this case (Pizer et al, 2010). Greater energy use, and
hence climate impacts, is negatively related to metro-level climate policy making (Zahran, 2008).
Pizer and colleagues (2010) show that there is greater per capita energy use in the South, while

31

Shipan and Lowry‘s (2001) show that Southern Democrat federal representatives are less likely to
support environmental policy than their party colleagues from other regions. After controlling for
other factors such as ideology and partisan politics, it is expected that states with relatively
greater regional harms from policy are less likely to engage in policy making, whether this is in
the form of active state EPAs or equivalent, strong climate action plans or federal representatives
supporting climate policy and ETP.
It is not expected that aggregate state economic impacts would dictate policy making.
Economic impacts are likely to influence policy making in a number of potentially contradictory
ways. If influential sectors within given states were likely to be negatively impacted by a federal
ETP, it is likely that these sectors would seek to influence the political process. Even though the
state as a whole may not be substantially negatively impacted, influential industries oppose ETP.
Direct regulatory burdens are also likely to play a more salient role in policy making at this level
than indirect costs. Although less accurate, direct impacts are more parsimonious and hence can
carry with them more persuasive weight to policy makers.
It is also expected that the states that benefit from climate policy – i.e. states that would
benefit from the reductions to the harms resulting from climate change and associated co-
pollutants – would also be more likely to engage in state-level climate policy making and support
federal level regulation. In other words, state-level policy makers and federal representatives
would be rationally motivated to enact climate change policies to the extent that the risk of
climate change could be reduced (Goulder, 2003; Victor, 2003).

32

DISSERTATION SUMMARY
Chapter Two – ―Computable General Equilibrium Model‖ – discusses the use of
computable general equilibrium modeling in environmental policy analysis, and presents a
detailed description of the USCGE model used for this analysis. Chapter Three – ―Emissions
Trading Policy Modeling‖ – describes the ETP mechanism and compares it with other policy
mechanisms used to address environmental externalities. Next this chapter describes the features
and assumptions used in the USCGE model to represent the impacts of an ETP on the U.S.
economy. Aggregate and sectoral results from the USCGE model are analyzed.
Chapter Four – ―Inter-Household Equity Modeling‖ – critically reviews the literature on
the household distribution of climate policy, with an emphasis on the modeling approaches used
and the policy scenarios examined. A detailed description of the USCGE model‘s distributional
element, which features detail for both income and consumption effects, is presented to inform
analysis of the model‘s household distributional impact results. These results are compared across
various measures of dispersion.
Chapter Five – ―Inter-Regional Equity Modeling‖ – explores the literature on the regional
distribution of climate policy costs, with reference to broader literatures on economic modeling
with multiple regions and the spatial impacts of climate change and climate policy costs. This
discussion informs analysis of the regionalized-CGE policy impact results, in this case across the
50 U.S. states and the District of Columbia.
Chapter Six – ―The Politics of Climate Policy‖ – explores the influence that the economic
impacts of climate policy can have on political decision making at the federal levels. This chapter
focuses on analysis of federal roll-call voting of climate and environmental bills between 2007
and 2011. Probit regressions examine the influence of state-level CGE modeling results, state-

33

level emissions rates, ideology, environmental policy voting on Congressional climate bill roll-
call voting. This section of the dissertation aims to contribute to bridging the gap between the
environmental economics and environmental political science fields by using economic impacts
estimations at the state level to inform analysis of climate change policy making by state
representatives. This chapter next discusses the development of climate policy since the 1980s.
Five trends are identified: 1) An increasing number of climate policies worldwide; 2) A move
towards adaptation; 3) A shift in policy tools – the emergence of emissions trading policy; 4) The
growing influence of policy analysis networks; 5) A broadening of analytic criteria – an
increasing role for economic sectors, regions, and equity. The dissertation concludes by
summarizing the findings, relating analyses to the broader literature and making policy
recommendations.

34

CHAPTER ONE:
COMPUTABLE GENERAL EQUILIBRIUM MODELING

―Among the many gadgets, instruments and artefacts in its care, London's
Science Museum holds a peculiar contraption that most resembles the work of
a deranged plumber. Yellow tubes connect together a number of tanks and
cisterns, around which coloured water can be pumped. Sluices and valves
govern the flow of liquid and makeshift meters record the water-levels.‖
The Economist, July 13
th
2006.

This passage from the The Economist magazine describes one of the first computational macro-
economic models. The ―MONIAC‖ hydraulic computer was created in 1949 by William Phillips,
an economist at the London School of Economics.
15
The model represents flows of money
throughout an economy: Income flows are subject to taxes, savings, and imports while demand is
stimulated by exports, government spending, and investment.
Economic models on today‘s powerful computers are able to capture far more subtle
interactions and complexities across the macroeconomy. These models are used by academics
and policy analysts to study numerous policy questions and fall into two broad categories.
Macroeconomic models, used largely by central banks, but increasingly by government agencies
at all levels, are the closest descendents of the Phillips hydraulic computer. These models analyze

15
Phillips is perhaps more famous within the field of economics for his work on the ―Phillips Curve,‖ a
representation of the relationship between inflation and unemployment.

35

continuous detailed indicators of regional, national and international economies. They are
particularly adept at capturing and projecting dynamic economic mechanisms, thus providing
policy makers with detailed images of the current and near-future economy.
The second category of economic models – computable general equilibrium (CGE)
models – are less concerned with the complexities of money and of the business cycle and instead
focus on the structure of production. CGE models are:
―multi-market simulation[s]…based on the simultaneous optimizing behavior of individual
consumers and firms, subject to economic account balances and resource constraints. [In other
words, they are] model[s] of the entire economy based on decisions by individual producers and
consumers in response to price signals within limits of available capital, labor, and natural
resources‖ (Rose, 2009).
This approach allows more sensitive treatment of transactions between economic sectors and
institutions (such as governments and households) and income and substitution effects as they
ripple across the economy. As such, CGE models have tended to be used in contexts such as
trade, tax, environment, and disasters, where significant policy changes (such as tariff reductions
or tax code changes) or external shocks (such as a natural disaster or technological changes) are
expected to transform production and transaction activity.
This dissertation uses CGE modeling to capture the economic impacts of an ETP, in part
because of the importance of production and transaction impacts across sectors and institutions,
especially with respect to the fuel and electricity sectors of the economy. CGE models are also
attractive because they can be tailored to answer specific policy questions. In this dissertation, the
USCGE model developed by Gbadebo Oladosu and Adam Rose (Rose et al., 2009; Oladosu and
Rose, 2007, Rose and Oladosu, 2002; Oladosu, 2000) is adapted to study the economic impacts of
ETP across households and regions of the U.S. economy.

36

This chapter presents a brief history of CGE policy modeling before discussing the range
of environmental and climate change policy questions analyzed with CGE modeling. Next the
USCGE model is presented in terms of the general model features. Details on model additions
and refinements for the purpose of modeling ETP are presented in the following chapter.
A BRIEF HISTORY OF CGE MODELING
Peter Dixon, a pioneer in CGE modeling at Monash University, Australia, argues that
three characteristics distinguish the CGE approach (2006). First, CGE models are computable, as
they represent the economy numerically. A numerical database – usually a Social Accounting
Matrix and substitution elasticity values – is used to calculate coefficients and parameters within
the model, as well as serving as year-zero data for the modeling process. Further data is usually
added to represent policy-relevant detail, while dynamic CGE models require larger data-sets
over time to enable projections.
Second, CGE models are general, i.e. they model the behavior of a comprehensive set of
interdependent economic actors. Households maximize utility while firms maximize profits and
minimize costs. CGE models highlight the role of market prices – whether for goods, services, or
factors – in the consumption and production decision making of households and firms. They can
also optimize the behavior of other institutions such as governments, unions, importers, and
exporters. Third, CGE models assume market equilibrium conditions. In other words, they
represent the influence of demand and supply decisions on commodity and factor prices.
Equations for commodity and factor prices are adjusted to ensure that aggregate demands do not
exceed total supplies.
The 19
th
-century French economist Leon Walras laid the foundation for large economic
models by emphasizing the inter-connectedness of markets across the economy. Such

37

interdependence is highlighted, for example, when a rise in the price of oil causes impacts such as
a reduction in output for customers of the oil industry, as well as increased investments in drilling
and research and development from oil companies. These immediate impacts have numerous
further impacts too. Corn farmers may benefit as substitutions towards non-gasoline cars raises
the demand and price for ethanol and corn products. Or American interest rates my decline
because foreign oil-rich countries are investing heavily in American Treasury bonds.
Walras‘ macroeconomic perspective was extended by the Russian-American Nobel
laureate economist Wassily Leontief. Leontief‘s papers in the 1930s and his 1941 book, The
Structure of the American Economy, contained the first detailed input-output (I-O) table, a matrix
depicting the flow of goods and services between US industries, households, and international
trading partners. Each cell of the matrix depicts a unique transaction, the output of one industry or
institution being the equivalent of an input for an industry or institution. Input-output analysis is
based on these transaction accounts for an economy and has been used effectively since
Leontief‘s time to examine numerous policy questions, e.g., the impact of a policy on a high-
polluting sector and how costs would be passed through the economy to transacting sectors.
There are numerous limitations of the I-O approach (Miller & Blair, 1985). For example, ratios of
transactions between sectors – essentially the production functions relationships – are assumed
fixed (i.e., linear proportional relationships). As such, substitutions away from a high-polluting
sector to a greener sector cannot be accurately modeled through the input-output approach.
Early I-O analysis was also limited by poor data and computational constraints.
Nonetheless, early modelers had a notable influence on policy making. Leontief‘s book was
translated into several languages and influenced economic planners in many countries (the
Soviets independently began utilizing I-O models in parallel to Leontief‘s original work). The

38

Norwegian economist Leif Johansen, widely considered the inventor of the CGE model, applied
his modeling for the Norway government. Economists using this approach at the Cowles
Commission contributed to Second World War rationing in the US. Such policy influence
troubled some. Leontief reported ―unconcealed alarm‖ among free-market supporters, who
argued that ―too close and too detailed an understanding of the structure of the economic machine
and its operation might encourage undesirable attempts to regulate its course‖ (Wassily Leontief
quoted in The Economist, 2006). In fact, in the early 1950s there was move to halt the
construction of Official U.S. government I-O tables, but interestingly it was leaders of some of
the largest U.S. firms that prevented this action.
Advances in I-O modeling and then CGE modeling came as a combination of advances in
data collection, computational power, and economic theory. National accounts data were first
developed in Depression-era Britain by notable economists such as Colin Clark and Richard
Stone. However, the United States published the first formal national accounts in 1947, with the
United Nations publishing the first international standard, ―A System of National Accounts and
Supporting Tables,‖ 5 years later. International standards have advanced since this time. In
contrast to the first attempts by Wassily Leonteif, which covered 17 industry sectors, the latest
US input-output tables can offer resolution of more than 500 industry sectors, which can be
disaggregated further.
Computer technology advances over the second half of the 20
th
century have been well
documented, and a similar growth in computer programming power for CGE models can be
traced. Despite important contributions to CGE modeling from the afore-mentioned Johansen
(1960), as well macro-economic mathematical programming model developed by Sandee (1960)
and Manne (1963) among others, the 1960s was a relatively fallow period for the field (Dixon,

39

2006), except for the construction of the first major regional I-O tables (Rose and Miernyk,
1989). During this period, ―general-equilibrium economists developed and refined theoretical
propositions on the existence, uniqueness, optimality and stability of solutions to general
equilibrium models‖ (Dixon, 2006). The economist Herbert Scarf was critical in applying these
theoretical propositions to computer programs, developing an algorithm that would render a
model solution within a finite number of iterations. Such groundwork contributed to the
emergence of mathematical programming software such as GAMS, GEMPACK, HERCULES,
and CASGEN (Bisschop and Meeraus, 1982; Brooke, Kendrick and Meeraus, 1988; Meeraus,
1983; Pearson, 1988; and Rutherford, 1985) through the 1970s and 1980s, which enabled
economists to develop CGE models without substantial training in computer programming or
equation-system solving.
16

CGE analysis became increasingly popular for both academics and policy makers from
the 1970s onwards. For instance, by the early 1980s, major surveys of the field had been
published (Shoven and Whalley, 1984). Alongside advances in modeling capability, CGE‘s stock
rose in part because of increased awareness of the limitations of alternative methods. Major
shocks to the world economy in the 1970s – including rising energy prices, changes to the
international monetary system, and real wage rate growth in Western economies – could not be
fully captured by econometric approaches, which are based upon historical data. The magnitude
of these shocks was unprecedented, as were the specific issues at play, and hence prior data could
not explain the current trends. In contrast, CGE analysis is particularly adept at identifying
counterfactual economic impacts – i.e. what would have happened under alternate conditions -
whether the analysis is looking backwards (i.e. analysis of historical events) or forwards (i.e.

16
Scarf‘s work was influential to CGE modeling in the US during the 1970s. Scarf‘s algorithm was largely
replaced by previously-developed techniques such as the Newton-Raphson and Euler algorithms (Dixon,
2006).

40

projections of future outcomes). For example, as Shoven and Whalley (1984) highlight, CGE had
been used extensively in tax and trade policy, which required counter-factual analysis. Moreover,
by focusing on the productive functions of the economy, CGE models are particularly adept at
capturing shocks such as dramatic changes in oil prices.
Meanwhile, other macroeconomic models that focused on Keynesian economics
principles, such as those developed by William Phillips discussed above, were being challenged
by economic theorists such as Robert Lucas who were more inclined to the neoclassical school of
thought. These economic theorists challenged the very notions underlying macroeconomic
models, especially those models used by monetary policy makers to anticipate economic changes.
This challenge was sufficiently profound that, according Christopher Sims of Princeton
University, the ―use of quantitative models as a guide to real-time policy advice was cast into
such deep disrepute that academic research on the topic nearly completely ceased‖ (The
Economist, 2006). The limitations of such Keynesian macroeconomic models could also be
overcome with the microeconomic assumptions that lay at the heart of CGE modeling. Indeed, as
macroeconomic models have been redesigned to account for the neo-classical critiques, they have
incorporated one of the ideas fundamental to CGE modeling: economy-wide changes are the
aggregation of sector- and institution-specific changes.
By the1990s, CGE modeling became established as a field of applied economics, as well
as gaining broader policy influence around the world. Surveys were published in leading journals
and detailed model descriptions were available in numerous textbooks (Dervis et al., 1982;
Shoven and Whalley, 1992; Dixon et al., 1982). CGE modeling became at once more
standardized – such as the Global Trade Analysis Project (GTAP), which combines data and a

41

network of scholars from over 50 countries – and more context-specific. CGE models have been
used to analyze questions including
effects on the following variables:
―macro, including measure of nation-wide or even global economic welfare;
industry;
regional variables;
labor market variables;
distributional variables; and
environmental variables;

of changes in

taxes, public consumption and social security payments;
tariffs and other interferences in international trade;
environmental policies;
technology;
international commodity prices and interest rates;
wage setting arrangements and union behavior; and
exploitation of mineral deposits (the Dutch disease).‖ (Dixon, 2006).
CGE modeling became particularly influential in a number of policy making areas, most
notably international trade policy and environmental policy. Numerous CGE models were
developed to examine the impacts of the 1987 Uruguay round of General Agreement on Trade
and Tariffs (now the World Trade Organization). When the North American Free-Trade
Agreement (NAFTA) debates arose in the early 1990s, these General Agreement on Trade and
Tariffs models could be easily-modified. As Devarajan and Robinson (2002) put it: ―[t]he result
was that, from the beginning and throughout the negotiations, high-quality economic analysis was
available on a timely basis to inform the debate.‖ These studies overwhelming found that impacts
to the US, Canada and Mexico would be positive, with Mexico benefitting the most (Burfisher,

42

Robinson, and Thierfelder, 2001). The weight of this combined analysis tipped the balance of
debate within the US towards passage of NAFTA (The Economist, 2006).
Of any country, CGE models have perhaps gained most influence on domestic policy in
Australia, and debates over trade policy were critical here too. As Peter Dixon asserts; ―[s]ince
the late 1970s, Australian policy makers have been calling for results from CGE models on
almost every economic issue. CGE studies are regularly debated in the media and in the
parliament.‖ Dixon argues that policy debates during the 1970s over removal of protectionist
trade barriers were heavily influenced by the ability of CGE modeling to identify winners and
losers across industry sectors and regions of the Australian economy. The ORANI model
developed by the IMPACT project was able to show that high tariffs caused a high real exchange
rate that favored import-competing industries such as textiles, clothing, footwear and motor
vehicles, and import-competing regions such as South Australia and Victoria. On the other hand,
tariff reductions would favor exporting industries such as wool, wheat, meat cattle and iron ore,
which tended to be concentrated in Queensland and Western Australia. Such analyses influenced
the policy debate such that previously protectionist Australia became a model for free trade.
Though useful and rigorous, CGE models are not without limitations. CGE models can
vary substantially in terms of fundamental equations, structure, operating system, and base data,
as well as context-specific assumptions. While many of these features have become standardized
across the literature, there is justifiable concern that replication of results is only possible when
the precise set of features used in any given study are also replicated. There is also concern that
uncertainty of particular variables might be compounded as it passes through the model. As a
consequence, the validity of results depends on researchers running sensitivity tests on all key
variables and assumptions, as well as presenting the core mechanisms and assumptions during the

43

peer review process. Most CGE models have been peer-reviewed, yet these models are highly
complex and require substantial training to use, limiting the number of individuals able to
evaluate and debate the assumptions made by any given model, as well as stoking criticisms that
CGE models are ―Black Boxes‖ (Sue Wing, 2004). Result verification could also be performed
by comparing model results ex post facto against empirically observable outcomes.
The limitations of CGE models highlight a broader tension within economic modeling
between detail and parsimony. Economic modeling aims to capture the resource-constrained,
welfare-maximizing realities of relevant societal actors, balancing the competing goals of
parsimony and detail with respect to the requirements of the research question. Detail often
prevails for climate policy because of the complex nature of the problem, yet this detail increases
variability in results and uncertainty as to policy outcomes. A recent meta-analysis of climate
policy economic impact studies shows that results can vary widely (Rose & Dormady, 2011).
However, much of the variance came from differing assumptions within the economic impact
modeling. For example, studies showing the largest negative economic impacts assume the
highest price per ton of emissions allowances or emissions tax rate. These studies likewise
assume the policies, when implemented, would require substantial reductions in emissions.
CGE MODELS IN ENVIRONMENTAL AND CLIMATE POLICY ANALYSIS
CGE models are often used in the environmental policy analysis field, not least because
they are able to identify the impact of changes in one area of the economy (e.g. regulation of the
energy sector) across multiple sectors and institutions within the economy, while accounting for
price changes and substitution effects. CGE models provide particularly useful analytical results
when the environmental problems studied impact a large number of industries or span across
large geographical areas. Issues such as acid rain (EPA, 2010), climate change (Conrad &

44

Schroder, 1993; Goulder et al., 1999), regional fishing quotas, national vehicle emissions controls
(Jorgenson & Wilcoxen, 1990), resource management issues (Bergman, 2005) – especially where
that industry plays a central role in the nation or region‘s economy – have all been analyzed in the
CGE context to great effect.
Analyses of acid rain and climate change policies have dominated this field beginning in
the 1990s. One of the earliest major CGE models – the GREEN model developed by the OECD
(Burniaux et al., 1992) – analyzed global climate policy issues. Single-country models were also
developed during this period. Hazilla and Kopp (1990) identified the cost of US environmental
regulations while Bergman (1990) estimated the economic impact to Sweden of replacing nuclear
power with other fuel sources while constraining emissions.
Bergman (2005) discusses numerous factors that have distinguished environmental policy
CGE models. Production sectors are often disaggregated to show greater resolution in high
polluting areas, especially electricity, transportation, fuels, metals, agriculture and chemicals.
Specific disaggregation schemes are dependent on the policy question – for example agriculture
being emphasized in climate policy analyses and forestry being disaggregated in acid rain
analyses (Nordhaus, 1994; Hill, 2001). The production function is generally a combination of
capital (K), labor (L), energy (E) and non-energy materials (M), or KLEM, at the higher levels of
the nesting structure (please see details in the USCGE model description below). In
environmental policy analysis, focus is placed on the energy nest, which is usually a combination
of electricity and fuels, before being disaggregated further down the nesting structure. Bergman
(2005) suggests that the elasticities of substitution between inputs are generally ―guesstimated,‖
being based along with the nesting structure on literature surveys of relevant econometric studies.

45

Bergman (2005) further suggests that emissions tend to be modeled as a proportion of
fossil fuel use. Hence emissions abatement is achieved via reductions in the use of fossil fuels
across industries, as well as inter-fuel substitutions. Technological change is the last factor
Bergman address which is salient to this study.
17
Technological change is particularly important
in long-run policies such as climate change and acid rain programs that are designed to allowed
industry the sufficient time to adjust practices. New technologies have the potential to contribute
to energy- and emissions-efficiency improvements during policy implementation, and hence are
often heralded as a solution to environmental problems. For example, carbon capture and storage
technologies have long been touted as a way for coal powered electricity generation plants to
produce the same amount of electricity with a fraction of the emissions. As such, the capacity of
technological change to either influence policy outcomes, or even be influenced by policy
implementation, is an important component of CGE modeling. According to Bergman (2005),
most studies assume that technological change is exogenously derived – that is, neither the policy
nor the working of the economy have any direct influence on the matter. As such, technological
change in relation ot enrgy use is usually modeled as ―Autonomous Energy Efficiency
Improvements‖ (AEEI) of between 0 and 2 per cent per year, on the basis of historical data. That
said, there are ongoing attempts to ―endogenize‖ technological change. For example, Goulder and
Schneider (1999) have influenced a stream of work that employs a market for research and
development services within the CGE model as a means to capture the impacts of induced
technological change.
CGE analysis is not appropriate for some areas of environmental policy analysis.
Numerous policy questions are focused on very local environmental problems, such as air and

17
Also discussed are environmental benefits and international trade. While these issues are both important
they are not modeled in detail here.

46

noise pollution in urban contexts. There are also issues may also only impact a small number of
people or industrial actors, such as CFCs. In these cases, regulated sectors could be greatly
impacted, yet the general equilibrium impacts would be very minor and hence difficult to capture
in the CGE context. The general impacts may also be overwhelmed by model assumptions or
changes to other variables in the model. CGE models are also ill-equipped to analyze policies that
affect only specific areas of given sectors. For example, there are numerous policies that target
only small and medium sized enterprises within a given sector. The production functions of these
areas may be quite distinct from the larger enterprises, and hence the assumptions and data
requirements may be difficult to obtain.
USCGE MODEL GENERAL CHARACTERISTICS
This analysis adapts the USCGE model developed by Gbadebo Oladosu and Adam Rose
(Rose et al., 2009; Oladosu and Rose, 2007, Rose and Oladosu, 2002; Oladosu, 2000) to analyze
the economic impacts of environmental policy and disasters. The resulting model consists of 65
producing sectors, along with mulitiple institutions: households (split into nine household income
groups), government (split into two groups of state and local, and federal), and external agents
(i.e. foreign producers). The model represents production activities as a series of nested constant
elasticity of substitution (CES) functions. For international trade, the model employs Armington
functions for imports and the constant elasticity of transformation function for exports.
Armington functions separate out imported and domestically produced goods, ideally to reflect
differential quality and consumer preferences. After governments collect taxes from labor and
capital income, the remaining income goes to households and foreign entities according to fixed
shares. Transfers also occur between institutions in the form of subsidies, social security
payments, and income taxes.

47

A Linear Expenditure System of aggregate commodities (such as Food, Housing, and
Gasoline) represents household consumption behavior, while a Leontief expenditure function
characterizes government consumption. Household and government borrowing and saving
functions are specified, and the consequent investments are allocated to finance capital goods.
Equilibrium conditions include the balancing of supply and demand across sectoral product
markets, while the labor market follows Keynesian assumptions to allow for an unemployment
equilibrium. Data from government and the academic literature is used to formulate key aspects
of this model: Social Accounting Matrices for national and selected states, as well as wage and
employment data.
18

In this analysis, two fundamental changes to the USCGE model enable detailed treatment
of ETP impacts on the electricity and fuel sectors. First, the Electricity Generation, Transmission,
and Distribution sector is disaggregated down to 10 sectors by energy source: Coal, Gas, Oil,
Other Fossil, Nuclear, Biomass, Geothermal, Hydro, Solar, and Wind.
19
USCGE is based on
social accounting matrix data from IMPLAN that does not disaggregate the electricity sector
beyond a single sector, hence an estimation approach developed by Marriot (2009) is
implemented using national U.S. EPA eGRID data (2012) for the electricity sector. Second, the
USCGE nesting structure is revised to allow for substitutions between the new electricity sectors,
as well as those between fuel sources.
Elasticity of substitution values play an important role in CGE models. As with each
stage of the USCGE model development, a literature review has been performed to ensure that
elasticity of substitution values are consistent with other studies (Rose, Oladosu, Lee, & Beeler
Asay, 2009). One complication here is that nesting structures can vary across CGE models also.

18
These data are acquired from IMPLAN, a national and regional economic accounts data provider
(IMPLAN, 2010) and government sources.
19
Please see Table 3, Base Case results below for the relative sizes of these sectors.

48

For example, the Phoenix model developed by Fisher-Vanden, Schu, Sue Wing, and Calvin
(2011) focuses electricity sector nesting on the substitutions between Base Load, Intermediate
Load, and Peak Load periods of electricity demand. Nonetheless, where nesting structures are
comparable, the elasticity of substitution values used in this analysis are consistent with those
used by Fisher-Vanden and colleagues.
USCGE MAIN MODEL
In line with CGE theory, producers are treated as profit maximizers. To implement this
mathematically, the duality principle is applied, which states that ―any concept defined in terms
of the properties of the production function has a ‗dual‘ definition in terms of the properties of the
cost function and vice versa‖ (Varian, 1992). As such, a cost function is minimized in the profit
equation. This allows for theoretical properties such as Shepard‘s Lemma to allow input demand
function derivation from cost functions. More practically, cost and price data required for the cost
function are readily available, while profit functions data requirements are difficult to obtain. The
use of cost data also allows for different physical measurements to be combined into a single
dollar value.
Each sector is assumed to be model by a representative producer. ―The aggregate profit
obtained by each production unit maximizing profit separately taking prices as given is the same
as that which would be obtained if they were to coordinate their decision‖ (Mas Colell, Whinson,
& Green, 1995). Aggregation issues are also relevant when considering substitution functions.
―Separability‖ is assumed, meaning that the marginal rate of substitution between any two factors
(see Figure 1 for the nesting structure) in a given group is independent from the marginal rates of
substitution elsewhere in the nesting structure.

49

Producer behavior in CGE models is usually represented by the constant elasticity of
substitution (CES) functional form. Examples of CES and CET (constant elasticity of
transformation, the corollary function using a negative elasticity of substitution, in this case to
represent shifts between domestic and foreign sales) functions are presented in equations 1) and
6) below. CES functions range from perfect to no substitution between factors. Perfect
substitution means that two factors or goods can be substituted without a change to utility, ceteris
paribus. This implies that an increase in the price of one good or factor would increase demand
for the other good or factor. Perfect substitution, also known as the Leontief or fixed input
coefficient function, is represented in Figure 2 by the straight line isoquant with an elasticity of
substitution value of 1(γ = 1). At the other extreme, no substitution is represented by the right-
angled isoquant with the negative infinity elasticity of substitution value (γ = -∞). In between is
unit elasticity of substitution, which corresponds to the Cobb-Douglass function, and is
represented in Figure 2 by a curved isoquant with an elasticity of substitution of zero (γ = 0).
The cost functions used in the USCGE model are constant returns to scale form, non-
separable, nested constant elasticity of substitution (NNCES), which is shown in equations 9 and
10 below. As shown in Figure 1, the nesting structure is divided into 9 levels. The top level
(―KELM‖) represents substitution possibilities between sub-aggregates of Capital (K), Labor (L),
Energy (E) and Materials inputs (M). Level 2 separates substitution possibilities into two groups
– an aggregation of Capital, Energy, and Labor inputs (KEL), and a material input aggregate (M).
Level 3 further separates the KEL nest into an aggregate of Capital and Energy inputs (KE) on
one side, and Labor inputs on the other (L). In addition, the materials nest is separated into three
further aggregates: 1) Services (S), which further disaggregates in Level 4 to Financial Services

50

(FS)
20
and Other Services (OS);
21
2) Manufactured Goods (M1), which disaggregates to Chemical
Materials (CM)
22
and Other Materials (OM)
23
; and 3) Transport, which disaggregates to
Transport Services (TR)
24
and Other Transport (OT).
25
Also in Level 4, Capital (K) and Energy
(E) are separated.
The Energy nest (E) is disaggregated with a high degree of resolution to highlight
impacts to GHG intensive sectors, especially Electricity Generation (ELEC) and Fuel (FUEL), in
Level 5. As shown in Figure 1, the top electricity nest in allows for substitution between Fossil
(EF) and Non-Fossil (ENF) generated electricity in Level 6. Level 7 then allows for substitution
between Coal generated electricity (EC) and Other fossil fuel generated electricity (EO) in the
fossil generated electricity area. The Other fossil fuel generated electricity sources are then
disaggregated into Gas (EO1), Oil (EO2), and Other Fossils (EO3). Substitution between Nuclear
(ENU) and Renewable generated electricity (ER) is represented in Level 6 of the non-fossil
generated electricity area (ENF). Here substitution between fuels (FUEL) is enabled first between
Coal (CL) and Non-coal (NCL), the latter of which is further divided in Level 7 between
Petroleum (PT) and Gas (GS). The combined effect of this second set of changes is to allow for

20
The Financial Services nest is an aggregate of the intermediate good inputs of Finance Banking and
Credit (BANK), Security Brokers (SECB), and Insurance (INSR).
21
The Other Services nest is an aggregate of the intermediate good inputs of Sanitary Services (SANT),
Wholesale Trade (WTRD), Retail Trade (RTRD), Real Estate (REST), Owner-Occupied Dwellings
(OODW), Hotel and Restaurants (HOTR), Personal Services (PSRV), Veterinary Services (VSRV), Waste
Management and Remediation (WAST), Other Business Services (OBSV), Entertainment (ENTR),
Education (EDUC), Medical Services (MEDC), Other Health and Social Services (OSOC), Federal
Military (FGML), Other Government (OGOV), and State and Local Government (SGGV).
22
The Chemical Materials nest is an aggregate of the intermediate good inputs from Chemicals
Manufacturing (MCHM), Private Water Utilities (PWAT), and Government Utilities (GVUT).
23
The Other Materials nest is an aggregate of the intermediate good inputs from Agriculture (ABEEF,
ADARY, AOLVS, APOUL, AFISH, AOTH), Mining (CRUD, OMIN), Construction (CNSR), Food
Manufacturing (MFML, MOML, MANM, MPTY, MFSH, MOFD), other Durable and Non-Durable
Manufacturing (MOND, MPRM, MORD, MSEM, MODR), Communications (COMC, INFO) and Non-
comparable Imports (NCMP).
24
The Transport Services nest consists of Air Transport (TAIR), Truck Transport (TRUK), Water
Transport (TWAT), and Rail Transport (TRAL).
25
The Other Transport nest consists of Other Transport (TOTH), Private Transit (TLTP), and Local Public
Transportation (TLTG).

51

inter-fuel substitution, both in terms of the direct fuel source and the electricity generating sectors
that consume a significant proportion of total fuel supplies.
Level 1
Level 2
Level 3
Level 4
Level 5
Level 6
Level 7
Level 8
Level 9
L = Labor OM = Other Materials EC = Electricity Generation Coal
K = Capital TR = Transport (Truck, Rail, Water and Air) EO = Electricity Generation Non-coal Fossil Fuels
E = Energy OT = Other Transport Services EO1 = Electricity Generation Gas
M = Materials
NCL = Non-Coal Fuel (Petrol and Natural
Gas)
EO2 = Electricity Generation Oil
M1 = Materials Sub PT = Petroleum Refining EO3 = Electricity Generation Other Fossil Fuels
T = Transport GS = Gas Utilities ENF = Electricity Generation Non-Fossil Fuels
S = Services ELEC = Electricity Generation ENU = Electricity Generation Nuclear
FS = Finance EF = Electricity Generation Fossil Fuels ER = Electricity Generation Renewables
CM = Chemicals
Figure 5: USCGE nesting structure for emissions trading policy simulations.

Figure 6: CES Production Function Isoquants
KELM
KEL M
S M1 T
FS OS CM OM TR OT
KE
ELEC
E
CL
FUEL
NCL
GS PT
K
L
ENF EF
EC EO
EO1
ENU
EO2 EO3
ER
Intermediate Goods

52

International trade is represented by the Armington relationship for imports (Equation 6)
and the constant elasticity of transformation for exports (Equation 1) below). As shown in
Equation 6), a CES function allows for demand substitution between domestic goods and
competitive imports. Equation 1 represents the corollary for substitutions between exports and
domestic markets to characterize the revenue-maximizing behavior of domestic firms. In line
with the small country assumption, import and export prices are fixed as equivalent to world
prices. The small country assumption is relaxed for individual sectors as part of the simulation
modeling, which is discussed further in the following chapter.
Labor and capital income payments from producing sectors are allocated to the nine
household income brackets (HH1-9). Labor income is subject to social security government taxes
(Equation 16), while capital income is subject to profit taxes by government and depreciation
charges and retained earnings functions by industries (Equations 18, 18.1 and 19). Labor and
capital income are distributed to households according to the Multi-Sector Income Distribution
Matrix (MSIDM), which uses exogenously sourced data to relates sectors of the economy to
income brackets in terms of both labor and capital income payments (Equations 17 and 20). The
MSIDM data are explicitly incorporated into the USCGE model labor and capital income
equations, ensuring that the sector-household income relationships are incorporated into the
production calculations (Rose, Stevens, & Davis, 1988; Rose, Wei, & Prager, 2012). Transfers
between institutions are also represented – please refer to Equation 24.
Household consumption is divided across numerous aggregate commodities (see Table
A3 in Appendix A) via a Linear Expenditure System. The cost of household services (Equations
29 and 30) incorporates household demands along with the changes to prices of composite goods
adjusted for household substitution elasticity values. These factors, along with the household
consumption data, inform expenditure shares on household services (Equations 31 and 32) and

53

household disposable income (SY
hh
, which is household labor and capital income less taxes and
transfers), combine to determine household demand and total purchases across these aggregate
commodity groups (Equations 33, 34 and 36). Finally, the household utility function (Equation
35) accounts for changes to household disposable income, commodity price changes, and
household substitution effects. The household utility function also sums across households to
represent total societal utility.
Government consumption is represented by a Leontief expenditure function. Household
and government savings are fixed proportions of disposable income (i.e. income following
adjustments for taxes and transfers) and are balanced by savings by foreign sources. Each of these
institutions also undertakes capital borrowing. Investments are financed by net institutional
savings plus depreciation charges and retained earnings (thus mirroring the deductions from
household capital income).
This counter-factual analysis first establishes a base case simulation with no policy
impacts. This study estimates the policy impact of a U.S. federal ETP from the current date until
the year 2020 (further details are provided in the following section), and hence a long-run closure
rule is implemented for both the base case and policy impact simulations, which assumes that
capital can move between sectors while labor factor use remains constant (Oladosu and Rose,
2007, Dixon and Rimmer, 2002). While this closure rule may appear to imply that employment
does not change between the present and 2020, a more accurate interpretation is that the base case
and policy simulations do not influence the underlying employment rate of the economy as, in the
long-run, flexible wages allow employees and employers to optimize labor output with respect to
market conditions (Dixon and Rimmer, 2002).

54

USCGE MODEL MATHEMATICAL EQUATIONS
Key
1. Variables – capitalized elements;
2. Parameters from pre-policy base data – capitalized elements with ―0‖ at the end;
3. Other parameters (shares, factors, etc) – lower case elements;
4. Arrays/Matrices – lower case subscript;
5. Specific cell/row or column within arrays/matrices – capital subscript.
For example, in the equation following this paragraph, representing Capital Income Formation,
the variable INC
K,s
refers to capital income (the subscript K) across all relevant institutions (s),
which in this case are the nine household income brackets (HH1-9). The parameters msidm
K,I,s

and re
K
both refer to the capital, the former being a matrix of sectors (i) and relevant institutions
(s). Both INC and FCPS are (post-policy) variables with corollary parameters INC0 and FCPS0
that represent pre-policy base data values.

CET between Exports (EXPS) and Domestic sales (DSL) for exporting sectors. Determines PRD
i
.
1)

1.1)
1.2)

55

where:
PRD
i
and PRD0
i
are output variables and pre-policy value respectively
26
across i sectors;
EXP is exports and DSL is domestic sales; sh are cost share parameters for exports EP
and domestic sales DS respectively; ρ are exogenously derived cost function exponents
for exports to the rest of the world (EOW).
Determines DSL
i
2)
where:
PD and PX are domestic and output price respectively; σ are exogenously derived cost
function exponents for exports to the rest of the world (EOW).
Determines EXPS
i
3)
where:
PE and PM are export and import prices respectively.
Determines PC
j
and PD
i

4)

26
Henceforth, the variable/pre-policy value distinction will be implied; variable, parameter, and sub- and
superscript definitions will be defined only once and implied thereon. A list of variable and parameter
definitions is provided in Table A1 below.

56

5)

where:
PC the price of domestic goods; mpr are make coefficients derived from the make matrix
(the supply matrix representing commodities produced by each industry).
CES between Imports (IMPS) and Domestic sales (DCQ) for importing sectors. Determines
SUP
i
27

6)

where:
SUP is the composite goods supply; sh are cost share parameters for imports MP and
domestic demand DD respectively, and follow the same calculation process as sh for EP
and DS above; ρ are exogenously derived cost function exponents for imports from the
rest of the world (MOW). SUP, DCQ, and DSL are all fixed at zero where the
corresponding values in any given sectors are zero. In other words, no transactions can
emerge between any given sectors/institutions pairings where they did not exist in the
base data.

27
Non-comparable imports are an exception, as imports equal supply.

57

Determines DCQ
gi

7)
where:
gi represents goods produced by sectors across the economy; PQ equals the composite
goods supply price; σ are exogenously derived cost function exponents for imports from
the rest of the world (MOW).
Determines IMPS
gi
8)
Import and Export taxes are set to zero. Elasticity values are multiplied by 0.5 and plus 0.001.
Efffac (Factor of Productivity for these purposes) are set to 1 for all sectors at the KELM level of
the nesting structure (these are changed for relevant sectors to reflect AEEI).
Determine PDMD
fi,i
and DMD
inpt,i
respectively.
9)
10)

58

where:
PDMD is the demand price; fi and inpt are composite factor inputs
28
, with fi representing
the upper level in a nest and inpt representing the lower level in a nest (e.g. KELM is fi to
the inpts of KEL and MAT); δ is the factor of productivity, set to 1 across all nests and
with respect to all sectors, except where changed at the KELM level of the nesting
structure for the purposes of modeling technology change. Demand and demand price
are fixed at zero where base data entry was zero (i.e. no new transactions between given
sectors can appear).
Determines DMD
KELM,i
11)
where:
DMD is determined for the top-level nest (KELM) only.

Determines PX(i)
12)
where:
tr is the sum of tax rates across government institutions.

28
Composite factor inputs are provided in Table X; the nesting structure in Figure 1 for provides detailed
relationships between composite factor inputs across nest levels.

59

PDMD for labor, capital, and goods equal PL, PK, and PQ respectively. PK equals the capital
return rate. Factor use of labor and capital
29
, FCU
f,i
, equal demand for labor and capital, DMD
f,i
,
and sales across goods and sectors, SAL
gi,i
, equals demand, DMD
gi,i
.
Determines net price PV
i

14)

where:
TAX is the sum of taxes collected by all government institutions.
Import and export prices are set as equal to world prices (except when small country assumption
is relaxed). Indirect taxes equal DMD(kelm,i) times PDMD(kelm,i) times the tax rate.
Emissions constraint function
15)

where:
TOTEMS is the emissions cap; emsfac and fuelfac are, respectively, industrial process and
fuel combustion emissions factors across all regulated industries (i.e. unregulated
industries are set to zero); fuel refers to commodities demanded from the Coal Mining

29
Labor and capital factors are represented as f in sub-scripts when combined.

60

(COAL), Crude Oil and Natural Gas (CRUD), Petroleum Refining (MPET), and Gas
Utilities (GASU) sectors.
Income allocation
Distributed Labor Income
16)
where:
FCPS are factor income distribution coefficients across sectors, in this case for labor
income, which are equal to PL the price of labor less the labor tax rate times the factor
use of labor across sectors.
Labor income allocation
17)

where:
INC
L,s
is labor income across s household income brackets; msidm
L,i,s
is the multi-sector
income distribution matrix for L labor income, representing shares of labor income by
sector paid to household income brackets.

61

Distributed Profit Income
18)

18.1)
where:
K refers to capital income; ENT refers to incomes paid to enterprises; TRNRC
ENT,za
are
transfers from government institutions (Federal Government Defence and Non-Defense,
and State Government) to ENT enterprises; ssup
i,ENT
refers to transactions between i
sectors and ENT enterprises from the institutional supply section of the social accounting
matrix; DEPR is capital depreciation and dpr is the depreciation rate parameter,
calculated by dividing capital payments from investments less indirect capital taxes by
the factor use of capital, all for pre-policy data.
Retained earnings, REAN
19)

where:
re is the retained earnings rate, calculated by dividing pre-policy retained capital earnings
(REAN) by capital factor use.

62

Profit income allocation
20)

where:
msidm
K,i,s
is the multi-sector income distribution matrix for K capital income,
representing shares of capital income by sector paid to household income brackets.
Federal and State government taxes on Labor income (social security) and Capital profits, TAX
f,gv
21)

Government income
22)

where:
t refers to the tax types indirect tax (tx), export tax (te) and import tax (tm); f refers to
factor inputs (labor and capital).
Government expenditure balance
23)

63

where:
INC
gv
is income across government institutions; GVSAV is government savings; and
TRNRC
za,gv
is transfers received by government institutions from government institutions
and foreign sources.
Government purchases are fixed as equal to pre-policy levels.
Transfers calculations
24)
where:
z and za are institutions engaging in transfer activity, including households (HH1-9),
government (Federal Government Defence and Non-Defense, and State Government),
enterprises (ENT). Additional detail for international transfers is provided via Rest of
World (ROW) and Stock Change (STK) functions.
Balance of payments of foreign countries
25)

26)

64

where:
TRNRC
za,ROW
are transfers to foreign sources from US households and federal
government institutions.
Household income, INC
hh

27)

28)

where:
HHBW
hh
and HHSV
hh
are household borrowings and savings across income brackets
respectively. HHBW and HHSV equal household income multiplied by the marginal
propensity to borrow and save (respectively) for each income bracket, which are derived
from pre-policy borrowing and saving as a ratio of total income. TRNRC
za,hh
are transfers
to households from households, government institutions, and foreign sources; TAX
HH,hh,gv

are household taxes across hh income brackets to gv government institutions. Household
tax equals household income multiplied by the tax rate for each government institution.
Household expenditure balance
Household savings or borrowings equal income multiplied by a marginal propensity to save and
borrow parameters across household brackets, which are derived from pre-policy saving and
borrowing as a ratio of total income.

65

Household Production Function
Unit cost of Household Services, PSRV
29)

30)

where:
hsrv are services purchased by households,
30
the parameter higshr are the shares of
household spending (for each income bracket) for each hsrv group that are spent on each
commodity (e.g. the share of the lowest income bracket‘s food spending that is spent on
fish); σ
hsrv,hh
are household substitution elasticity values; HDMD is household demand for
commodities; HDSRV is the total household expenditure on hsrv household service
groups across hh household income brackets.
Share of inputs into services, HIDEM
gi,hsrv,hh
(31 is intermediate variable used to calculate 32)
31)

32)

30
Services are grouped into Food, Housing, Gasoline, Public Transport, Other Transport, Medical,
Household Goods, Other Goods, Other Services, Water, Electricity, and Other Fuels – see Table A3 for
further details.

66

where:
The variable HGISH
gi,hsrv,hh
are shares of household spending (for each income bracket)
for each hsrv group that are spent on each commodity.
Demand for inputs into services
33)
Total input purchases
34)

Parameter HHCAL calculated from various sources including SAL0, mpet spend hh table,
disposable income.
Utility function
35)

36)

67

where:
UTILI
hh
is utility per household income bracket; SY is disposable income; PSRV is the
unit cost of household services; hhcal
SPEXPD,hsrv,hh
are HDSRV (total household
expenditure on hsrv household service groups across hh household income brackets)
adjusted for income substitution elasticity values; hhcal
MSHARE,hsrv1,hh
are the shares of
household disposable income (by income bracket) spent on each hsrv commodity group,
adjusted for income substitution elasticity values.
Objective function
37)

where:
PRODU is gross domestic product; PRD
i
is gross sectoral product; and PX
i
is the output
price for each sector.
Total savings
38)

68

where:
SKT represents stock change.
Investment demand equals investment (INVEST) times pre-policy investment parameter.
Investment demand equals investment demand times Capital consumption matrix (cac)
parameters (and summed across industries). Investment price equals quantity price times Capital
consumption matrix (cac) parameters (and summed across goods). Stock change (SKT) equals
pre-policy stock change parameters times supply (SUP).

69

CHAPTER TWO:
EMISSIONS TRADING POLICY MODELING

―Anyone who has the audacity to flip on a light switch will be forced to pay higher
energy bills thanks to this new tax increase.‖
John Boehner on the Waxman-Markey 2009 Climate Change Bill
―Now, I'll be doing something a little more rewarding, more environmentally conscious.
I'll be trying to help customers save some money; that's a good feeling. And this looks
like a field that will be growing.‖
Scott Newman, Energy Conservation Engineer, November 2009

How much do ETPs cost the economy? And who pays these costs? These are two of the
stimulating questions behind this dissertation, and more specifically this chapter. However, as the
quotes above highlight, these questions have both political and practical consequences. John
Boehner is promoting the Republican Party‘s position on ETP, which is that it amounts to a tax
on energy consumers, and would negatively impact the economy as a whole. This quote was in
fact linked to an MIT study (Paltsev, Reilly, Jacoby, Gurgel, & Metcalf, 2007), which appeared to
estimate a possible $366 billion in government revenues from the Waxman-Markey bill. Dividing
this figure by the number of US households gives a crude estimate of $3,100 cost per household,
which was widely quoted at the time, both in the press and by Republicans such as John Boehner.

70

The original authors of the study were sufficiently aggrieved by this ―misrepresentation‖
as to write to John Boehner to explain why households would instead pay only $340 on average.
The authors argue that, while the $366 billion in government revenues figure is not incorrect, the
study instead provided a range of possible revenues of between $100 billion and $500 billion per
year. However, the most important distinction is that the study redistributes government revenues
back to households as part of the program. The purpose of such a design is to ensure that lower
income households – who spend a larger proportion of incomes on electricity, fuel, and other
GHG-intensive goods – are compensated for the inequitable costs of climate policy.
31
With such
revenue redistribution, households on average are negatively impacted by $340, though these
figures fluctuate over the years of the study, 2015-2050, and according to the make-up and
location of the household
32
(Paltsev et al., 2007).
This incident is an interesting example of the ways in which economic research can be
translated and mistranslated into the political arena. This is at times political opportunism in order
to support a particular viewpoint. Yet it is also reflective of uncertainty and ambiguity within the
economic literature. Economic systems are large, complex and dynamic, and, as shown in
Chapter 2, there are multitudinous approaches used to capture the impacts of environmental
policy alone. Modeling approaches aside, most recent analyses of U.S. federal climate policy
suggest that whether the macroeconomic effects of ETP are positive or negative depends on the
policy context and modeling assumptions. For example, Montgomery and colleagues (2009) find
a total negative impact of $730 billion for the proposed Waxman-Markey congressional bill. Yet
when Paltsev and colleagues (2009) combine the Waxman-Markey bill with the federal stimulus,

31
This issue is discussed in depth in the Chapter Four.
32
Larger households that consume more energy would pay more, as would those located in areas with
higher rates of GHG consumption.

71

they estimate a total $1.39 trillion negative impact.
33
A recent meta-analysis of climate policy
economic impact studies across a broader range of contexts shows that results can vary widely
(Rose & Dormady, 2011). The meta-analysis found that much of the variance in results was
caused by differing assumptions within the economic impact modeling, as opposed to the model
characteristics or deficits. For example, studies showing the largest negative economic impacts
assume the highest price per ton of emissions allowances or emissions tax rate. These studies
likewise assume the policies, when implemented, would require substantial reductions in
emissions.
Economic models of ETP include numerous assumptions because they are capturing
complex systems. In ETPs, the government allocates each regulated entity a fixed number of
emissions allowances per time period. At the end of each time period, regulated entities must
return emissions allowances equal to their GHG emissions. If the regulated entity cannot reduce
emissions down to their cap level, it may purchase emissions allowances from regulated
industries with spare allowances.
34
ETPs therefore place a regulatory burden on polluting sectors,
especially those for which emissions reduction is above the market price for emissions
allowances. Conversely, some regulated entities may benefit if they can invest in emissions
reductions at a per-unit rate which are below market price. Numerous emissions reductions
strategies are open to regulated entities. Regulated entities could reduce output, invest in
emissions reductions technology (e.g. energy-efficient light bulbs, renewable energy power
sources, potential carbon capture and storage technologies), improve emissions efficiency of
current practices (e.g. electricity-use management systems, industrial process improvements), or
engage in land management approaches (e.g. afforestation, soil sequestration).

33
As highlighted below, the model approach is also critical, as Ross and colleagues (2008) also find a
slightly negative impact for a similar policy using an alternative CGE model, ADAGE.
34
Failure to comply with this process results in fines in excess of the emissions allowance price, thus
emissions allowance purchase would always be preferable to fine payment.

72

In line with the general equilibrium economics presented in the previous chapter, ETPs
are likely to have impacts across the economy, some positive and some negative. Regulated
industries – e.g. electricity generation – are likely to bear the greatest burden of policy
implementation, though some facilities within those industries – e.g. recently built power plants
with efficient fuel switching process – may even benefit from the changes. However,
consequential impacts would ripple across the economy. Prices for downstream customers of
regulated industries would likely increase – in our example, this would include all sectors using
electricity, as well as households and government institutions – while upstream customers would
experience reductions in demand – e.g. coal mining and other fuel sources. Hence, while John
Boehner is probably correct to say in the above quote that electricity consumers are likely to
experience price increases as a result of an ETP, the magnitude of the price increase is important.
And while some job would likely be cut as a result of an ETP, there are also likely to be increases
in employment in other sectors, as exemplified by the second quote at the beginning of this
chapter.
The ripple effects from a policy regulating the electricity generation sector could counter-
balance or amplify one another, because of the centrality of the sector within the macro-economy.
Counter-balancing could occur when lower demand for raw fuel materials would reduce prices,
which could offset some of the increase in higher electricity prices to consumers. Amplification
could occur when households paying higher electricity bills could also be impacted by higher
costs from goods produced by industries experiencing higher electricity costs themselves.
Computable general equilibrium modeling has often been used in the ETP arena, and is used in
this dissertation, precisely because of its ability to capture many of these ripple effects.

73

This chapter provides a brief overview of the development of emissions trading as an
academic solution to the problem of environmental externalities before exploring the broad range
of ETPs implemented worldwide. The following section explores the numerous issues raised
regarding ETP. This discussion also highlights the various design mechanisms that can be used in
ETP to solve these issues. This discussion of design mechanisms also informs the modeling
approach taken in this dissertation. The modeling approach, as presented in the last section of this
chapter, draws upon recent academic literature and identifies innovative modeling features in the
CGE context.
POLICY APPROACHES FOR INTERNALIZING EXTERNALITIES
ETPs are one of numerous approaches used to ―internalize externalities.‖ The concept of
economic externalities was first highlighted by the English economist Arthur Pigou in his book
The Economics of Welfare (1920). ―Incidental uncharged disservices,‖ or negative externalities,
are the harmful effects of production which the producer has no incentive to pay for. For
example, a factory built in a densely populated area is likely to increase congestion and cause
light, noise and air pollution. Each of these creates a cost for the local residents that the
industrialist has no direct charge for, in terms of inputs to production. Industrialists of course also
provide incidental uncharged social benefits, for example if they invest in new roads that others
use, or if they build housing for employees that also raises the value of other homes in the area.
While both types of externalities have received attention, negative externalities have
concerned academics and policy makers the most because of the potential for serious harm to be
done. Thus it is argued that a factory should not be built if the social costs are sufficiently large.
The question then becomes one of what is deemed ―sufficiently large‖ by the polity, which in
America today is a complex and dynamic interactive process of academic and public debate,

74

policy making, policy implementation, bureaucracy, and judicial functions. As such, academic
scholarship on the issue of negative environmental externalities is found in areas ranging from
environmental economics and environmental law through to public administration, organizational
behavior, and risk perception.
Numerous approaches have been used to address the issue of negative environmental
externalities. These approaches can focus on behavioral incentives as well as compensatory
mechanisms, and in some cases the two elements are integrated. Arthur Pigou highlighted that
taxes are used to internalize externalities and cited the example of countries taxing alcohol to pay
for policing costs. The incentive to drink alcohol is reduced as the price is higher, while society at
large is compensated for the policing costs of those who may still decide to drink. The challenge
with any such tax on negative externalities is to identify the correct level of tax to levy, how
much should be compensated, and to whom. The corollary to this approach is to incentivize
behavior with lower negative externalities. For example, subsidies could be provided to factories
to locate outside of a dense neighborhood. This would obviously create a different set of costs
and transfers between relevant actors, with locals paying for the factory to move elsewhere rather
than the factory paying locals for the right to pollute in their area, which may not be feasible
given resource constraints.
Other approaches have focused on behavioral incentives, such as command-and-control
policies, which issue closure orders or levy fines when particular compliance targets are not met.
Such fines may be placed into a pot for potential compensation; however, the aggregation of
piecemeal fines is unlikely to match total revenue drawn from a tax. Command-and-control
policies also allow ―free‖ pollution up to the emissions reducing requirement, yet under the tax
approach, no emissions are without cost. In part because of this, command-and-control policies
are potentially open to lawsuits filed by industry against bureaucrats because of perceived

75

arbitrary implementation of fines. The courts are also an avenue for compensation suits filed by
individuals harmed by industrial pollution. In addition the courts serve as a check on the political
and bureaucratic process. Bureaucracies such as the Environmental Protection Agency may file
suits against industry to enforce statutes or act in lieu of policy decision making. Individuals or
interest groups may also file suit against government agencies to ensure that they enforce statutes.
Other pollution reduction approaches do not include a punitive element and instead focus
on information and moral suasion, including best practices sharing between industrial entities and
media campaigns – such as the ―flex your power‖ public campaign to encourage more efficient
use of energy in homes. While these approaches can achieve pollution reduction, there are only
limited numbers of so-called ―win-win‖ opportunities available, whereby it is economically
efficient to reduce emissions. In addition, the voluntary nature of moral suasion and information
campaign approaches means that only a limited number of industrial entities or members of the
public are likely to be influenced.
ETPs are one of these numerous policy and legal mechanisms that internalize
externalities. The English economist Ronald Coase in his 1960 article ―The Problem of Social
Cost‖ first identified that a market for externalities could be introduced by treating pollution as a
property right that could be traded. Coase showed that pollution markets can ensure that social
welfare is maximized – i.e. pollution rights are used in the most efficient way – as long as
pollution rights can be clearly established and monitored, and transaction costs for trading
pollution rights are negligible. It did not take long for this idea to be applied to more practical
policy designs. In 1966, the American economist Tom Crocker first published a design for an air
pollution rights market – ―The Structuring of Atmospheric Pollution Control Systems‖, while in
1968 the Canadian economist John Dales first proposed a market for water pollution rights in
Pollution, Property and Prices. Baumol and Oates (1971) and Montgomery (1972) then provided

76

the mathematical proofs to support the case that a uniform pollution permit price would meet
particular environmental targets in an economically efficient manner.
ETPs also focus on behavioral incentives. Entities with cost-effective pollution
reductions opportunities are encouraged to take them, while other entities are facing the
equivalent of a tax. Unlike a tax system, the proceeds are not necessarily paid to government, and
instead can be directly allocated to pollution reduction. This last sentence is intentionally vague.
To be more precise, government does not benefit when pollution permits are freely distributed to
industry – according to an appropriate rule such as historical emissions rates. However,
government can also auction pollution permits in order to raise revenue. This approach is
essentially equivalent to a government-administered tax.
While ETP and tax approaches may differ in terms of where funds are distributed, both
are generally far more cost-effective at achieving total pollution reduction than the command-
and-control and legal-compensation approaches discussed above. Or put another way, ETP and
tax approaches can achieve a better environment at the same cost as command-and-control
approaches. However, there are important exceptions. These approaches are no longer more cost
effective when policy monitoring and enforcement costs exceed the cost savings compared to
command-and-control systems. Monitoring and enforcement costs increase as the number of
polluting entities increases, the geographically spread of entities widens, or when pollution
sources are mobile – such as vehicles. In these cases where monitoring and enforcement costs are
increased, administrators face a complex trade-off between the rate of monitoring and
enforcement and the rate of cheating among regulated entities. As such, command-and-control
approaches can be more cost-effective when large numbers of small polluters are regulated – such
as launderettes – or when vehicles are regulated. This is partly why the favored approach for

77

motor vehicles is a combination of fuel efficiency standards for new vehicles and baseline smog
checks for older vehicles (Baumol and Oates, 1975).
TRANSLATING EMISSIONS TRADING POLICIES FROM ACADEMIA TO POLITICS
As economist Tom Tietenberg (2008) highlights, three major events enabled ETP to
move from the academic setting into the political sphere. The first event, in 1976, was the result
of the failure for numerous regions to attain targets mandated in the Clean Air Act of 1963. In
order to meet clean air targets under the current regime, the EPA would have had to curtail
growth, an unpopular move among political representatives of these regions. The EPA instead
changed the rule, allowing new businesses to trade emissions rights with existing businesses. This
approach allowed for economic growth to be achieved alongside improvements in air quality.
While the size of the program allowed for economies of scale, these were overwhelmed by
inefficiencies elsewhere; the government monitored emissions reduction certification and permit
trading too closely, leading to inflated transactions costs and a reduction in economic efficiency
(Dudek and Palmisano, 1988).
The second event was the 1990 Acid Rain Program, which implemented a major sulfur
dioxide emissions trading program across the United States. Numerous previous attempts at
regulating acid rain pollutants had stalled in Congress. The emissions trading element reduced
costs sufficiently in order to make the bill politically feasible. Emissions allowances could be
obtained by private transactions, where the prices were not revealed to the market, and
government auction of new allowances. This second, innovative feature enabled a market price to
be established, and hence provided important signals to the market regarding whether and when
to invest in emissions reductions technology. Moreover, administrators were able to cheaply

78

monitor the impact of emissions reductions targets on industry – a low price signals that targets
are being met easily and vice versa for a high price.
The third major event was the 1997 Kyoto Protocol on Climate Change, an international
agreement for legally binding GHG emissions reduction targets in industrial countries (largely
Western and former-Soviet countries). In order for industrial countries to be able to meet these
targets, three kinds of international emissions trading were allowed. Alongside emissions traded
between industrial countries (―Emissions Trading‖), these countries could also receive credit for
financed emissions reduction projects in other industrial countries (―Joint Implementation‖) or
developing countries (―Clean Development Mechanism‖). This policy was important because it
was the first to address climate change. It also had important consequences with ETPs
subsequently implemented at all levels, ranging from companies such as BP to regions such as the
European Union Emissions Trading System, the largest ETP in the world.
EMISSIONS TRADING POLICIES IN PRACTICE
As shown in Table 6, ETPs have been implemented at numerous levels of governance,
from the level of the company up to the international sphere. The international oil companies BP
(Akhurst, 2003) and Shell have both implemented internal ETP for carbon emissions as part of a
broader effort to green their production practices and public image. A voluntary and primarily
US-business based scheme has also been in operation for GHG emissions through the Chicago
Climate Exchange.
Turning to government regulation, there have been numerous sub-national ETP. The
trailblazers were in the US, with the Los Angeles RECLAIM program, which regulated Nitrous
Oxides and Sulfur Oxides from 1994 onwards, and the Chicago ERMS program for volatile
organic materials (Evans, 2006) from 2000. 11 states in the North East implemented the Ozone

79

Transport Commission in 1999, which was subsequently extended out to the broader region
through the State Implementation Plan; both were part of the Nitrous Oxide Budget Program.
Like the Sulfur Dioxide trading policy detailed above, these actions were in response to the US
federal government‘s Clean Air Act Amendments of 1990.
More recently, ETP for GHGs were proposed in numerous states and regions in the US
and Australia, largely in response to these countries‘ decision to opt out of the UN Kyoto climate
change protocol (Abate, 2005-6). ETPs in New South Wales, Australia, the RGGI in North
America,
35
and the State of California are three examples of sub-national government programs
currently in operation, yet numerous more at both the state (New Mexico, Illinois, Florida,
Massachusetts, Oregon) and regional (Western Climate Initiative, Midwestern Greenhouse Gas
Regional Accord [Rose, Wei, Wennberg, & Peterson, 2009]) levels have been debated in recent
years.
Table 6: Emissions trading policies in deliberation and in operation.
Venue Spatial area Deliberation Design Operation
US Sulfur Dioxide National 1990 1995
Chicago, US City-region 1993 2000
Los Angeles, US (RECLAIM) City-region 1994
Chile National
Texas, US Sub-national state 2005
UN Kyoto Protocol International 1997 1997
BP, UK Company 1997 1997 1998
New Jersey, US Sub-national state 1998 1998
Shell, UK Company 1998 1999 2002
UK National 1998 2000 2001
Canada National 1998
Norway National 1998 2002 2005
New South Wales, Australia Sub-national state 1998 2003 2003
Denmark National 1999 1999 2000
European Union International region 1999 2003 2005
Chicago Climate Exchange (CCX) Company 2000 2001 2003
Massachusetts, US Sub-national state 2001 2001
NEG/ECP 2001 2002
NAFTA International region 2001
Japan Voluntary National 2002 2002 2005

35
The Regional Greenhouse Gas Initiative is a cooperative of 10 states in North East US; www.rggi.org.

80

US Congress National 2003 2009
RGGI, US Sub-national region 2003 2003 2009
Western Climate Initiative, US Sub-national region 2004 2007
Oregon, US Sub-national state 2004
Japan National 2005 2005 2006
New Mexico, US Sub-national state 2005
California, US Sub-national state 2005 2006 2013
Illinois, US Sub-national state 2006
Florida, US Sub-national state 2007 2008
New Zealand National 2007 2007 2008
South Korea National 2007 2008
Ontario-Quebec, Canada Sub-national region 2008
Australia National 2009
Tokyo City-region 2009 2010
Adapted from Betsill and Hoffmann (2011)
At the national level, numerous European countries operated ETP for GHGs in the first
half of the 2000s decade. These were later all incorporated into the European Union Emissions
Trading Scheme discussed above. New Zealand and Japan have also introduced ETPs (Jaffe,
Ranson, & Stavins, 2009; Betsill and Hoffman, 2011), while the Canada, Australia, and South
Korea governments have all debated them. Since then, there have been two major debates around
introducing a US federal emissions trading scheme for GHGs. The 2008 Liebermann-Warner
Senate Bill was defeated in June of that year. The 2009 Waxman-Markey Assembly Bill passed a
year later, but the Senate version – the Kerry-Boxer Bill – has since stalled.
It is notable that, in response to the lack of legislative action on climate change from
Congress, courts and environmental campaigners have stepped in. The Supreme Court had
already ruled in 2007 that the U.S. Environmental Protection Agency is required under the Clean
Air Act to regulate GHGs if they pose a risk to human health. In response the Environmental
Protection Agency implemented new fuel-economy standards for vehicles identified in 2009, and
proposed carbon emissions limits in 2012, both of which have since been further upheld in court
proceedings. The latter policy has focused on carbon pollution from new power plants, such that
new coal power plants without carbon capture and storage capabilities are not able to be built. As

81

U.S. Energy Information Agency projections suggest that no new coal powered plants were going
to become operational between 2017 and 2035 (EIA, 2012), in some respects this policy could
have less of an impact than the ETPs proposed by Congress in 2009. However, equivalent
emissions reductions could have been achieved at a lower cost elsewhere in the economy, as
would have been the case for the congressional proposals. As such, Congress‘ failure to enact
climate change legislation in 2009 has led to an inferior outcome for the U.S. economy.
EMISSIONS TRADING POLICY DESIGN PROBLEMS AND SOLUTIONS
There are numerous issues raised by the implementation of ETPs, some of which can be
addressed with specific policy design mechanisms. These issues include the potential for:
monitoring and enforcement costs to increase transaction costs and possibly outweigh
benefits,
allowance markets and auctions to be manipulated by powerful industrial actors,
emissions leakage to unregulated regions,
mismatches between regulated areas and externality impacts,
co-pollutant ―hot-spots‖ to develop,
lower income groups or racial groups to be disproportionately impacted by the policy
in terms of consumption or income and employment,
mothballing of regulated entities, leading to significant job losses in regions,
regulatory uncertainty.
ETPs can be designed to address these issues with design mechanisms such as:
which sectors must comply (coverage),
the cap level,

82

the distribution of the cap level over time,
the allowance allocation process,
whether or not allowances can be borrowed or banked,
whether allowance price floors or ceilings are adopted,
the extent to which offsets (i.e. emissions reductions from another policy or program)
can be used.
However, implementing these design options may create consequential impacts also. This section
explores each of the emissions trading issues, ways in which they may be address, and the
potential consequential impacts these may create.
MONITORING AND ENFORCEMENT COSTS
As discussed above, it is possible that monitoring and enforcement costs may increase
transaction costs such that alternative methods such as command-and-control may prove to be
more cost effective. It may be possible to reduce monitoring costs through mechanisms such as
self-reporting of emissions levels or randomized monitoring checks. One of the benefits of ETPs
is the ability to reveal prices through auctions, and arguably the same logic can be applied to self-
reporting in the monitoring of emissions levels. Primarily, emissions rates are relatively easy to
monitor because industries also regularly report inputs and outputs from which emissions levels
can be estimated. When combined with random spot checks, the incentive to under-report would
be low. Industrial entities may have some incentive to cheat the system in a single time period
ETP. However, over numerous periods the incentive to cheat is reduced because successive
under-reporting of emissions levels would be difficult and hence would simply mean costs being
displaced over time.

83

MARKET POWER IN ALLOWANCE MARKETS AND AUCTIONS
Market power is potentially a problem in situations where, because of their dominance in
a market, one or many industrial actors can leverage financial profit from the allowance market or
auctions through price manipulation. The extent to which such market power can occur is largely
dependent on the market share of any given company, or the ability of multiple companies to gain
a significant market share and form a monopsony through collusion. Such actions undermine the
assumptions of perfect competition within a market place and can lead to sub-optimal outcomes
in terms of economic efficiency, as well as an inequitable distribution of funds towards the
dominant firm or firms. This has been shown to be possible in computer and mathematical
simulations as well as experimental game studies of ETPs (Dormady, 2012). The same problems
can afflict both emissions allowance trading markets and allowance auctions.
Market dominance is possible in ETPs in particular when only utility companies are
regulated because this industry tends to be populated by a limited number of firms. One solution
therefore is to design the market in such a way that single or groups of firms are not able to
dominate market share. In the case of GHG regulation, for example, including a broader number
of sectors in the policy can help to reduce the market share of utility firms. Increasing the
geographical spread of a policy can also help to dilute the market share utility firms may currently
leverage on a particular locality. There is also the possibly that market dominance in the
emissions market could spill over into markets for electricity or energy more broadly.
A major contributor to the market dominance problem is asymmetric information
between industrial actors and administrators. This violates another assumption of well-
functioning markets: perfect information. When emissions trades are private and there is little
price information provided to other regulated entities and administrators, market dominance

84

actions are both hidden and incentivized. Therefore another possible solution to the problem of
market power is ensure that allowance trades or auction data are both reported and publicly
documented. While this facility in itself may not combat monopolistic or collusive behavior, it
would enable administrators to identify and punish such behavior if appropriate. Up-to-date and
publicly available trade information also improves decision making among industrial actors such
that market efficiency is improved. Such publicly available information has been touted as a
positive feature of the Acid Rain Program‘s sulfur emissions trading system, despite private
transactions also being enabled in that system. However, the RGGI has been criticized for having
paucity of publicly available information for its auction system (Dormady, 2012).
EMISSIONS LEAKAGE TO UNREGULATED REGIONS AND SECTORS
Emissions leakage can occur when regulated pollutants are not confined to the regions
and sectors covered by the regulation. Unilateral regulations in any given region or country could
cause polluting industries to move to regions with less regulations. Purchasers may also substitute
towards relatively cheaper goods and services from lower-regulation regions. This problem is not
limited to ETP, and is especially concerning in the field of climate change because sources and
impacts are both global. However, because ETP has become a popular approach for responding to
climate change worldwide, the question has often been studied in this context (Winchester,
Caron, and Rausch, 2012). The degree of leakage appears to be largely influenced by the ease of
substitution between electricity and energy sources between regions. For example, California‘s
Global Warming Solutions Act (AB32) attempts to minimize emissions leakage to neighboring
states, with which California freely trades electricity, by placing constraints on the extent to
which Californian power plants can enable leakage to electricity generators with mutual owners
within the Western Interconnection.

85

California is an interesting case because, in addition to significant inter-state electricity
trading, the region is also closely tied to Pacific Rim trading countries, the majority of which have
limited environmental regulations. As such, there is the potential for Californian regulations to
create emissions leakages to countries around the Pacific. This may result in higher overall GHG
emissions, because such countries also have lower limits on other industrial pollutants, or are
operating with less fuel-efficient technologies in industry and transportation. In addition, simply
transporting goods from Pacific countries to California may also increase net emissions.
Within a regulated region, there is also the danger that economic sectors will substitute
towards goods and services from unregulated yet polluting sectors. For example, should
electricity prices rise sufficiently, other industries and households may seek to produce energy
―in-house‖ with unregulated – and possibly dirtier – machinery.
36
This is potentially a problem
where ETP only covers larger industrial entities – because increased monitoring and enforcement
costs for smaller entities can undermine cost-efficiencies. However, the net effect here is likely to
be small because of the importance of the electricity generation sector in GHG emissions and the
economies of scale in that industry, which mean that prices would have to rise significantly
before widespread substitutions towards smaller in-house electricity production operation would
occur.
In the case of GHG emissions, regulation would ideally be applied globally to avoid the
problem of emissions leakage. The 1997 Kyoto protocol attempted to address this issue by
seeking international cooperation on GHG emissions reductions target. However, the agreement
has been limited by two inter-related factors. First, some countries such as the U.S. did not ratify

36
An example from a different context highlights the ability of businesses to substitute away from large
scale electricity, should the conditions dictate. The state-run electricity grid in India is notoriously
unreliable and experiences frequent black-outs. In response many businesses have invested in diesel
electricity generators, despite the per unit energy cost being three times higher.

86

the agreement, while others such as Canada have since pulled out. This is arguably due to
conservative politicians in both countries being reluctant to adhere to international agreements in
general, and skeptical of the climate change problem in particular.
37
However, it is also due to the
second factor, the Kyoto protocol‘s lack of regulation for developing countries. This led
politicians in some Western countries to raise concerns about whether leakage towards countries
such as China and India would occur. Subsequent international climate policy negotiations have
stalled because Western countries wanted for China and India to sign up to emissions targets
while these latter countries resisted.
MISMATCHES BETWEEN REGULATED AREAS AND EXTERNALITY IMPACTS
This issue has important similarities with the issue of emissions leakage, however the
mismatch is between the regulated region and the impacted area. The level of spatial externality
refers to the size of geographical area that is harmed by each air pollutant. For example, by
leading to worldwide climate change, GHGs are said to have global externalities, while the acid
rain pollutants – Nitrous Oxides, Sulfur Oxides, and volatile organic compounds – harm human
and natural life across watershed regions, and particulate matter impacts human health in
localized areas. The externalities that ETPs aim to reduce may extend far beyond the space
covered by the policy. For example, GHGs would ideally be matched by a global emissions
trading scheme, assuming transaction costs are held constant
38
. However, the current ad-hoc
collection of sub-national regional, national, and inter-governmental operations described above

37
Canada pulled out of Kyoto in 2011 after the conservative party gained power, while a switch of political
office in the opposite direction in Australia lead to that country ratifying Kyoto for the first time in 2007.
38
There are two opposing forces when it comes to transaction costs. On the one hand, a larger scheme‘s
economies of scale would reduce the marginal transaction costs. However, on the other hand, the greater
complexity of trading between the established laws of different countries would increase marginal
transaction costs.

87

does not match the level of spatial externality, creating perverse incentives for other countries to
benefit from free-riding (Weiner, 2006-7).
The geographical coverage of climate policy regulation – as highlighted in the discussion
of the Kyoto protocol above – is a major issue in terms of international equity. For example,
developing countries have consistently argued during international climate change conferences
that historical responsibility for industrial GHG production, and hence emissions mitigation, lies
with the Western world.
39
Moreover, the future costs of adaptation are more likely to be
experienced by vulnerable developing countries (for example, in Africa – Collier et al, 2008).
This argument depends on a particular notion of equity, a historical version of what Adam Rose
and colleagues (1998) refer to as ―sovereignty‖ equity, which states that ―[a]ll nations have an
equal right to pollute and to be protected from pollution‖ (1998: 30).
This notion of international ―sovereignty‖ equity can both conflict and overlap with
individual equity – or what Rose and colleagues refer to as ―egalitarian‖ equity, which states that
―all people have an equal right to pollute or to be protected from pollution‖ (1998: 30). The
vulnerability of individuals to pollution is likely to be correlated to some extent to the
vulnerability of nations. For example, in terms of climate change, the poorest and most vulnerable
nations are likely to house the poorest and most vulnerable individuals. However, sole reliance on
international variation disguises the dramatic gap between rich and poor within countries.
Hurricane Katrina highlighted the vulnerability to weather of the poor within the world‘s richest
nation. The ability to adapt to climate change will depend on financial resources, regardless of the
country. Hence, the current policy spatial mismatch of ETPs might be equitable in terms of
international ―sovereignty,‖ yet be inequitable in terms of individual ―egalitarian‖ equity.

39
See Zhang (2008) for a more detailed discussion of China‘s reluctance to engage in a global emissions
trading scheme.

88

ETP design features could help alleviate these equity concerns, while still ultimately
achieving an outcome whereby emissions leakage incentives are reduced along with total
emissions. Particular countries could have less-stringent caps placed on them, both in terms of the
current level and how this cap changes over time. For example, a given country could be allowed
to increase emissions for a number of years before having the cap gradually tighten and reduce
over time. This would be equivalent to granting countries additional allowances. While it might
be ideal for developing countries to ―free-ride‖ in a world where only developed countries are
regulated, it is also possible for developing countries to benefit from a global ETP when
compared with a world without any regulation, even when the harms of climate change are not
accounted for. This is because developing countries may have a comparative advantage in
emissions reductions markets, such that emissions reductions technology investments from
developed countries would create more growth than would have occurred otherwise. For
example, this is easy to imagine with respect to forest offsets or renewable energy production
where land factor inputs are plentiful in developing countries.
Currently numerous ETPs have been enacted worldwide, and there is the question of
whether linking these policies would improve or undermine cost efficiencies (Jaffe, Ranson, &
Stavins, 2009). Works in the literature also consider links to other policies in general (Fullerton,
2002), and links to financial markets (Sandor, Walsh, & Marques, 2002). Offsets come in
numerous forms, including the planting of trees as so-called ―carbon sinks,‖ and incentivized
emissions reductions in developing countries, and are beneficial because they provide relatively
low cost sources of emissions reductions. However, offsets accounting a serious issue. Questions
have been raised as to whether emissions reductions would have been undertaken without the
incentives from offsets, or to what extent forestry should be seen as a carbon-capturing, given that
carbon would be released back into the atmosphere should the tree fall down or be burned. Under

89

the Kyoto Protocol, the Clean Development Mechanism and Joint Implementation programs offer
the broadest offset accounting structures, and aim to provide avenues for Western countries to
offset emissions reductions. For example, the European Union Emissions Trading Scheme allows
offsets through these programs, while Japan has undertaken the majority of its emissions
reductions by purchasing cheap offsets through these same programs.
ANCILLARY POLLUTANTS
Emission policy spatial mismatch can also occur through indirectly regulated air
pollutants, which are often called ancillary pollutants. Ancillary pollutants are commonplace as
most industrial processes emit a range of pollutants concomitantly. Yet the proportional mix of
pollutants can vary substantially between industries, so that regulation on one pollutant across all
industries would lead to disproportionate emissions of ancillary pollutants. This is of particular
concern when the GHGsare the regulated pollutant and the ancillary pollutants are creating
locally and regionally harmful externalities.
ETPs can exacerbate these problems because they allow for individual entities to increase
pollution as long as the overall level of pollution is reduced (through the trading of pollution
rights permits). As described above, those entities with the highest marginal abatement costs are
likely to be able continue polluting at current or growing levels, creating so-called ―hotspots‖ for
ancillary pollutants (Drury et al, 1998-1999; Chinn, 1999). Thus, while a global emissions trading
scheme would be optimal if the sole criterion is the reduction of GHG emissions, this may not be
the case if ancillary pollutants are brought into the criteria mix.
Until recently, empirical evidence of this phenomenon was limited, and findings were
mixed. Drury and colleagues (1998-1999) cite ―preliminary analysis‖ from environmental justice
scholar Manuel Pastor that the major polluters within the Los Angeles RECLAIM program, and

90

those purchasing the greatest number of pollution allowances, were located in poor and ethnic
minority areas, creating disproportionate health risks in those communities. However, this claim
does not appear to have been published and environmental law scholar Byron Swift finds no
evidence of disproportionate harms resulting from the ETPs that emerged from the 1990 Clean
Air Act Amendments. Moreover, recent work by Evan Ringquist (2011) at Indiana University
found that the Sulfur Dioxide ETP has not concentrated these emissions in black or Hispanic
communities, or poorly educated communities. Instead, black and Hispanic communities have
lower than average rates of sulfur dioxide. As such, while there have been no analysis of the
impact of ancillary pollutants, the fact that there is no evidence of ETPs increasing the regulated
pollutant suggests that the impact for ancillary pollutants will be the same.
ECONOMIC EQUITY IMPACTS
There is also the potential problem of policy-cost mismatch, which is manifest by those
in lower income brackets and people of color paying disproportionately for the costs of the
policy, through lost earnings, higher consumer prices, or varying returns on investment. This
problem becomes a spatial issue if disproportionately impacted industries and social groups are
clustered in a given area. There is a growing literature on the ―incidence‖ of ETP across income
brackets, in other words whether rich or poor sections of society are likely to pay
disproportionately the costs of the policy. An emissions trading scheme would be viewed as
―progressive‖ if it favored the poor, and ―regressive‖ if it favored the rich. The findings of papers
are mixed and highly dependent on the modeling assumptions made regarding policy mechanisms
and economic context.
Early studies of this phenomenon relating to environmental regulation in general used
more basic input-output models, and tended to find that policies increased costs which were then

91

largely passed on to the household consumer. As lower income household spend a greater
proportion of their income on the affected goods, it was suggested that they were bearing
inequitable levels of the costs. However, this is not the full picture. Wealthier households with a
greater level of income via company stocks would also be affected. Moreover, there are feedback
effects that input-output analysis does not address. More recent CGE analyses have identified a
more nuanced image of costs distribution, which in particular has enabled comparison of different
ETP options. For example, auctioned credits tend to be less regressive than ―grandfathered‖
credits as the government is able to redistribute the revenue from the former to the ―losers,‖
though this latter practice results in less efficient outcomes (Parry & Williams, 2010). Moreover,
CGE modeling allows the comparison of the impacts across regions (Burtraw et al 2009, Hassett
et al 2009).
However, there remain gaps within this literature. For example, there have been only
limited attempts to assess the impact of different policy mechanisms on equity outcomes. Burtraw
and colleagues (2009) provide an in-depth discussion of the potential trade-offs here, but do not
back it up with a coherent analytical framework. Paul, Burtraw, and Palmer (2008) overcome the
methodological deficiencies of the previous paper by running an economic simulation model of
allocation types specifically, and find that auctions are superior to free allocation in terms of their
ability to compensate the losers of cap and trade policies, especially lower level electricity
consumers. However, these impacts depend on whether the energy markets in question are
regulated or fully competitive. Future research could combine these approaches to determine a
fuller understanding of compensatory regimes that was also methodologically rigorous.
In a similar vein, there is little discussion of how policy mechanisms might influence
outcomes in terms of broader analytical criteria. Income equity is often considered across
households, yet rarely discussed in terms of race. The Coase theorem is often cited as a means to

92

square the three competing criteria of environment, equity, and economy, as it states that the
allocation of pollution rights will lead to the same (efficient) aggregate economic outcomes
regardless of the distribution of those rights. This is certainly true on the whole, yet further
analysis of how specific policy mechanisms influence these criteria is limited.
Another mechanism for setting differential cap levels is to distribute allowances unevenly
across sectors, and this could have equity consequences with respect to both inter-household and
inter-regional equity.
40
However, a more lively area of debate is the mode through which
allowances are distributed. ―Grandfathering‖, also known as ―contingent allocation‖ (Bushnell &
Chen, 2012) distributes allowances for free according to the historical emissions rates of the
regulated entities. For example, if the cap level for an industry is determined to be 10 percent
below current levels, then entities within that industry would be allocated allowances equivalent
to 90 percent of their current emissions at no charge. In contrast, auctions distribute allowances
by charging entities at a price identified during the bidding process.
41
The bidding process is
governed by the same economic principles as the market for permit trading; under competition,
polluting entities are likely to bid up to a price equal to the per-unit cost of investment in emission
reductions.
Industry is likely to favor the grandfather allocation approach because, in contrast, the
auction process transfers resources from industries to government. On the other hand there are
clear benefits to government and those reliant on government services if auctions can generate
substantial additional revenue, what has been called the ―double dividend‖ (Goulder, 1995). The

40
The issue of the equity impacts of ETP are the theme of Chapter 5 (inter-household equity) and Chapter 6
(inter-regional equity), and hence are presented in greater detail there.
41
There are numerous styles of auctions that each result in unique levels of industry transfers to
government.

93

question of where such revenue should be allocated is one with both inter-household and inter-
regional equity impacts.
UNCERTAINTY
Another major issue to consider is that of uncertainty. Reflecting the uncertainty over the
costs of climate change, the costs of ETP face uncertainty over emissions levels (and hence
allocation levels), compliance rates, and political ―credible commitment.‖ Stranlund, Costello &
Chavez (2005) also explore the relationship between permit banking and enforcement, yet focus
on the related issue of violation penalties. Their simpler model assumes inter-period permit
banking and focuses on the position in the process at which penalties should be administered for
non-compliance. In line with others, they find that a high penalty for emissions violations can
have a negative effect on enforcement. However, they propose high penalties for false reporting
of emissions levels in order to achieve high compliance and cost-effective regulation.
Murphy and Stranlund (2006) the direct and indirect effects of enforcement on ETPs.
They take a more general approach to this issue by comparing the direct impacts of enforcement
on the number of violations with the countervailing effects that occur through the subsequent
increase in the market price. They find that increased enforcement in general does decrease the
number of violations. And while this is greater than the effect of the market permit price change,
this secondary effect does reduce the overall efficacy of enforcement. These findings have
interesting applications for transaction costs and the relationship to principle-agent relationships.
The paper highlights the trade-offs between different principles-agent relationships in terms of the
transaction costs of compliance. The findings suggest that regulators must be clear on their
compliance targets and the level of efficiency they desire before determining the principal-agent
incentive structures.

94

The issue of monitoring has been discussed above in the literature (Konishi, 2005;
Mrozek & Keeler, 2004). Chen and Liu (2005) take their analysis in a different direction. They
compare the two regulator positions of a commitment to auditing and no commitment to auditing.
They find that the lack of a commitment to auditing in fact results in increased auditing by the
regulator. Hence this is a less efficient position to adopt. This highlights again the requirement of
credible commitment on the part of the regulator when establishing the institutional framework.
MODELING EMISSIONS TRADING POLICY IN THE USCGE FRAMEWORK
This paper uses the U.S. federal context because of the importance a potential U.S.
climate policy could have for global climate change policy. The U.S. currently contributes a
significant proportion of worldwide GHG emissions; 16 percent in 2011according to the
European Commission‘s Joint Research Centre (Reuters, 2012). There have been numerous failed
attempts to enact U.S. federal level GHG reduction policies in recent years,
42
the most recent of
which – in 2009 – featured ETPs as their basis.
A summary of model features is presented in Table 7. This paper compares two ETP
designs: 1) a broad-based policy design that includes the majority of sectors contributing to
national emissions output; and 2) a narrow-based policy design that features only the electricity
generating sectors. The broad-based policy design reflects broadly the sectors included in the
Kerry-Lieberman Senate Bill of 2009, with the exception of construction and other mining, both
of which have relatively small emissions levels.
43
A similarly broad based policy was used in the
first phases of the European Union Emissions Trading Scheme and in California‘s Global

42
The politics of U.S. climate policy are discussed further in Chapter 6.
43
Although the Kerry-Lieberman Bill has a notable emissions allowance allocation program, these initial
estimations are instead based on the scenario of allowances being freely granted to regulated entities
according to their historical emissions for the year 2005. The free granting of emissions allowances to
regulated industries also implies that no revenue is generated from the program.

95

Warming Solutions Act emissions trading program. The narrow-based policy design reflects the
approach used in the RGGI across 10 U.S. and Canadian states.
42

Table 7: Summary of USCGE emissions trading model features
Model
element
Description of notable features
Computable
General
Equilibrium
(USCGE)
model
Developed by Gbadebo Oladosu and Adam Rose for environmental policy analysis
(Oladosu and Rose, 2007; Rose and Oladosu, 2002; Oladosu 2000) and for terrorism
analysis (Rose, Oladosu, Lee, & Beeler Asay, 2009). Tailored here for ETP modeling with
following features:
1. Electricity sector disaggregated from 2 sectors to 10. New electricity sector estimated
with a combination of U.S. EPA eGRID data (2012) and U.S. Census data (2012).
2. Revised electricity and fuel sector nesting structure presented in Figure 1. Recent
literature reviewed for elasticity values; sensitivity tests performed.
3. Long-run closure rule fixes unemployment rate and rate of return for capital and allows
capital factor use and labor wage rate to vary.
ETP model
element
Two policy designs compared: Broad-based regulation of most emitting sectors (based on
Kerry-Lieberman senate bill of 2009) and narrow-based regulation of electricity generating
sectors (similar to RGGI program in U.S. Northeastern states).
Carbon emissions per sector calculated with U.S. EPA emissions inventory data (2009) and
USCGE model fuel input data. Other GHG emissions will be included forthwith.
Emissions constraint placed on regulated sector emissions. Trading is the difference
between pre-determined emissions levels per sector and model results. Trading net profits
and losses are redistributed to relevant sectors via changes to indirect tax.
Zero profit condition relaxed for single sector (Medical Services – MEDC) to offset
emissions constraint and balance number of equations.
Equity impact
model
element
Multi-Sector Income Distribution Matrix (Rose, Wei and Prager, 2012) calculated using a
U.S. Economic Census (2011), U.S. IRS (2011), and U.S. BLS (2011) data; serves as basis
for labor and capital income allocation from sectors to households.
Household consumption effects determined by combination of factors including incomes,
savings rates, commodity prices, and household preferences.
Distributional impacts assessed in terms of relative and absolute measures, reflecting multi-
dimensional equity considerations.

Within each policy design three emissions cap level scenarios are presented. These cap
levels reflect required emissions reductions for regulated sectors of 200, 300 and 400 Tg CO
2
and
correspond to approximately 6.7, 9.7, and 12.7 percent of total industrial emissions. As the
USCGE model is comparative-static as opposed to dynamic, this simulation reflects annualized
emissions reductions across each year of the program. This stands in contrast to the differential
emissions reductions schedules usually applied in policies, whereby low reduction targets are

96

encouraged initially before gradually increasing reduction requirements to desired levels, thus
allowing time for capital and labor factor use to move between sectors and for new technologies
to proliferate across the market.
Other dynamic factors are also neglected by this comparative-static approach. For
example, in an ETP, cheaper emissions reductions investments are likely to be targeted first. This
would provide a relative advantage to early-adopters, or sectors regulated in the earlier phases of
the program. On the other hand, there are likely to be technology improvements which tend to
decrease the relative price of emissions reductions over time, hence providing relative higher
costs for early adopters and sectors regulated in the earlier phases of the program. These opposing
forces are also subject to the forces of inflation and discounting. This analysis assumes that these
dynamic factors balance one another out in general. Instead, this analysis is focused on the
differential between counterfactual emissions (from the U.S. EPA emissions inventory) and the
emissions caps. This analysis also does not currently address other dynamic features of ETP, such
as banking and borrowing of emissions allowances; emissions reductions are assessed in
aggregate rather than with respect to time.
The emissions constraint equation consists of two elements. The industrial process
emissions element of the constraint incorporates an industrial process emissions coefficient for
each sector which is equal to emissions divided by base case output levels. Post-policy industrial
process emissions equal post-policy output multiplied by the industrial process emissions
coefficient for each sector (see the ―TOTEMS‖ Equation below). The fuel combustion element of
the constraint uses this same logic, but replaces output with the sector demands for inputs from
the four fuel sectors, i.e. Coal Mining, Oil and Gas Extraction, Petroleum Refining, and Gas
Utilities. Fuel combustion coefficients are calculated by dividing fuel-combustion emissions data
from the U.S. EPA by weighted demand for fuel input from each sector. The weightings here are

97

based on three factors: 1) Carbon emissions factors for different fuel types; 2) Sector categories –
i.e. electricity, transport, industrial, and commercial
44
; and 3) Adjustments to account for sectors
– i.e. the four fuel sectors and the chemical manufacture sector – using fuel inputs in their
production process.

Where:
TOTEMS is the emissions cap; emsfac and fuelfac are, respectively, industrial process and fuel
combustion emissions factors across all regulated industries (i.e. unregulated industries are set to
zero); fuel refers to commodities demanded from the Coal Mining (COAL), Crude Oil and
Natural Gas (CRUD), Petroleum Refining (MPET), and Gas Utilities (GASU) sectors. Emissions
allowance trading is the difference between pre-determined emissions levels per sector – in this
case an equal emissions reduction across each regulated sector – and model results. The net
profits and losses per sector from these implied emissions allowance trades are then redistributed
to relevant sectors via changes to indirect tax.
Table 8 presents total (industrial process and fuel combustion) emissions for each sector
regulated in the broad-based policy simulation; narrow-based policy sectors (i.e. fossil fuel
generated electricity sectors) are highlighted with bold font. Electricity generation sectors
contribute to over half of U.S. industrial (as opposed to residential) emissions, with Coal-
generated electricity accounting for 43.2 per cent of emissions alone. Petroleum refining is the
largest manufacturing sector CO
2
emitter, with 9.4 per cent of industrial emissions – because this
sector powers industrial processes directly with fuel combustion (MECS, 2006).

44
This ensures that EPA emissions inventory data control totals for these sector categories are met.

98

In CGE modeling, the number of equations and variables should balance to ensure that
the completely specified model is square. When the emissions constraint is added to the model,
an equation elsewhere should be removed to balance the model. To this purpose, this paper
relaxes the Zero Profit Assumption for the Medical Services sector (MEDC).
Conservation practices are often referred to as ―no regrets‖ options, whereby reductions
in emissions can be achieved at no cost, or even with net cost savings. Cost-beneficial options are
by definition economically efficient, and therefore the presence of conservation options appears
to be contrary the classical economic assumptions upon which CGE models are based. However,
there are numerous reasons, usually referred to as market failures,‖ why conservation options
may be cost-beneficial and not currently adopted into practice. Conservation options may have
long pay-back periods, and hence lead to decision makers with short-term investment horizons
choosing other options. Moreover, uncertainties surrounding future market trends, technology
development, and policy implementation highlight the potential for caution with respect to
investing in no regrets strategies on the part of relevant decision makers. There are also
behavioral factors relevant to emissions reductions that may present a more complex set of
decision making criteria – such as people enjoying less efficient lighting or overpowered vehicles,
or traditional business and institutional practices – and hence decrease the likelihood of change
towards more economically efficient outcomes. It is therefore possible for conservation practices
to be implemented with the effect of some companies benefitting from the presence of
government intervention.
Conservation practices are modeled in this study solely by assuming improved
electricity-use efficiency in the manufacturing sector when emissions are capped. The rationale
here is that only those sectors regulated under the ETP – of which the manufacturing sectors are a
significant group – would be incentivized to engage in electricity demand management

99

practices.
45
Moreover, Bloomberg New Energy Finance marginal abatement cost curve data for
the US (BNEF, 2012) highlights the importance of electricity demand management as a relatively

Table 8: Emissions and output for sectors regulated in broad-
based policy (narrow-based policy sectors in bold), 2007
46

Sectors
Emissions
(Tg CO
2
)
Proportion
of industrial
emissions
Output
($Bn)
CO
2

emissions
per $000's
output
COAL 9.53 0.21% 31.42 0.30
CRUD 75.62 1.66% 260.61 0.29
MFML 0.30 0.01% 34.70 0.01
MOML 0.57 0.01% 52.69 0.01
MANM 1.30 0.03% 114.07 0.01
MPTY 0.71 0.02% 54.59 0.01
MFSH 0.27 0.01% 11.84 0.02
MOFD 9.06 0.20% 356.43 0.03
MCHM 100.24 2.20% 962.21 0.10
MPET 424.93 9.35% 595.22 0.71
MOND 96.74 2.13% 827.68 0.12
MPRM 116.52 2.56% 249.71 0.47
MORD 0.04 0.00% 2.91 0.01
MSEM 2.09 0.05% 670.43 0.00
MODR 118.56 2.61% 2421.33 0.05
ELCL 1961.83 43.16% 155.38 12.63
ELGS 372.70 8.20% 79.98 4.66
ELOL 55.14 1.21% 8.43 6.54
ELOF 9.19 0.20% 0.88 10.48
GASU 34.87 0.77% 97.28 0.36
Source: U.S Environmental Protection Agency (2007)
Emissions Inventory.
low-cost strategy for emissions abatement. Offsets are included in the modeling approach by
allocating emissions reductions potentials to the Other Agriculture (AOTH) sector.
47
There are

45
Non-regulated sectors electricity demand is influenced primarily by electricity price changes within the
model.
46
Note that Table 2 only includes carbon emissions, which account for about 72% of U.S. greenhouse gas
emissions.
47
It is important to note that the Agriculture Other (AOTH) sector includes Forestry as well as Agricultural
Services. The current resolution of the Agriculture Other (AOTH) sector means that the offsets are being
attributed to both functions, as opposed to the Forestry sector only, which would be ideal.

100

substantial opportunities for emissions reductions in forestry, cheap reductions that would not be
as desirable without government intervention.
Technological change is modeled as exogenous and reflects only Autonomous Energy
Efficiency Improvements (AEEI). Technological change would ideally be represented as an
endogenous element of the model, following the logic that policy changes would likely spur new
technology implementation above and beyond the business-as-usual condition. Because of the
complexity of this process, including investments in research and development, technological
adoption by first users, and learning and diffusion across the market, those attempting to model
technological change endogenously have tended to create new sectors and substitution
possibilities to represent one step of this process – such as research and development (Goulder
and Schneider, 1999). However, the majority of CGE models represent technological change
exogenously, and AEEI is a common approach in the environmental policy analysis area. In this
study, AEEI is implemented as a factor of productivity in the constant elasticity of substitution
nests for the electricity sector. Reflecting empirical studies, AEEI is commonly estimated to
range between 0 and 2 percent per year, and the baseline for this study is 1 percent, with a 0
percent change simulation run as a sensitivity test.
GENERAL EQUILIBRIUM IMPACTS OF EMISSIONS TRADING POLICY
The aggregate economic impacts are small and negative for most simulations (Table 9),
corresponding to the majority of the literature, and hence providing some validation for the model
results. In Table 9 the aggregate economic impacts range from a 0.0% to negative 4.8% impact.
These results align with similar analyses in the literature; Montgomery and colleagues (2009) find
a total negative impact of $730 billion for the proposed Waxman-Markey congressional bill,
which equates to an approximate 3% negative impact to the U.S. economy. Paltsev and

101

colleagues (2009) combine the Waxman-Market bill with the federal stimulus, and estimate a
total $1.39 trillion negative impact, which equates to an approximate negative impact 5.6%.
48

More broadly, a meta-analysis of climate policy impact studies across numerous contexts (Rose
and Dormady, 2011) found that economic impacts ranged from positive 3% to negative 5%.
Table 9: Comparison of broad-based and narrow-based policy results across emissions cap levels
(Case 1). Percent changes to output and emissions by sector.

Broad-based policy

Narrow-based policy
Base Case
Level

Emissions reduction (%)

Emissions reduction (%)
6.7 9.7 12.7 6.7 9.7 12.7
Sector
s

PR EM

PR EM PR EM PR EM

PR EM PR EM PR EM
ABEEF

37 9

0.8 1.2 -0.5 -0.6 -1.7 -2.5

1.0 1.5 -0.4 -0.5 -1.8 -2.5
ADARY

39 9

0.7 1.1 -1.2 -1.8 -3.2 -4.6

1.0 1.5 -1.0 -1.5 -3.2 -4.6
AOLVS

24 4

0.9 1.9 -0.4 -0.9 -1.7 -3.6

1.0 2.2 -0.4 -0.8 -1.8 -3.8
APOUL

29 2

0.7 2.0 -0.9 1.1 -2.6 0.1

1.0 2.1 -0.8 1.2 -2.6 0.1
AFISH

4 1

2.6 4.6 3.4 6.5 4.1 8.3

2.5 4.3 3.3 6.4 4.2 8.4
AOTH

192 25

0.8 1.6 0.0 1.2 -0.8 0.8

0.6 0.0 -0.4 0.0 -1.3 0.1
COAL

30 10

-2.3 -3.2 -3.4 -4.7 -4.5 -6.1

-2.1 -3.0 -3.3 -4.6 -4.5 -6.2
CRUD

261 77

-0.5 -0.5 -1.8 -1.8 -3.1 -3.2

-0.2 -0.2 -1.6 -1.7 -3.0 -3.1
OMIN

147 18

-6.9 -10.2 -6.8 -10.1 -6.8 -10.0

-6.9 -10.1 -6.8 -10.0 -6.8 -10.0
CNSR

1,474 106

-6.6 -13.5 -6.8 -14.0 -7.0 -14.5

-6.6 -13.4 -6.7 -13.9 -7.0 -14.5
MFML

35 0

0.6 2.8 -1.3 -3.6 -3.2 -9.6

0.9 3.9 -1.1 -2.9 -3.2 -9.5
MOML

52 1

0.8 2.2 -1.2 -1.3 -3.3 -4.8

1.1 2.8 -1.0 -0.9 -3.3 -4.8
MANM

113 1

0.8 2.7 -0.5 -0.6 -1.8 -3.8

1.0 3.2 -0.4 -0.3 -1.8 -3.8
MPTY

55 1

0.8 6.7 -0.2 -0.1 -1.4 -6.6

1.0 7.9 -0.1 0.6 -1.3 -6.5
MFSH

12 0

0.7 6.0 0.3 2.7 -0.3 -0.9

0.8 6.5 0.3 3.1 -0.3 -0.8
MOFD

350 9

0.9 2.1 -0.7 -0.5 -2.3 -3.1

1.1 2.5 -0.6 -0.3 -2.3 -3.1
MCHM

968 102

0.9 0.9 0.5 0.5 0.1 0.0

0.9 1.0 0.6 0.5 0.1 0.0
MPET

598 434

-0.7 -0.8 -2.6 -2.7 -4.4 -4.5

-0.4 -0.4 -2.4 -2.4 -4.3 -4.4
MOND

826 98

0.3 0.5 -0.7 -1.1 -1.7 -2.6

0.5 0.8 -0.6 -0.9 -1.6 -2.6
MPRM

253 118

1.6 1.3 3.9 3.7 6.1 6.1

1.2 1.0 3.7 3.6 6.3 6.3
MORD

3 0

0.8 -3.6 1.4 -3.4 2.1 -3.1

0.7 -3.7 1.4 -3.4 2.1 -3.1
MSEM

673 2

-1.0 -1.7 0.7 -0.8 2.5 0.1

-1.3 -1.8 0.5 -0.9 2.5 0.1
MODR

2,439 120

-0.4 -1.2 0.5 -0.3 1.5 0.6

-0.6 -1.4 0.4 -0.4 1.5 0.6
TAIR

118 164

0.3 0.5 0.7 0.6 1.2 0.9

0.2 0.5 0.7 0.7 1.2 1.0
TRUK

256 419

-0.1 -0.3 -2.0 -3.9 -3.8 -7.3

0.2 0.3 -1.8 -3.5 -3.8 -7.2
TWAT

36 51

0.3 0.3 -0.5 -1.6 -1.0 -3.1

0.4 0.6 -0.4 -1.4 -1.0 -3.1
TRAL

54 51

-0.2 -0.5 -1.0 -2.4 -1.8 -4.1

-0.1 -0.2 -0.9 -2.1 -1.8 -4.0
TOTH

286 45

-0.1 -0.2 -1.1 -3.1 -2.0 -5.8

0.1 0.3 -0.9 -2.8 -2.0 -5.8
TLTP

32 9

0.0 0.0 -1.7 -4.0 -3.4 -7.8

0.3 0.6 -1.5 -3.6 -3.3 -7.8
COMC

678 2

0.1 0.2 -3.0 -4.2 -6.0 -8.4

0.6 0.9 -2.7 -3.7 -5.9 -8.4
INFO

508 3

-1.0 -1.7 -2.9 -5.2 -4.7 -8.5

-0.7 -1.1 -2.7 -4.8 -4.7 -8.5
ELCL

139 1,877

-5.4 -5.4 -9.6 -9.5 -13.8 -13.6

-4.8 -4.8 -9.3 -9.2 -13.9 -13.7
ELGS

82 378

-3.7 -3.8 -7.2 -7.3 -10.6 -10.9

-3.1 -3.2 -6.8 -7.0 -10.5 -10.7
ELOL

4 41

-5.5 -5.5 -11.5 -11.4 -16.8 -16.7

-4.5 -4.5 -10.6 -10.6 -16.3 -16.2
ELOF

1 9

-5.9 -6.0 -9.9 -9.9 -14.0 -13.9

-5.3 -5.4 -9.7 -9.6 -14.0 -13.9
ELNU

85 0

-2.1 0.0 -4.5 0.0 -7.0 0.0

-1.7 0.0 -4.2 0.0 -6.8 0.0

48
As highlighted below, the model approach is also critical, as Ross and colleagues (2008) also find a
slightly negative impact for a similar policy using an alternative CGE model, ADAGE.

102

ELBM

6 0

-0.6 0.0 -5.1 0.0 -9.8 0.0

0.2 0.0 -4.5 0.0 -9.3 0.0
ELGT

1 0

-0.4 0.0 -5.1 0.0 -10.1 0.0

0.4 0.0 -4.6 0.0 -9.6 0.0
ELHY

7 0

-0.1 0.0 -2.8 0.0 8.7 0.0

0.4 0.0 -2.5 0.0 -5.5 0.0
ELSL

0 0

-0.4 0.0 -5.1 0.0 -10.1 0.0

0.4 0.0 -4.6 0.0 -9.6 0.0
ELWD

3 0

-0.6 0.0 -5.1 0.0 -9.8 0.0

0.1 0.0 -4.5 0.0 -9.3 0.0
GASU

98 35

0.1 0.1 -1.3 -1.3 -2.6 -2.7

0.3 0.3 -1.1 -1.2 -2.6 -2.7
PWAT

7 0

0.0 -0.6 -0.6 -1.5 0.1 -0.8

0.1 -0.4 -0.5 -1.4 -1.3 -2.4
SANT

7 2

0.0 -0.2 -1.0 -1.6 -2.3 -3.4

0.1 0.0 -0.9 -1.5 -2.1 -3.0
WTRD

1,170 27

-0.3 -0.8 -2.2 -6.2 -4.1 -11.3

0.0 0.1 -2.0 -5.7 -4.0 -11.2
RTRD

1,304 38

0.1 0.2 -4.0 -11.6 -8.0 -22.4

0.7 2.2 -3.5 -10.4 -7.9 -22.2
REST

1,072 10

-0.5 -0.7 -2.3 -2.6 -4.2 -4.6

-0.2 -0.4 -2.1 -2.4 -4.2 -4.5
BANK

809 3

0.3 0.7 -3.0 -4.5 -6.2 -9.5

0.8 1.5 -2.6 -4.0 -6.2 -9.4
SECB

375 1

0.1 0.8 -0.5 -2.9 -1.2 -6.4

0.3 1.4 -0.5 -2.6 -1.2 -6.3
INSR

600 0

0.3 1.0 -3.0 -6.6 -6.3 -13.7

0.9 2.3 -2.7 -5.8 -6.3 -13.6
OODW

1,011 0

0.2 0.4 -3.0 -2.5 -6.0 -5.2

0.8 0.9 -2.7 -2.2 -6.0 -5.1
HOTR

671 11

0.3 0.5 -1.3 -2.2 -3.1 -5.2

0.5 1.0 -1.1 -1.9 -3.0 -5.2
PSRV

153 5

0.9 1.8 -3.4 -6.1 -7.9 -13.6

1.6 3.1 -3.0 -5.3 -7.9 -13.6
VSRV

30 0

1.0 2.4 -2.9 -6.8 -7.0 -15.6

1.6 4.0 -2.5 -5.9 -7.0 -15.5
WAST

68 47

0.0 0.0 -0.8 -1.2 -1.6 -2.3

0.1 0.1 -0.7 -1.0 -1.6 -2.3
OBSV

3,224 48

-0.3 -1.0 -1.4 -4.0 -2.5 -7.0

-0.2 -0.5 -1.3 -3.6 -2.5 -6.9
ENTR

204 2

0.6 1.4 -2.8 -6.0 -6.4 -13.3

1.1 2.7 -2.5 -5.3 -6.4 -13.2
EDUC

750 1

0.1 0.6 -0.4 -4.5 -0.9 -9.3

0.2 1.4 -0.4 -3.9 -0.9 -9.2
MEDC

1,333 11

* * * * * *

* * * * * *
OSOC

185 14

0.5 1.5 -2.7 -8.0 -6.3 -17.2

1.0 3.1 -2.4 -7.0 -6.2 -17.1
TLTG

26 9

0.2 0.5 0.3 0.5 0.4 0.5

0.2 0.6 0.3 0.6 0.4 0.6
GVUT

44 0

0.1 0.0 -0.6 0.0 -1.6 0.0

0.2 0.0 -0.5 0.0 -1.4 0.0
FGML

172 0

0.0 0.0 0.0 0.0 0.0 0.0

0.0 0.0 0.0 0.0 0.0 0.0
OGOV

311 43

0.1 0.4 -0.3 -3.5 -0.7 -7.2

0.2 1.0 -0.3 -3.0 -0.7 -7.1
SGGV

460 0

0.0 0.0 0.0 0.0 0.0 0.0

0.0 0.0 0.0 0.0 0.0 0.0
Total

24,994 4,504

-0.6 -5.3 -3.0 -9.7 -4.8 -13.8

0.0 -4.8 -2.8 -9.5 -4.7 -13.8
ELEC

327 2,305

-3.9 -5.1 -7.4 -9.2 -10.7 -13.2

-5.2 -6.7 -7.1 -8.9 -10.9 -13.3
EF

226 2,305

-4.8 -9.5 -8.7 -13.5 -12.7 -17.5

-6.3 -11.1 -8.4 -13.2 -12.7 -17.6
ENF

102 0

-1.8 0.0 -4.4 0.0 -6.2 0.0

-2.8 0.0 -4.1 0.0 -6.9 0.0
Kerry

7,093 3,312

-0.2 -6.7 -0.3 -9.7 -0.4 -12.7

-0.3 -7.9 -0.3 -9.5 -0.4 -12.8
PR = Output; EM = Carbon emissions; ELEC = Electricity generation sectors; EF = Electricity Generation Fossil Fuel sectors; ENF =
Electricity Generation Non-Fossil Fuel sectors; KL = Sectors included in the Kerry-Lieberman (Broad-based) simulation.
†
The Other
Agriculture sector (AOTH) incorporates offsets within the CGE analysis, though these are not reported here. Please see the breakdown
of emissions by sector groupings in Table 10 for offsets results. *Zero profit condition constraints on the Medical Sector (MEDC)
were relaxed to facilitate model processing; hence credence cannot be given to results for this sector.

Results at the sector level also provide important scope for model validation with respect
to economic theory and literature results. It is expected that regulated sectors would bear the
greatest regulatory burden from an ETP. As shown in Table 9, this is the case especially for the
fossil-fuel generated electricity sectors, which have the largest emissions per output levels. As
shown in the ―EF‖ row towards the foot of Table 9, the fossil-fuel generated electricity sectors are
negatively impacted by between 4.8% and 12.7% for the most stringent emissions cap levels in
both of the policy designs. Of the fossil-fuel generated electricity sectors, the oil-generated

103

electricity sector is particularly impacted across all of the ETP scenarios. While it is unexpected
that this sector is proportionally more impacted in terms of both output and emissions than the
coal-generated electricity sector, in level terms the coal-generating sector is more greatly
impacted on both accounts, which suggests a scale effect.
Reductions to fossil-generated electricity output are reflected in reductions in fuel sectors
output, especially Coal mining (COAL), which sees reductions in output of between 2.1% and
4.5%. For most cases, these negative impacts are greater than the other fuel producing sectors.
The Crude Oil (CRUD) and Petroleum Refining (MPET) sectors are negatively impacted by
0.5%-3.1% and 0.4%-4.4% respectively, and the Natural Gas (GASU) sector is impacted between
positive 0.3% and negative 2.6%. These results reflect the differential emissions intensities of
each fuel type and demonstrate the capacity for inter-fuel substitutions within the emissions
trading model. Fuel sector results follow intuition more closely that the electricity sub-sectors
because of the direct link between combustion emissions of regulated sectors and fuel demands in
the emissions constraint function.
The results in Table 9 also provide evidence that the Coase theorem does not hold in the
general equilibrium setting. The Coase theorem states that economic efficiency is the same
regardless of which actors receive emissions rights, i.e., individual entity mitigation levels are
invariant. However, there are contrasting results for the broad-based and narrow-based ETP – for
which equivalent total emissions reduction are required, but emissions rights are distributed
differently. This highlights the limitations of applying the Coase theorem in the general
equilibrium setting, where small changes in one area of the model can influence changes across
the economy. The question then becomes: Which emissions rights allocation schedule is the more
economically efficient? As shown in Table 9, the broad-based ETP design renders a less onerous
impact on the economy than the narrow-based policy design, with the exception of the least

104

stringent emissions cap in Table 9. This result highlights the fact that the least stringent cap is
non-binding to the emissions cap level. Aside from that exception, the results follow the intuition
that more economically efficient emissions reductions opportunities arise when emissions trading
is possible across a greater number of economic sectors.
The general equilibrium effects also have important implications for emissions
reductions, as shown in Table 10. While the required reductions are equivalent across both policy
designs, total economy-wide emissions reductions are greater in the narrow-based scenario for the
least stringent cap simulation (6.7% required reductions in emissions). However, when the
emissions cap are binding in both broad-based and narrow-based policy simulations – 9.7% and
12.7% required reductions – the emissions reductions are broadly equivalent. It is also notable
that offsets are a large proportion of total emissions reductions, which reflects the relatively low
cost of offsets. The numbers of offsets is similar across each simulation because offsets are
limited to just over 100 Tg of CO
2
emissions under the policy assumptions, following similar
constraints in the 2009 Waxman-Markey and Kerry-Lieberman Congress Bills.
Table 10: Comparison of broad-based and narrow-based policy results across emissions cap
levels. Emission reductions levels for groups of regulated and non-regulated sectors.

Broad-based policy Narrow-based policy

Emissions reduction (%)

Emissions reduction (%)

6.7 9.7 12.7

6.7 9.7 12.7
Sectors

EM share EM share EM share

EM share EM share EM share
Broad-based
regulated*

-120.9 50.7 -221.7 60.3 -322.5 65.0

-104.9 48.4 -214.2 60.1 -323.8 65.3
Electricity -118.7 49.8 -211.2 57.4 -304.0 61.3 -104.3 48.1 -204.9 57.4 -305.9 61.7
Coal -101.4 42.6 -177.9 48.4 -254.8 51.4 -89.7 41.3 -173.3 48.6 -257.4 51.9
Gas -14.5 6.1 -27.8 7.6 -41.2 8.3 -12.3 5.7 -26.4 7.4 -40.7 8.2
Oil -2.2 0.9 -4.6 1.3 -6.8 1.4 -1.8 0.8 -4.3 1.2 -6.6 1.3
Other Fossil -0.5 0.2 -0.9 0.2 -1.2 0.3 -0.5 0.2 -0.9 0.2 -1.2 0.3
Non-electric -2.2 0.9 -10.5 2.8 -18.5 3.7 -0.6 0.3 -9.3 2.6 -17.9 3.6
Offsets (Agro) -101.2 42.5 -100.3 27.3 -99.5 20.1 -100.9 46.5 -100.0 28.0 -99.0 20.0
Non-regulated
sectors

-16.3 6.8 -45.9 12.5 -74.0 14.9

-11.1 5.1 -42.5 11.9 -73.2 14.8
Total
reduction

-238.3 100 -367.9 100 -495.6 100

-287.6 100 -358.0 100 -498.2 100
*Regulated sectors in broad-based policy simulation are compared across both simulations for sake of consistency.

105

Table 11: Implied net emissions trading and allowance prices ($/ton) across policy caps.

Broad-based policy Narrow-based policy

Emissions reduction (%) Emissions reduction (%)
Sector Base 6.7 9.7 12.7 6.7 9.7 12.7
AOTH 0.00 -100.8 -99.8 -98.9 -100.4 -99.9 -98.9
COAL 9.67 0.3 0.5 0.6 0.0 0.0 0.0
CRUD 76.75 4.8 6.1 7.4 0.0 0.0 0.0
MFML 0.30 0.0 0.0 0.0 0.0 0.0 0.0
MOML 0.58 0.1 0.0 0.0 0.0 0.0 0.0
MANM 1.31 0.1 0.1 0.1 0.0 0.0 0.0
MPTY 0.72 0.1 0.1 0.0 0.0 0.0 0.0
MFSH 0.27 0.0 0.0 0.0 0.0 0.0 0.0
MOFD 9.09 0.8 0.8 0.9 0.0 0.0 0.0
MCHM 101.71 7.8 10.4 13.0 0.0 0.0 0.0
MPET 433.72 26.0 30.9 36.0 0.0 0.0 0.0
MOND 98.13 7.1 8.6 10.0 0.0 0.0 0.0
MPRM 117.93 9.5 15.9 22.4 0.0 0.0 0.0
MORD 0.04 0.0 0.0 0.0 0.0 0.0 0.0
MSEM 2.11 0.1 0.2 0.3 0.0 0.0 0.0
MODR 119.56 6.5 11.2 15.9 0.0 0.0 0.0
ELCL 1876.77 23.5 3.2 -17.3 76.5 75.3 72.8
ELGS 378.50 10.8 9.1 7.3 22.3 23.3 25.4
ELOL 40.64 0.4 -0.6 -1.6 1.3 1.0 0.4
ELOF 8.94 0.1 0.0 -0.1 0.3 0.3 0.3
ELNU 0.00 0.0 0.0 0.0 0.0 0.0 0.0
ELBM 0.00 0.0 0.0 0.0 0.0 0.0 0.0
ELGT 0.00 0.0 0.0 0.0 0.0 0.0 0.0
ELHY 0.00 0.0 0.0 0.0 0.0 0.0 0.0
ELSL 0.00 0.0 0.0 0.0 0.0 0.0 0.0
ELWD 0.00 0.0 0.0 0.0 0.0 0.0 0.0
GASU 35.28 2.4 3.0 3.5 0.0 3.8 4.2
Total 0.0 0.0 0.0 0.0 0.0 0.0
Emissions Allowance
Price
12.0 14.0 16.5 6.9 11.5 20.5

Table 11 presents the implied net emissions trading and allowance prices across each
simulation. These values are the difference between regulated sectors each reducing emissions at
an equivalent level – i.e. emissions reductions targets without trading; for example 6.7 percent for
the least stringent broad-based emissions cap – and the actual emissions reductions rendered by
the simulation. As highlighted in the previous paragraph, the offsets provided by forestry growth
within the Other Agriculture (AOTH) sector are the predominant net seller of emissions
allowances. This is largely because an emissions reduction policy without trading would
implicitly not include offsets. The notable counter-intuitive result here is that the Coal-powered
electricity generation sector (ELCL) becomes a net seller of emissions allowances as the

106

emissions cap reaches the most stringent cap level. Because coal is the most carbon-intensive
fuel, it is expected that coal-powered generation would be relatively impacted by an emissions
reduction policy, and would therefore also feature relatively more expensive emissions reductions
possibilities. However, it is possible that the marginal abatement cost curve of different sectors
cross. For example, abatement costs for the Coal-powered electricity generation sector (ELCL)
could be relatively high when reducing fewer emissions, yet become relatively cheaper when
larger numbers of emissions are being abated because of scale economies.
TESTING ASSUMPTIONS
A recent meta-analysis of climate policy economic impact studies across a broader range
of contexts shows that results can vary widely (Rose & Dormady, 2011). The meta-analysis found
that much of the variance in results was caused by differing assumptions within the economic
impact modeling, as opposed to the model characteristics or deficits. For example, studies
showing the largest negative economic impacts assume the highest price per ton of emissions
allowances or emissions tax rate. These studies likewise assume the policies, when implemented,
would require substantial reductions in emissions. This study confirms the importance of model
assumptions and policy design in influencing aggregate economic results.
Alongside the comparison of broad-based and narrow-based policies presented above,
sensitivity tests are also performed for the impact of the policy-induced manufacturing electricity-
use efficiency improvements. Case 1 is the above simulation, whereby manufacturing electricity-
use efficiency improvements are assumed to be 5% per year. In Case 2, manufacturing electricity-
use efficiency improvements are assumed to be 3.5% per year. In Case 3, it is assume that there
are no improvements in manufacturing electricity-use efficiency.

107

Comparing the results for Cases 2 and 3 (Table 12) and Case 1 is instructive. It is first
notable that including the manfacturing electricity-use improvements assumptions softens the
negative impact to the economy as a whole. For example, for the most stringent cap level, the
aggregate negative impact is 4.8% for Case 1, compared with 5.6% for Case 2 and 6.5% for Case
3. Also in line with intuition, output for the electricity generating sectors is less negative when the
manufacturing electricity-use improvements are included. This follows the logic that, as
manufacturing is using electricity more efficiently, electricity demand is reduced, thus reducing
output in the electricity generating sectors. For the most stringent cap level, the Case 1 impacts to
output are negative 11%, and negative 10.8% and 8.7% respectively for Cases 2 and 3. The non-
fossil generating sectors are particularly impacted by the introduction of this assumption.
Whereas fossil-generated electricity is impacted to a very similar extent (only 2 percentage points
cover each of Case 1, 2, and 3), non-fossil generated electricity is negative 6.2% in Case 1, but
only negative 3.6% in Case 3. The combination of the two main results in this paragraph
(aggregate output and electricity sector output) implies that non-electricity sectors are impacted
relatively favorably when the manufacturing electricity-use improvement assumption is
introduced. In fact, the results are less dramatic across the board with these assumptions, with
some sectors such as Fisheries (AFISH) more positively impacted, and most sectors – such as the
remainder Agricultural sectors (ABEEF, ADARY, AOLVS, APOUL, AOTH) – more negatively
impacted when the manufacturing electricity-use improvement assumption is included.

108

Table 12: Comparison of electricity conservation case results across broad-based policy
emissions cap levels (Cases 2 and 3). Percent changes to output and emissions by sector.

Case 2

Case 3
Base Case
Level

Emissions reduction (%)

Emissions reduction (%)

6.7 9.7
12.7
6.7 9.7 12.7
Sector
s

PR EM

PR EM PR EM PR EM

PR EM PR EM PR EM
ABEEF

37 9

-1.2 -1.7 -1.3 -1.8 -2.7 -3.8

-0.8 -1.1 -2.1 -3.0 -3.6 -5.1
ADARY

39 9

-2.2 -3.2 -2.3 -3.3 -4.4 -6.4

-1.4 -2.1 -3.5 -5.0 -5.7 -8.1
AOLVS

24 4

-1.2 -2.5 -1.3 -2.6 -2.6 -5.5

-0.7 -1.5 -2.0 -4.3 -3.4 -7.1
APOUL

29 2

-1.9 0.2 -2.0 0.2 -3.8 -0.9

-1.3 0.2 -3.1 -0.8 -4.9 -1.9
AFISH

4 1

3.7 7.3 3.8 7.4 4.5 9.3

3.4 6.4 4.1 8.3 4.9 10.2
AOTH

192 25

-1.0 0.1 -1.0 0.1 -2.0 -0.5

-0.7 0.1 -1.6 -0.4 -2.6 -1.0
COAL

30 10

-2.8 -3.8 -2.8 -3.9 -3.8 -5.3

-1.2 -1.7 -2.3 -3.1 -3.3 -4.4
CRUD

261 77

-2.4 -2.4 -2.4 -2.5 -3.8 -3.8

-1.8 -1.9 -3.2 -3.2 -4.5 -4.6
OMIN

147 18

-6.8 -10.0 -6.8 -10.0 -6.8 -10.0

-7.0 -10.2 -6.9 -10.2 -6.9 -10.1
CNSR

1,474 106

-6.8 -14.2 -6.8 -14.2 -7.1 -14.8

-6.8 -14.1 -7.0 -14.6 -7.3 -15.2
MFML

35 0

-2.2 -6.9 -2.3 -7.2 -4.4 -13.3

-1.5 -4.9 -3.5 -10.9 -5.6 -16.9
MOML

52 1

-2.3 -3.3 -2.4 -3.5 -4.6 -7.3

-1.5 -2.4 -3.6 -6.0 -5.9 -9.8
MANM

113 1

-1.2 -2.6 -1.3 -2.8 -2.6 -6.1

-0.8 -1.9 -2.1 -5.1 -3.5 -8.4
MPTY

55 1

-1.0 -4.9 -1.0 -5.3 -2.2 -11.8

-0.8 -4.4 -1.9 -10.7 -3.2 -16.9
MFSH

12 0

-0.2 -0.5 -0.2 -0.7 -0.8 -4.5

-0.2 -1.0 -0.7 -4.5 -1.4 -8.4
MOFD

350 9

-1.6 -2.1 -1.7 -2.3 -3.3 -5.0

-1.0 -1.5 -2.7 -4.1 -4.4 -6.9
MCHM

968 102

0.0 0.0 0.0 -0.1 -0.5 -0.6

-0.2 -0.2 -0.7 -0.7 -1.1 -1.3
MPET

598 434

-3.5 -3.5 -3.6 -3.6 -5.4 -5.5

-2.8 -2.9 -4.7 -4.7 -6.5 -6.6
MOND

826 98

-1.3 -2.0 -1.3 -2.1 -2.3 -3.7

-1.0 -1.7 -2.0 -3.3 -3.0 -4.9
MPRM

253 118

4.3 4.4 4.5 4.5 6.9 7.1

2.4 2.4 4.8 5.0 7.3 7.6
MORD

3 0

1.6 -1.5 1.6 -1.4 2.3 -1.1

1.1 0.6 1.7 0.8 2.4 1.1
MSEM

673 2

1.5 0.0 1.6 0.1 3.5 1.1

0.5 -0.3 2.4 0.7 4.4 1.8
MODR

2,439 120

0.8 0.2 0.9 0.3 1.9 1.2

0.2 -0.1 1.2 0.8 2.2 1.7
TAIR

118 164

0.9 0.7 1.0 0.8 1.5 1.1

0.7 0.6 1.2 0.9 1.8 1.3
TRUK

256 419

-2.9 -5.5 -3.0 -5.7 -4.8 -9.2

-2.1 -4.1 -4.0 -7.7 -5.9 -11.1
TWAT

36 51

-0.7 -2.3 -0.8 -2.4 -1.3 -3.9

-0.5 -1.6 -1.0 -3.2 -1.5 -4.6
TRAL

54 51

-1.4 -3.1 -1.4 -3.2 -2.2 -5.0

-1.0 -2.3 -1.8 -4.1 -2.6 -5.9
TOTH

286 45

-1.5 -4.3 -1.6 -4.5 -2.6 -7.3

-1.1 -3.1 -2.1 -6.0 -3.2 -8.8
TLTP

32 9

-2.4 -5.8 -2.5 -6.0 -4.3 -10.0

-1.7 -4.1 -3.5 -8.1 -5.3 -12.1
COMC

678 2

-4.4 -6.2 -4.5 -6.4 -7.7 -10.8

-3.0 -4.4 -6.2 -8.8 -9.4 -13.2
INFO

508 3

-3.7 -6.8 -3.8 -6.9 -5.8 -10.4

-2.9 -5.4 -4.9 -8.8 -6.9 -12.3
ELCL

139 1,877

-9.2 -9.0 -9.4 -9.2 -13.6 -13.3

-4.9 -4.8 -9.1 -8.9 -13.3 -13.0
ELGS

82 378

-7.1 -7.3 -7.3 -7.4 -10.7 -10.9

-4.0 -4.1 -7.4 -7.6 -10.9 -11.1
ELOL

4 41

-11.6 -11.5 -11.9 -11.8 -17.3 -17.2

-7.4 -7.3 -13.0 -12.9 -18.4 -18.2
ELOF

1 9

-9.2 -9.1 -9.4 -9.3 -13.4 -13.2

-4.6 -4.5 -8.6 -8.4 -12.6 -12.4
ELNU

85 0

-4.6 0.0 -4.7 0.0 -7.3 0.0

0.8 0.0 -1.5 0.0 -4.0 0.0
ELBM

6 0

-6.9 0.0 -7.1 0.0 -11.8 0.0

0.3 0.0 -4.5 0.0 -9.5 0.0
ELGT

1 0

-7.0 0.0 -7.3 0.0 -12.1 0.0

0.6 0.0 -4.5 0.0 -9.6 0.0
ELHY

7 0

12.2 0.0 11.9 0.0 6.5 0.0

21.2 0.0 15.1 0.0 9.2 0.0
ELSL

0 0

-7.0 0.0 -7.3 0.0 -12.1 0.0

0.5 0.0 -4.5 0.0 -9.6 0.0
ELWD

3 0

-6.8 0.0 -7.1 0.0 -11.7 0.0

0.2 0.0 -4.5 0.0 -9.4 0.0
GASU

98 35

-1.9 -1.9 -1.9 -2.0 -3.3 -3.4

-1.3 -1.4 -2.7 -2.8 -4.0 -4.2
PWAT

7 0

0.5 -0.1 0.5 -0.2 -0.4 -1.3

1.3 0.8 0.5 -0.3 -0.4 -1.4
SANT

7 2

-1.8 -2.6 -1.8 -2.7 -2.9 -4.1

-1.4 -2.1 -2.4 -3.5 -3.5 -5.0
WTRD

1,170 27

-3.1 -8.7 -3.2 -8.9 -5.1 -14.1

-2.3 -6.6 -4.3 -11.8 -6.2 -16.9
RTRD

1,304 38

-5.8 -16.6 -6.0 -17.2 -10.2 -28.1

-4.0 -11.7 -8.2 -22.9 -12.4 -33.6
REST

1,072 10

-3.2 -3.5 -3.3 -3.6 -5.4 -5.7

-2.4 -2.7 -4.4 -4.7 -6.5 -7.0
BANK

809 3

-4.5 -6.8 -4.6 -7.1 -8.1 -12.3

-3.0 -4.7 -6.4 -9.8 -9.9 -15.0
SECB

375 1

-0.9 -4.6 -0.9 -4.8 -1.6 -8.2

-0.6 -3.1 -1.2 -6.6 -1.9 -9.9

109

INSR

600 0

-4.5 -9.9 -4.7 -10.3 -8.2 -17.6

-3.1 -6.8 -6.5 -14.2 -10.0 -21.3
OODW

1,011 0

-4.4 -3.8 -4.6 -3.9 -7.6 -6.7

-3.1 -2.6 -6.1 -5.4 -9.1 -8.0
HOTR

671 11

-2.1 -3.6 -2.2 -3.8 -4.2 -7.0

-1.3 -2.3 -3.2 -5.4 -5.5 -8.9
PSRV

153 5

-5.5 -9.6 -5.7 -10.1 -10.5 -17.9

-3.5 -6.3 -8.2 -14.1 -13.1 -21.9
VSRV

30 0

-4.8 -11.0 -5.1 -11.5 -9.5 -20.5

-3.1 -7.2 -7.4 -16.3 -12.0 -25.2
WAST

68 47

-1.2 -1.7 -1.3 -1.8 -2.1 -3.0

-0.9 -1.3 -1.7 -2.5 -2.6 -3.7
OBSV

3,224 48

-1.9 -5.4 -2.0 -5.6 -3.1 -8.7

-1.5 -4.2 -2.6 -7.3 -3.8 -10.5
ENTR

204 2

-4.5 -9.5 -4.7 -9.9 -8.5 -17.3

-2.9 -6.3 -6.6 -13.7 -10.7 -21.2
EDUC

750 1

-0.6 -6.6 -0.7 -6.9 -1.2 -11.8

-0.4 -4.5 -1.0 -9.5 -1.5 -14.4
MEDC

1,333 11

* * * * * *

* * * * * *
OSOC

185 14

-4.3 -12.2 -4.5 -12.7 -8.5 -22.4

-2.8 -8.1 -6.5 -17.7 -10.8 -27.4
TLTG

26 9

0.3 0.5 0.3 0.5 0.4 0.6

0.3 0.4 0.4 0.5 0.5 0.5
GVUT

44 0

-1.1 0.0 -1.2 0.0 -2.0 0.0

-0.8 0.0 -1.6 0.0 -2.4 0.0
FGML

172 0

0.0 0.0 0.0 0.0 0.0 0.0

0.0 0.0 0.0 0.0 0.0 0.0
OGOV

311 43

-0.5 -5.2 -0.5 -5.4 -1.0 -9.3

-0.3 -3.6 -0.8 -7.5 -1.3 -11.3
SGGV

460 0

0.0 0.0 0.0 0.0 0.0 0.0

0.0 0.0 0.0 0.0 0.0 0.0
Total

24,994 4,504

-3.9 -8.3 -4.0 -9.7 -5.6 -14.1

-3.1 -6.0 -4.9 -9.7 -6.5 -14.4
ELEC

327 2,305

-7.0 -8.8 -7.1 -9.0 -10.8 -13.0

-2.5 -4.7 -6.1 -8.7 -9.7 -12.8
EF

226 2,305

-8.5 -13.1 -8.7 -13.3 -12.6 -17.3

-4.6 -9.0 -8.5 -13.0 -12.5 -17.0
ENF

102 0

-3.6 0.0 -3.8 0.0 -6.8 0.0

2.2 0.0 -0.6 0.0 -3.6 0.0
Kerry

7,093 3,312

-0.4 -9.6 -0.4 -9.7 -0.5 -12.7

-0.5 -6.7 -0.6 -9.7 -0.7 -12.7
PR = Output; EM = Carbon emissions; ELEC = Electricity generation sectors; EF = Electricity Generation Fossil Fuel sectors; ENF =
Electricity Generation Non-Fossil Fuel sectors; KL = Sectors included in the Kerry-Lieberman (Broad-based) simulation.
†
The Other
Agriculture sector (AOTH) incorporates offsets within the CGE analysis, though these are not reported here. Please see the breakdown
of emissions by sector groupings in Table 10 for offsets results. *Zero profit condition constraints on the Medical Sector (MEDC)
were relaxed to facilitate model processing; hence credence cannot be given to results for this sector.

Table 13 reports the impact to output per ton of emissions reduced for Cases 1, 2, and 3.
Counter-intuitively, the impact to output per ton abated is lower as the cap level becomes more
stringent. The impacts to the electricity-generating sectors are increasing per ton of carbon
abated. In contrast, the impacts to regulated non-electricity sectors are reducing substantially in
costs as the cap is tightened – thus explaining the reduction to regulated sectors overall. Indeed, in
Case 1, the regulated non-electricity sectors are growing in output while reducing emissions
levels. This is at least partly explained by the presence of the manufacturing electricity-use
efficiency improvements, as the regulated non-electricity sectors in Cases 2 and 3 are not growing
in output, and impact to output per ton is much higher than even the electricity generating sectors.

110

Table 13: Impact to output per ton of emissions reduced
($/ton CO
2
)

for regulated sectors and sub-groupings.

Case 1: Electricity efficiency for manufacturing
(regulated, non-electricity) at 5%; AEEI at 1%.

Emissions reduction (%)
Sector Groups 6.7 9.7 12.7
Regulated Sectors 137.7 108.3 96.0
Electricity 106.9 114.6 115.0
Coal 74.7 75.1 75.3
Gas 211.0 211.8 212.3
Oil 86.7 87.4 87.2
Other Fossil 90.2 91.1 91.6
Non-electricity 1818.1 -18.6* -215.3*
*Negative abatement costs may seem counter-intuitive, but these cases are
possible when sectors reduce emissions while increasing output. **The Gas
and Regulated Non-electricty sectors increases emissions in the base case,
and hence do not have a calculable abatement cost.

Case 2: Electricity efficiency for manufacturing
(regulated, non-electricity) at 3.5%; AEEI at 1%.

Emissions reduction (%)
Sector Groups 6.7 9.7 12.7
Regulated Sectors 141.4 139.7 119.1
Electricity 94.3 94.3 94.8
Coal 75.4 75.4 75.5
Gas 212.1 212.1 212.5
Oil 87.1 87.1 87.1
Other Fossil 91.7 91.7 92.0
Non-electricity 523.6 493.5 136.1

Case 3: No electricity efficiency for manufacturing; AEEI at 1%.

Emissions reduction (%)
Sector Groups 6.7 9.7 12.7
Regulated Sectors 274.7 181.0 147.2
Electricity 95.9 95.7 95.7
Coal 76.2 75.9 75.8
Gas 212.3 212.4 212.8
Oil 88.0 87.5 87.3
Other Fossil 93.0 92.6 92.6
Non-electricity 1,828.5 934.0 530.0

111

Tables 14 and 15 show that findings in the previous simulations are extended across a
broader range of emissions cap levels. Table 14 highlights the relative stability of impacts to
output per emissions abated for the electricity generating sectors. Indeed, for the Oil-generated
electricity sector, the impact to output per ton of carbon abated is almost constant as the cap level
increases. In contrast there is more movement for the Gas-generated and Other-fossil-generated
electricity sectors. Table 15 similarly confirms the counter-intuitive result highlighted above,
namely that Coal-generated electricity (ELCL) becomes a net seller of emissions allowances. It is
only the Gas-generated electricity (ELGS) sector that remains a net purchaser of emissions
allowances among electricity producers. Meanwhile, the Chemical manufacturing (MCHM),
Petroleum refining (MPET), Primary metals (MPRM), and Other durables manufacturing
(MODR) sectors all increase the number of emissions allowances purchased as the emissions cap
is tightened.
Table 14: Impact to output ($/ton CO
2
)

for regulated sectors and sub-groupings.

Emissions reduction (%)
Sector Groups 6 8 10 12 14 16
Regulated Sectors 130.9 109.3 98.4 92.9 89.8 88.2
Electricity 91.1 92.5 93.3 93.8 94.2 94.5
Coal 74.5 74.8 75.1 75.2 75.3 75.5
Gas 210.9 211.4 211.8 212.2 212.5 212.8
Oil 86.7 86.9 87.0 87.0 87.0 87.0
Other Fossil 89.9 90.6 91.1 91.4 91.7 91.9
Non-electricity 7,562.8 301.8 -131.7* -270.6* -336.3* -370.9*
*Negative abatement costs may seem counter-intuitive, but these cases are possible when sectors reduce emissions
while increasing output. **The Gas and Regulated Non-electricity sectors increases emissions in the base case, and
hence do not have a calculable abatement cost.

112

Table 15: Implied net emissions trading and allowance prices ($/ton) for Case 1 scenarios.

Emissions reduction (%)
Base 6 8 10 12 14 16
AOTH 0.00 -100.9 -100.4 -99.8 -99.1 -98.5 -97.9
COAL 9.67 0.3 0.4 0.5 0.6 0.7 0.8
CRUD 76.75 4.4 5.4 6.2 7.1 7.9 8.8
MFML 0.30 0.0 0.0 0.0 0.0 0.0 0.0
MOML 0.58 0.1 0.1 0.0 0.0 0.0 0.0
MANM 1.31 0.1 0.1 0.1 0.1 0.1 0.1
MPTY 0.72 0.1 0.1 0.1 0.1 0.0 0.0
MFSH 0.27 0.0 0.0 0.0 0.0 0.0 0.0
MOFD 9.09 0.8 0.8 0.8 0.9 0.9 0.9
MCHM 101.71 7.1 8.9 10.7 12.4 14.0 15.7
MPET 433.72 24.2 28.1 31.5 34.8 38.1 41.5
MOND 98.13 6.7 7.8 8.7 9.6 10.5 11.4
MPRM 117.93 8.2 12.2 16.5 20.8 25.2 29.6
MORD 0.04 0.0 0.0 0.0 0.0 0.0 0.0
MSEM 2.11 0.1 0.1 0.2 0.3 0.3 0.4
MODR 119.56 5.6 8.6 11.7 14.8 17.9 21.0
ELCL 1876.77 21.9 14.9 1.6 -11.7 -25.9 -39.3
ELGS 378.50 10.1 10.1 9.0 7.8 6.6 5.4
ELOL 40.64 0.5 0.0 -0.7 -1.3 -2.0 -2.5
ELOF 8.94 0.0 0.0 0.0 -0.1 -0.1 -0.2
ELNU 0.00 0.0 0.0 0.0 0.0 0.0 0.0
ELBM 0.00 0.0 0.0 0.0 0.0 0.0 0.0
ELGT 0.00 0.0 0.0 0.0 0.0 0.0 0.0
ELHY 0.00 0.0 0.0 0.0 0.0 0.0 0.0
ELSL 0.00 0.0 0.0 0.0 0.0 0.0 0.0
ELWD 0.00 0.0 0.0 0.0 0.0 0.0 0.0
GASU 35.28 2.2 2.7 3.0 3.4 3.8 4.2
Total -8.43 0.00 0.00 0.00 0.00 0.00
Emissions Allowance Price 11.7 12.7 14.1 15.7 17.6 19.5

CONCLUSION
To conclude, let us examine the results in light of the proposed hypotheses in the
Introduction Chapter. The first hypotheses – H1: Aggregate results will be less than negative 1%
impact on output (Rose & Dormady, 2011; Rose, Wei & Prager, 2012) – is not supported by the
results. Instead, the impacts are larger for more stringent emissions reductions levels, thus
supporting the second hypothesis – H2: Coase Theorem: Aggregate results will be largely
unaffected by the allocation schedule (Coase, 1960). However, it is notable that the aggregate
economic impacts with respect to emissions allowance allocation – i.e. the broad-based vs.

113

narrow-based policy results – were not substantially different. This reflects the claims of the
Coase Theorem, that aggregate social welfare should be unaffected by emissions rights
allocations. The results were also shown to be somewhat sensitive to key assumptions regarding
electricity-use efficiency improvements in the manufacturing sectors. While the macro-economic
impacts were not substantially altered by changes to these assumptions, the impacts with respect
to particular sectors were. This highlights the importance of policy design being sensitive to
modeling assumptions.
The third set of hypotheses related to sectoral impacts: H3a: GHG-intensive sectors are
likely to bear the greatest regulatory burden; H3b: Downstream and upstream sectors would
absorb some of this burden. Hypothesis H3a was shown to be correct: regulated sectors in both
policy designs (broad-based and narrow-based policies) were the most negatively impacted
sectors in percentage terms. Non-regulated sectors have not only absorbed some of the burden,
per hypothesis H3b, but experience large negative shocks to output also. These results highlight
the interconnectedness of economies CGE aims to represent. Shocks to one area of the economy,
especially one as fundamental as the electricity generation sector, will have potentially substantial
impacts on other sectors.
Hypotheses H4a and H4b aimed to highlight the potential benefits to particular sectors:
H4a: Green technology and government sectors are likely to benefit relatively more from ETP;
H4b: Sectors importing and exporting green technology and GHG-intensive substitutes are likely
to benefit relatively from ETP. Of the renewable electricity generation sectors, only the hydro-
generated electricity sector increased output. Government sectors fared relatively well in the
simulations.

114

CHAPTER THREE:
INTER-HOUSEHOLD EQUITY MODELING

―By closing Redcar's annual capacity of three million tons of steel, Corus will
produce six million fewer tons of CO2. That means more carbon allowances,
which could translate into about $300 million a year if credits hit $50. Corus is
essentially being paid to lay off British workers.‖
Wall Street Journal, December 17 2009
―Tomorrow morning I will feel like I have been to a wake‖
John Wakelin, Former Redcar Steelworks Employee. February 19 2010

The above quote is describing one of the controversial elements of ETP, a policy mechanism used
to reduce industrial GHG emissions. The steel plant in Redcar, located in England‘s Northeast
region, was ―mothballed‖ as a consequence of the European Union Emissions Trading Scheme‘s
(EUETS) implementation in 2005. The mothballing of the Corus Redcar plant provides an
example of the potentially negative impacts from ETP. Following the logic of ETP, Corus‘
Redcar plant was simply one of the cheaper places to reduce emissions.
The mothballing of the Corus Redcar plant provides an example of the potentially
negative impacts from ETP. The closure of the Redcar plant also raises the important equity
considerations explored in this dissertation. In closing the Redcar plant, Corus has laid off
numerous, likely middle-class workers, whose income also contributed to the local economy.

115

Meanwhile, wealthier capital investors in Corus would see no impact to their income from the
company. The question then becomes: Is this situation reflective of a broader reality that ETPs
inadvertently transfer funds from lower to higher economic income brackets? In other words,
what are the distributional impacts of ETP?
ENVIRONMENTAL AND CLIMATE EQUITY
These questions highlight broader policy debates around the issue of environmental
inequalities. Since the 1970s there has been concern that the ―sources of potential environmental
risk may be concentrated among racial and ethnic minorities and the poor‖ (Ringquist, 2005).
Sociologist Robert Bullard first studied landfills and hazardous materials sites in Houston, Texas,
finding that all were within or nearby African-American communities (Bullard, 1983). In
numerous studies, Bullard continued to find environmental inequalities across communities in the
US south. Since Bullard‘s pioneering a substantial environmental justice literature has
demonstrated environmental inequalities with respect to race – though not necessarily class
(Ringquist, 2005) – within the US. In tandem to a growing academic literature, grassroots
community campaigns began to develop around a variety of environmental justice issues
(Schweitzer and Stephenson, 2007). These campaigns have gained traction, with successful
public policy influence (Ringquist, 2005), not least the U.S. Environmental Protection Agency‘s
recent efforts under Director Lisa Jackson to incorporate environmental justice concerns into the
rule making process:
―Environmental Justice is the fair treatment and meaningful involvement of all
people regardless of race, color, national origin, or income with respect to the
development, implementation, and enforcement of environmental laws,
regulations, and policies. EPA has this goal for all communities and persons
across this Nation. It will be achieved when everyone enjoys the same degree of
protection from environmental and health hazards and equal access to the

116

decision-making process to have a healthy environment in which to live, learn,
and work.‖ (EPA Environmental Justice website, 2013).
The potential environmental inequalities of climate change are global in nature, with
countries, as well as locations within the US, likely to be impacted with substantial variation. One
of the most discussed divides in the climate change literature is the gap between ―North and
South,‖ or ―Developed and Developing‖ countries (Adger et al. 2006). Such distinctions, while
highlighting important differences between impacts to countries as sovereign units, can imply that
there are deterministic correlations between region and economic power; e.g., that all
communities within Sub-Saharan African countries are vulnerable to climate changes while
Western European countries will have the resources with which to adapt. Such crude distinctions
obscure the important regional (e.g. EU vs. ASIA) and sub-regional (e.g. within North America)
differences – which are discussed further in the following chapter, as well as the differential
climate impacts according to economic class, race, and gender. Such distinctions may also belie
the interconnected nature of both regional and global economies. While a climate change-induced
drought in the plains of North America would have a particularly pernicious impact on the
economies of the US, Canada, and Mexico, the impacts would be felt across the globe in terms of
agricultural commodity prices and beyond.
To the extent that poorer communities inhabit and derive livelihoods from those local
environments vulnerable to climate change – such as low-lying deltas or drought-vulnerable
plains – it follows that poorer communities are at greater threat from climate change. Even if
there was no correlation between the presence of poorer communities and vulnerable locations,
these communities would be more vulnerable because of their inability to adapt their economies
and infrastructure to climate change, and hence protect themselves against threats.

117

As the Stern Review has shown, identifying the future economic costs of climate change
is subject to substantial uncertainty. Extreme weather events are likely to contribute to a
significant proportion of the increased costs of climate change. However, parsing out the impact
due to climate change as opposed to other variables is challenging. This reality is acknowledged
in the IPCC (2007) report, which states that: ―Aside from major extreme events and thresholds,
climate change is seldom the main factor in considering stresses on the sustainability of industry,
settlements and society.‖ Data on natural disasters nonetheless provide useful insights in this
respect. As risk analysis scholars Howard Kunreuther and Erwann Michel-Kerjan observe, the
last 60 years have lead to a marked increase in the economic and insured losses from large natural
catastrophes worldwide (Figure 1). In 2005, 99.7 percent of catastrophic losses worldwide were
caused by weather-related events, and the insured losses from weather-related events increased
from an average of $3 billion per year between 1970 and 1990 to an average of $16 billion per
year between the years 1990 and 2004 (in equivalent dollars). The authors attribute this increase
to a combination of climate change and increased development in hazard-prone areas of the
world; there is certainly a correlation between the increase in the costs of extreme weather events
and increased average global temperatures over this period. However, no attempt is made to
identify the magnitude of change that different factors may account for.

118

Figure 7: Economic and insured losses from great natural catastrophes worldwide, 1950-2007
(Kunreuther and Michel-Kerjan, 2009).
Attributing the costs of climate change across income groups within society is more
challenging still. Nonetheless, more general statements are supportable. For example, as the IPCC
(2007) asserts: ―Poor communities can be especially vulnerable, in particular those concentrated
in relatively high-risk areas. Adverse health impacts will be greatest in low-income countries.
Those at greater risk include, in all countries, the urban poor, the elderly and children, traditional
societies, subsistence farmers, and coastal populations.‖ Moreover, it is clear that when poor
communities are caught up in extreme weather events, the outcomes are often disturbing. As
sociologist Kathleen Tierney observes, the lack of disposable income can place a substantial
burden on poor communities during extreme weather events. In the US, community evacuation
plans often assume that residents have access to private transportation. ―In New Orleans, for
example, it was recognized that approximately 300,000 residents would not be able to evacuate
on their own in a timely manner under the threat of a hurricane. Instead, they were dependent
upon governmental capacity to bring in vehicles (and drivers) to assist with evacuation – capacity

119

that was nonexistent in Hurricane Katrina.‖ Should households have access to private
transportation, fuel prices in the aftermath of extreme weather events are likely to rise steeply,
causing further vulnerability. Moreover, low income households are also more vulnerable to
income losses from taking time off work.
Alongside insufficient evacuation plans, government policies have often faced criticism
for favoring more affluent communities. For example, Tierney points out that disaster assistance
programs are often based on a ―one housing unit, one head of household‖ policy, which unfairly
discriminates against those residing in multi-family households. Moreover, programs
administering small business loans following disasters have been criticized for favoring white,
Hispanic, and Asian community businesses over their African-American community counter-
parts. Extreme heat waves are more likely cause suffering in communities of color and the poor
(Morello-Frosch, Pastor, Sadd, and Shonkoff, 2009). For example, ―African Americans in Los
Angeles are nearly twice as likely to die from a heat wave than other Los Angeles residents‖
while poor families are less likely than others to have access to cooling systems (Morello-Frosch
et al., 2009).
Alongside the inequities observed with respect to extreme weather events, there less
dramatic, yet possible just as harmful, day-to-day climate change inequalities. There is already
substantial evidence across a broad environmental justice literature that communities of color and
the poor are more likely to live close to polluting industrial entities, hazardous waste sites, and
transportation systems (Ringquist, 2005). Increased temperatures would speed up the chemical
processes that form smog, a major contributor to deaths associated with air-pollution (Jacobson,
2008. Alongside these disparate health impacts, climate change is likely to increase the costs of
necessity goods – which poorer households spend a larger proportion of their income on – reduce

120

job opportunities in sectors employing low-income workers, and increase the price of extreme
weather insurance, thus pricing out poorer sections of society (Morello-Frosch et al, 2009).
EQUITY IN ENVIRONMENTAL AND CLIMATE POLICY
Assuming that climate costs will be borne disproportionately by minority and poorer
communities, a ―do nothing‖ climate policy approach would implicitly be inequitable. The
question would then become one of which policy framework would best address the inequitable
harms of climate change, while also accounting for inequitable harms from policy intervention. It
is often argued in the economics literature that, in the ideal, a policy regime would effectively
compensate those poorer households relatively harmed by transferring funds from those wealthier
households relatively benefitting (or less harmed). The problem is how to implement a clean
compensation regime in a messy world and political system. For example, it could be argued that
governments already compensate poorer households through progressive taxation and transfer
systems, which are criticized for being complex, bureaucratic, and ineffective with respect to
matching funds to policy goals. It could also be argued that compensation regimes are insufficient
responses to the problem. The costs of environmentally-related poor health can be estimated in
monetary terms, but is it possible to truly account for the harms that such impediments can bring
to individuals and their families?
This last position would suggest that instead of compensation, focus should be place on
reducing the initial environmental and climate harms. This does appear to be most common
policy response to environmental harms; as detailed in the previous chapter, governments have
employed a mixture of strategies including emissions reductions mandates, emissions taxes, and
emissions trading programs. Yet such programs are not without potentially inequitable impacts.
As the quotes at the beginning of the chapter highlight, the European Union Emissions Trading

121

System has lead to some lower income households being negatively impacted while wealthier
capital owners are, at least, unharmed economically. Indeed, this is one of the six mechanisms
Don Fullerton (2008) identifies as potentially contributing to the inequitable impact of climate
policy:
49

1. Emissions rights are granted to wealthy capital owners,
2. Wages and jobs are reduced for poorer households when abatement technologies are
capital intensive, (or if income payments to poorer households are more likely to
come from GHG intensive industries),
3. (Poor, less educated and) unemployed individuals are less able to transition to the
new green economy,
4. Increases in fossil-fuel prices (or any GHG intensive commodities), which are a
higher proportion of low-income budgets,
5. Poor households may value current consumption greater than incremental
environmental improvements,
6. Air-quality improvements may increase property prices, benefitting wealthier
property owners over poorer renters (not to mention the potentially uneven health
impacts potentially caused by pollution changes with respect to ancillary pollutants).
Rose and Oladosu (2007: 4-5) highlight additional factors contributing to inequitable climate
policy impacts:
1. ―The extent to which general equilibrium effect are taken into account to capture
production/income distribution/consumption in response to the policy. For example a
large decrease in coal production may have a disproportionate effect on income of
high-wage unionized miners, but the decrease in their consumption may be for
products that are characterized by a predominant number of low-wage earners.

49
These six mechanisms are paraphrased from Don Fullerton‘s work. Those sections in parentheses have
been added by this author. Oladosu and Rose (2007

122

2. The extent to which dynamic effects are taken in account (e.g. with respect to savings
and investment).
3. The extent to which demographic considerations pertaining to household composition
are taken into account; related to this is the demarcation of income groups, especially
at the highest and lowest levels.
4. The type of revenue recycling (including lump-sum transfers) and in contrast to
alternatives such as budget deficit reduction and individual and corporate tax relief.‖
Most recent studies of the equity impacts of climate policy focus on the fourth of
Fullerton‘s mechanisms (2008), i.e. the welfare changes with respect to consumption (Burtraw et
al., 2009; Hassett et al., 2009; Metcalf et al., 2008; Metcalf, 2009; Parry, Sigman, Walls, &
Williams, 2005). Climate policies can impact household consumption by causing price increases
for GHG-intensive products and their downstream counterparts. As shown in the 2011 Consumer
Expenditure Survey data in Table 16, items such as ―Utilities, fuels, etc‖ and ―Food,‖ which are
particularly vulnerable to emission control price changes, are a larger proportion of lower-income
household budgets. Interestingly, the lowest income brackets spend a relatively low proportion of
their income on ―Gasoline and motor oil,‖ though spending on this item is nonetheless lower for
those with income more than $70,000. It is also notable that those earning more than $70,000
spend a significantly greater than average proportion of their income on ―Insurance and
pensions.‖ This reflects an important correlation between income levels and savings rates.
Increased disposable income after necessity goods are purchased allows for U.S. households to
invest their future.
Analyses have confirmed that low income households tend to be more vulnerable to price
changes of GHG-intensive products because they spend a greater proportion of their income on
these goods, especially heating, transport, and foods (Burtraw et al., 2009; Hassett et al., 2009;
Metcalf et al., 2008; Metcalf, 2009; Parry, Sigman, Walls, & Williams, 2005). Household
consumption is used as a proxy for lifetime income following Milton Freidman‘s ―permanent
income‖ hypothesis, and hence aims to the welfare changes for those long-term poor which are of

123

interest to many policy makers. This approach also avoids one problem with income data, the
alternative measure welfare, whereby single-year income data may reflect temporary fluctuations
in income.
Table 16: Shares of household budgets spent on commodity items, by income bracket.
Item Avg. >$5k
$5-
10k
$10-
15k
$15-
20k
$20-
30k
$30-
40k
$40-
50k
$50-
70k
>$70k
Food 13.0 15.0 17.2 16.4 15.1 14.1 14.5 12.8 13.1 12.0
Alcoholic beverages 0.9 1.0 1.1 0.7 0.6 0.7 0.9 0.8 0.8 1.0
Housing 33.8 37.4 38.9 42.5 39.7 38.6 36.5 36.1 33.8 31.4
Utilities, fuels, etc 7.5 9.1 9.9 11.4 11.0 9.8 9.1 9.1 8.0 6.1
Apparel and services 3.5 4.9 4.5 3.7 3.3 3.1 4.0 2.9 3.5 3.5
Transportation 16.7 15.1 15.3 12.8 16.2 15.3 17.4 17.7 18.0 16.6
Gasoline, motor oil 5.3 5.0 5.3 5.9 6.0 6.5 6.1 6.6 5.9 4.7
Public, other trans. 1.0 1.0 0.7 0.9 0.7 0.6 0.8 0.9 0.8 1.3
Health care 6.7 5.4 5.3 7.9 8.2 8.7 7.5 8.2 7.4 5.8
Education 2.1 8.3 4.5 1.7 1.4 1.7 1.2 1.3 1.3 2.4
Tobacco products 0.7 1.3 1.5 1.7 1.3 1.2 1.0 1.0 0.8 0.4
Insurance and pensions 10.9 1.4 1.4 2.1 2.7 4.9 6.4 8.2 9.7 14.8
Source: Consumer Expenditure Survey, 2011 (http://www.bls.gov/cex/tables.htm).
The permanent income hypothesis has been shown to be empirically inaccurate, and
hence other studies have analyzed the welfare impacts of climate policy via Fullerton‘s second
and third mechanisms, i.e. income and employment changes. ETP would be regressive if lower
income households are more dependent on greenhouse-gas intensive industries for income, or are
less able to transition to emerging green industries (Rose, Wei and Prager, 2012). Those studies
with a broader definition of income distribution impacts often use computable general
equilibrium (CGE) modeling, which can incorporate both consumption and income changes
(Fullerton & Heutal, 2007; Oladosu & Rose, 2007; Rausch et al., 2010). Using this
comprehensive approach, Oladosu and Rose (2007) find a regional-level policy would be
relatively progressive, imposing greater costs on higher than lower income brackets. Like Parry
and Williams (2010), this paper combines both consumption and income effects in order to

124

achieve a more realistic treatment of welfare changes, as well as a triangulated and multi-
dimensional image of equity.
Given these likely inequalities of climate policy, the Schultze principle of ―do no direct
harm‖ (1977) would certainly imply that compensation mechanisms should be incorporated into
climate policy mechanisms. This principle is based on deeper rationales. For instance, Hochman
(1974) argues that, as citizen behavior responds to current policy frameworks, compensation for
groups harmed by policy change could be critical in upholding the legitimacy of government
institutions. ETP presents further opportunities for policymakers to balance economic efficiency
and equity goals. The Coase theorem asserts that emissions markets can achieve equivalent
economic outcomes regardless of which actors receive the emissions rights. This presents an
opportunity for ETP to achieve equity goals. Emissions rights could be, for example, allocated
directly to lower income households, or to economic sectors which employ lower income
workers. Another option is for government revenues raised through allowance sales to industry to
be redistributed to lower income households to offset any uneven policy cost distribution.
However, each of these options would also result in consequential ―second-order‖ policy
costs, both in terms of economic efficiency and equity. Allocating emissions rights to poorer
households as opposed to industry would increase industrial costs and likely reduce economic
efficiency. For example, Rose, Wei and Prager (2012) found that while revenue redistribution
would result in more equitable outcomes, there was an significant trade-off with economic
efficiency, reflecting results elsewhere (Parry and Williams, 2010). Equally, poorer households
benefit from an initial stimulus may consume commodities in such a way that benefits ―trickle
up‖ to wealthier households. Hence the equity improvements from the initial allocation may be
quickly softened over time.

125

Because of the variability in equity indicators and policy context
50
discussed above,
recent studies necessarily present only a partial analysis of ETP. Further variability in policy
analysis is provided by the modeling approach used. Parry and Williams (2010) use mathematical
models specifically designed to capture the consumption impacts of climate policy. Detailed
electricity market models – such as the HAIKU model – have been used to capture the important
impacts of electricity prices on consumption decisions (Burtraw et al., 2011), across both income
brackets and regions of the U.S. Rose, Wei, and Prager (2012) capture the income effects of
climate policy in the California context using a regional macro-econometric model (REMI).
However, the majority of studies in this area have used CGE modeling (Goettle & Fawcett, 2009;
Oladosu and Rose, 2007; Paltsev et al, 2009; Rausch et al., 2010).
This dissertation explores variability in the policy type and design mechanisms by
holding constant the geographic policy context – the U.S. federal level – and the CGE modeling
approach. This paper enhances the USCGE model of the U.S. economy developed by Gbadebo
Oladosu and Adam Rose (Rose et al, 2009; Oladosu and Rose, 2007, Rose and Oladosu, 2002;
Oladosu, 2000) to compare two types of ETP: 1) a narrow-based policy that focuses on regulating
the electricity generation sector
51
; and 2) a broad-based policy similar to the Kerry-Lieberman
senate bill of 2009, which centered on an ETP.
52
This paper captures both income and
consumption in an integrated approach in order to assess the equity impacts of ETPs.

50
In addition to examining different policy designs, studies have also examined different geographic
contexts for ETP, including California (Rose, Wei and Prager, 2012; Roland-Holst, 2010), the U.S. federal
level (Burtraw et al, 2011), and Australia (Adams, 2007) among numerous others.
51
This narrow-based scenario is similar to the Regional Greenhouse Gas Initiative policy design, where the
electricity generation sector is the focus of regulation.
52
This broad-based scenario is similar to the European Union Emissions Trading System and the California
Global Warming Solutions, where polluting manufacturing sectors are regulated alongside electricity
generation sectors.

126

Table 17: Summary of model features designed to capture equity impacts
Model
element
Description of notable features
Equity
impact model
element
Multi-Sector Income Distribution Matrix (Rose, Wei and Prager, 2012) calculated using a
U.S. Economic Census (2011), U.S. IRS (2011), and U.S. BLS (2011) data; serves as basis
for labor and capital income allocation from sectors to households.
Household consumption effects determined by combination of factors including incomes,
savings rates, commodity prices, and household preferences.
Distributional impacts assessed in terms of relative and absolute measures, reflecting multi-
dimensional equity considerations.

As shown in Table 17, this study also uses a comprehensive approach to analyze equity
changes by combining both income and consumption effects in the CGE context. A Multi-Sector
Income Distribution Matrix (MSIDM) provides resolution to the income distribution which is
found only in a handful of studies in the literature (Rose et al., 1988; Oladosu & Rose, 2007;
Rose, Wei & Prager, 2012). The MSIDM relates sectors of the economy to income brackets in
terms of both labor and capital income payments, yet is not readily available in government data
sources, and must be constructed with data from the various sources.
53
These matrices enable
high-resolution analyses of inter-household distribution of economic impacts at the national level.
The MSIDM data are incorporated endogenous into the USCGE model labor and capital income
equations, ensuring that the sector-household income relationships are incorporated into the
production calculations.
Tables 18 and 19 present the respective labor and capital income distributions for the
most carbon intensive sectors of the economy. If we assume a first-order world, we would expect
to see an ETP negatively impact directly-regulated sectors. As such labor incomes for the lowest
income brackets would be negatively and disproportionately impacted because of the

53
Economic data will also be acquired from Bureau of Labor Statistics (occupations and employment by
industry, wages, consumer expenditure), U.S. IRS (income and tax), U.S. Bureau of the Census
(households per income bracket), and Bureau of Economic Analysis (sectoral output) to produce the
MSIDM that sits at the core of the distributional analysis.

127

concentration of low incomes in the petroleum refining (MPET), non-durable manufacturing
(MOND), chemical manufacturing (MCHM), and primary metals (MPRM). However, these may
be offset by negative impacts to capital incomes, which tend to be clustered in the higher income
brackets, both for carbon intensive sectors, as well as the economy as a whole. This analysis
captures both effects simultaneously.
The USCGE model assumes that household consumption responds to income changes at
different levels for each household income bracket. After governments collect taxes from labor
and capital income, the remaining income goes to households and foreign entities according to
the MSIDM fixed shares. Transfers also occur between institutions in the form of subsidies,
social security payments, and income taxes. The resulting incomes are spent by households on
commodities. A Linear Expenditure System of aggregate commodities (such as Food, Housing,
and Gasoline) represents household consumption behavior. Commodity prices also adjust in
response to the policy shock and household demands are subject to substitution functions in
response to these price changes.

Table 18: Labor income distribution for sectors with largest carbon emissions.
Sector LT10k 10-15k 15-25k 25-35k 35-50k 50-75k 75-100k 100-150k 150k+
ELEC 2.67% 0.91% 0.60% 1.85% 6.09% 18.10% 26.02% 19.57% 24.21%
MPET 10.84% 0.00% 0.58% 1.75% 5.93% 16.36% 20.47% 15.59% 28.48%
MOND 10.32% 0.00% 4.74% 10.06% 15.98% 20.25% 12.66% 7.31% 18.68%
MCHM 10.05% 1.17% 1.10% 3.23% 7.48% 17.01% 15.38% 12.63% 31.94%
MPRM 19.83% 0.52% 0.96% 6.77% 15.94% 23.45% 13.05% 7.14% 12.35%
MODR 9.43% 0.34% 2.14% 7.59% 14.71% 19.76% 14.81% 10.44% 20.77%
Economy
wide 2.15% 2.07% 5.91% 7.15% 10.62% 15.96% 13.24% 11.23% 31.66%
ELEC = Electricity generation; MPET = Petroleum refining; MOND = Non-durable goods manufacturing; MCHM = Chemicals
manufacturing; MPRM = Primary metals manufacturing; MODR = Durable goods manufacturing.

128

Table 19: Capital income distribution for sectors with largest carbon emissions.
Sector LT10k 10-15k 15-25k 25-35k 35-50k 50-75k 75-100k 100-150k 150k+
ELEC 0.00% 0.60% 1.22% 1.26% 1.98% 4.13% 4.38% 6.19% 80.23%
MPET 0.00% 0.42% 1.00% 0.99% 1.74% 3.47% 4.14% 6.01% 82.22%
MOND 0.00% 0.38% 0.95% 0.94% 1.67% 3.39% 4.06% 5.91% 82.69%
MCHM 0.00% 0.32% 0.86% 0.88% 1.58% 3.26% 3.94% 5.77% 83.39%
MPRM 0.00% 0.31% 0.85% 0.87% 1.57% 3.25% 3.92% 5.75% 83.47%
MODR 0.00% 0.32% 0.86% 0.88% 1.58% 3.26% 3.94% 5.76% 83.41%
Economy
wide 0.00% 0.57% 1.19% 1.15% 1.90% 3.79% 4.23% 6.04% 81.14%
ELEC = Electricity generation; MPET = Petroleum refining; MOND = Non-durable goods manufacturing; MCHM = Chemicals
manufacturing; MPRM = Primary metals manufacturing; MODR = Durable goods manufacturing.
Another debate in the literature which is related to the income and consumption
distinction discussed above is whether to use relative or absolute measures of poverty and welfare
change. Most countries within the Organization for Economic Cooperation and Development
(OECD, an international organization comprising the most developed nations) use relative
measures of poverty, categorizing those in poverty as households earning a percentage
(commonly 60 percent) of median income. Relative measures have theoretical limitations,
however. If all individuals across the distribution were to double their income, the relative wealth
would remain the same, with the same number in poverty, despite clear welfare gains across the
board. The U.S. avoids this particular problem because it is the only major developed country to
use absolute measures in its definition of poverty. Relative measures do reveal some important
aspects, however. Perception matters, and the relative wealth of an individual‘s peers plays a
significant role in influencing happiness levels, a finding present in both survey results
(Blanchflower and Oswald, 2004) and brain scan imagining (Fliessbach et al., 2007).
EQUITY IMPACTS OF EMISSIONS TRADING POLICY
The results in Table 20 are interesting to analyze with respect to the Coase theorem. The
Coase theorem states that economic efficiency is the same regardless of which actors receive

129

emissions rights. In this case, whether the emissions rights are allocated to the electricity-
generating sectors only (as in the narrow-based policy) or a larger number of sectors (as in the
broad-based policy) appears to have minimal impact on gross output.
54
The Coase theorem also
implies that economic impacts of externality rights allocations may be uneven across society. It
stands to reason that when rights are concentrated in the hands of a particular group – for
example, capital owners – that group is likely to benefit in comparison to an even allocation of
rights – for example, per household. The same logic applies with respect to allocations across
different sectors of the economy. We see that in the results here, with electricity generation
sector, where non-fossil electricity generation sectors are more negatively impacted in the
narrow-based policy (see row ―ENF‖ in Table 20). Are these different impacts by sector reflected
in terms of income and consumption across the economy as a whole?

Table 20: Comparison of broad-based and narrow-based policy results across emissions cap
levels. Percent changes to output and emissions by sector.

Broad-based policy

Narrow-based policy
Base Case
Level

Emissions reduction (%)

Emissions reduction (%)

6.7 9.7
12.7
6.7 9.7 12.7
Sector
s

PR EM

PR EM PR EM PR EM

PR EM PR EM PR EM
Total

24,994 4,504

-0.6 -5.3 -3.0 -9.7 -4.8 -13.8

0.0 -4.8 -2.8 -9.5 -4.7 -13.8
ELEC

327 2,305

-3.9 -5.1 -7.4 -9.2 -10.7 -13.2

-5.2 -6.7 -7.1 -8.9 -10.9 -13.3
EF

226 2,305

-4.8 -9.5 -8.7 -13.5 -12.7 -17.5

-6.3 -11.1 -8.4 -13.2 -12.7 -17.6
ENF

102 0

-1.8 0.0 -4.4 0.0 -6.2 0.0

-2.8 0.0 -4.1 0.0 -6.9 0.0
Kerry

7,093 3,312

-0.2 -6.7 -0.3 -9.7 -0.4 -12.7

-0.3 -7.9 -0.3 -9.5 -0.4 -12.8

PRD = Output; EMS = Carbon emissions; ELEC = Electricity generation sectors; EF = Electricity Generation Fossil
Fuel sectors; ENF = Electricity Generation Non-Fossil Fuel sectors; KL = Sectors included in the Kerry-Lieberman
(Broad-based) simulation.
†
The Other Agriculture sector (AOTH) incorporates offsets within the CGE analysis, though
these are not reported here. Please see the breakdown of emissions by sector groupings in Table 10 for offsets results.
*Zero profit condition constraints on the Medical Sector (MEDC) were relaxed to facilitate model processing; hence
credence cannot be given to results for this sector.

54
As discussed in the previous chapter, the exception here is the least stringent cap of the narrow-based
policy, which is non-binding to the emissions constraint.

130

Table 21: Comparison of broad-based and narrow-based policy results across emissions cap
levels. Level ($Bn) and percent changes to household income distribution by bracket and Gini
coefficients.

Income
brackets
Base
case
Broad-based policy Narrow-based policy
Emissions reduction (%) Emissions reduction (%)
6.7 9.7 12.7 6.7 9.7 12.7
Level %Δ Level %Δ Level %Δ Level %Δ Level %Δ Level %Δ
LT10k 142.8 139.5 -2.3 135.3 -5.2 131.5 -7.9 140.3 -1.8 135.7 -4.9 131.5 -7.9
10-15k 147.7 145.4 -1.5 136.0 -7.9 127.2 -13.8 146.9 -0.6 136.7 -7.4 127.1 -13.9
15-25k 412.6 418.3 1.4 374.0 -9.4 334.1 -19.0 425.9 3.2 378.4 -8.3 334.6 -18.9
25-35k 493.4 492.5 -0.2 458.2 -7.1 427.3 -13.4 498.3 1.0 461.6 -6.4 427.7 -13.3
35-50k 730.1 734.2 0.6 697.9 -4.4 665.3 -8.9 740.5 1.4 701.5 -3.9 665.8 -8.8
50-75k 1105.7 1125.3 1.8 1082.8 -2.1 1045.1 -5.5 1132.7 2.4 1087.1 -1.7 1045.7 -5.4
75-100k 942.0 956.8 1.6 928.2 -1.5 903.4 -4.1 961.9 2.1 931.1 -1.2 903.9 -4.0
100-150k 839.5 867.5 3.3 842.8 0.4 821.3 -2.2 871.9 3.9 845.2 0.7 821.7 -2.1
150k+ 3753.6 3777.4 0.6 3776.4 0.6 3785.6 0.9 3778.5 0.7 3775.6 0.6 3784.2 0.8
Total 8567.3 8657.0 1.0 8431.5 -1.6 8240.9 -3.8 8696.8 1.5 8452.9 -1.3 8242.3 -3.8
Gini 0.5684 0.5693 0.0009 0.5791 0.0107 0.5887 0.0203 0.5677 -0.0007 0.5781 0.0097 0.5885 0.0201

INCOME IMPACTS TO EQUITY
As shown in Table 21, while the narrow-based policy income distribution impacts are
more regressive than the broad-based policy impacts, the differences are not substantial. The
negative impacts are greater across all income brackets except the highest income bracket when
we compare the most stringent cap levels (12.7% required reduction in emissions). For example,
the 10-15k is negatively impacted by 14% in the broad-based policy, yet negatively impacted
14.3% in the narrow-based policy. Lower income brackets are more negatively affected in the
narrow-based policy; such an increase in inequality is captured by the relative increase in the Gini
coefficient (0.0203 in the broad-based policy compared with 0.0208 in the narrow-based policy).
The other notable trend is that, for both emissions allocations approaches, inequality
increases as the cap becomes more stringent. In the least stringent cap levels for both the broad-
based policy and the narrow-based policy, the income effects are positive for the higher income

131

brackets. In the aggregate, the impacts of the least stringent caps are either slightly positive or
zero, as shown in the Total row of Table 21. However, this belies a negative impact to the lower
income groups. Such inequity is captured by the Gini coefficient, which shows an increase in
inequality for even the least stringent policy in both cases. Moreover, as the cap level becomes
more stringent, the negative impacts to the lower income brackets increase while the positive
impacts to the higher income brackets are retained. This is largely because the higher income
brackets have greater proportions income payments coming from capital sources, which are also
relatively larger for sectors with low GHG intensity.
The results in Tables 22 and 23 provide further assessment of the causes behind income
payment differentials across income brackets. Tables 22 and 23 present the 12 sectors with the
largest changes in payments to labor and capital income, respectively, for the broad-based policy.
If we focus on the simulations presented here, the positive impacts to the higher income brackets
appear to be caused by a high concentration of labor income payments to the $150k+ bracket
from the Medical services (MEDC) sector. The State and local government (SGGV) sector also
has high income payments to the highest income brackets, though the distribution is relatively
more even here. Nonetheless, the combined positive impact for these two sectors outweighs the
negative impact to the highest earners in the Construction (CNSR), Other business services
(OBSV) and Education (EDUC) among others.
It is notable that most of these sectors are not regulated by the broad-based policy. In
other words the regulatory burden placed on carbon-intensive sectors appears to have been passed
on to households via labor and capital income payments in non-regulated sectors. This highlights
the idea that general equilibrium policy impacts are not necessarily predictable a priori. The one
exception is the Coal-generated electricity (ELCL), which renders a large negative impact to
capital payments in this scenario.

132

Table 22: Level change in labor income distribution for broad-based policy for 9.7%
emissions reductions.
Sector LT10k 10-15k 15-25k 25-35k 35-50k 50-75k 75-100k 100-150k 150k+ Total
SGGV 0.984 0.214 5.905 14.656 40.767 82.924 71.992 63.115 88.309 368.867
MEDC 0.317 0.003 1.412 4.784 9.402 19.860 22.531 20.776 161.256 240.340
INFO -0.341 -0.037 -0.341 -0.356 -0.635 -1.223 -1.286 -1.700 -6.791 -12.711
INSR -0.008 -0.204 -0.075 -0.517 -1.190 -2.411 -2.416 -1.890 -5.464 -14.176
BANK -0.110 -0.060 -0.305 -1.053 -2.077 -2.584 -1.896 -2.106 -7.442 -17.633
WTRD 0.000 -0.714 -0.535 -1.274 -2.345 -3.254 -2.875 -2.813 -8.537 -22.347
OGOV -0.188 -0.013 -0.363 -0.901 -2.505 -5.095 -4.424 -3.878 -5.426 -22.793
HOTR -0.133 -0.022 -17.657 -7.633 -4.304 -2.386 -0.919 -0.611 -0.943 -34.607
RTRD -0.136 -1.009 -12.614 -10.235 -7.641 -5.078 -2.408 -2.021 -6.362 -47.504
EDUC -0.126 -0.049 -0.986 -2.069 -5.703 -11.467 -10.044 -8.747 -11.808 -50.999
OBSV -1.624 -1.933 -3.854 -5.512 -7.549 -11.631 -9.995 -11.276 -38.918 -92.292
CNSR -0.035 -3.245 -3.639 -16.499 -35.275 -61.081 -54.180 -32.405 -47.168 -253.526
Total
Labor
Income -7.478 -12.296 -39.073 -35.795 -32.373 -23.541 -12.668 5.447 82.539 -75.240
Total
Income -7.478 -11.712 -38.586 -35.239 -32.163 -22.933 -13.728 3.247 22.770 -135.822

Table 23: Level change in capital income distribution for broad-based policy at 9.7%
emissions reductions cap level.
Sector LT10k 10-15k 15-25k 25-35k 35-50k 50-75k 75-100k 100-150k 150k+ Total
SGGV 0.000 0.975 1.320 1.372 1.577 3.353 2.073 2.334 3.276 16.280
MEDC 0.000 0.102 0.211 0.194 0.335 0.631 0.741 1.057 12.998 16.270
ELCL 0.000 -0.022 -0.044 -0.046 -0.072 -0.150 -0.159 -0.224 -2.907 -3.623
INSR 0.000 -0.014 -0.035 -0.036 -0.064 -0.130 -0.156 -0.227 -3.221 -3.883
COMC 0.000 -0.016 -0.038 -0.038 -0.067 -0.135 -0.162 -0.235 -3.268 -3.959
WTRD 0.000 -0.021 -0.053 -0.053 -0.095 -0.195 -0.234 -0.342 -4.854 -5.846
BANK 0.000 -0.018 -0.050 -0.052 -0.094 -0.197 -0.239 -0.351 -5.171 -6.170
REST 0.000 -0.018 -0.053 -0.056 -0.100 -0.211 -0.256 -0.377 -5.596 -6.667
OBSV 0.000 -0.022 -0.051 -0.048 -0.084 -0.183 -0.241 -0.407 -8.733 -9.770
OODW 0.000 -0.028 -0.082 -0.087 -0.156 -0.330 -0.400 -0.588 -8.725 -10.396
RTRD 0.000 -0.047 -0.111 -0.108 -0.190 -0.376 -0.449 -0.650 -8.791 -10.723
CNSR 0.000 -0.102 -0.199 -0.178 -0.303 -0.554 -0.645 -0.910 -10.448 -13.339
Total
Labor
Income 0.000 0.584 0.487 0.556 0.210 0.608 -1.060 -2.201 -59.768 -60.582
Total
Income -7.478 -11.712 -38.586 -35.239 -32.163 -22.933 -13.728 3.247 22.770 -135.822

133

CONSUMPTION IMPACTS TO EQUITY
Important equity effects are also revealed by the consumption impacts in this analysis. In
the USCGE model, changes to income influence consumption choices of households; these model
components are reflected in the results here. For example, as shown in Table 21 above, in the
least stringent policy (6.7% reduction in emissions), only the two lowest income brackets were
negatively impacted in terms of total income. These impacts are softened by the intermediate
calculation stage – i.e. disposable income – which essentially adjusts income payments to
households from sectors according to borrowings, savings, and transfers. As highlighted by the
Gini coefficient, the distribution of disposable income is much more equitable than that for total
income. Nonetheless, the policy impact to equity is actually very similar for total income and
disposable income.
Table 24: Comparison of broad-based and narrow-based policy results across emissions cap
levels. Level ($Bn) and percent changes to household disposable income by bracket and Gini
coefficients.

Income
brackets
Base
case
Broad-based policy Narrow-based policy
Emissions reduction (%) Emissions reduction (%)
6.7 9.7 12.7 6.7 9.7 12.7
Level %Δ Level %Δ Level %Δ Level %Δ Level %Δ Level %Δ
LT10k 545.1 534.7 -1.9 519.3 -4.7 505.4 -7.3 537.3 -1.4 520.8 -4.5 505.6 -7.3
10-15k 396.1 394.0 -0.5 376.6 -4.9 360.8 -8.9 396.8 0.2 378.1 -4.6 360.7 -8.9
15-25k 697.7 704.4 1.0 655.2 -6.1 611.3 -12.4 712.8 2.2 660.1 -5.4 611.8 -12.3
25-35k 713.2 713.4 0.0 679.5 -4.7 649.2 -9.0 719.3 0.9 682.8 -4.3 649.6 -8.9
35-50k 1090.0 1094.8 0.4 1053.0 -3.4 1015.9 -6.8 1102.0 1.1 1057.2 -3.0 1016.4 -6.8
50-75k 1210.2 1217.0 0.6 1176.8 -2.8 1141.0 -5.7 1224.1 1.1 1180.8 -2.4 1141.5 -5.7
75-100k 978.6 984.0 0.6 955.1 -2.4 929.7 -5.0 989.0 1.1 958.0 -2.1 930.2 -4.9
100-150k 613.5 631.4 2.9 614.0 0.1 598.9 -2.4 634.4 3.4 615.7 0.4 599.1 -2.3
150k+ 3057.0 3076.1 0.6 3073.0 0.5 3078.3 0.7 3077.4 0.7 3072.6 0.5 3077.2 0.7
Gini 0.4121 0.4146 0.0025 0.4236 0.0115 0.4325 0.0204 0.4132 0.0011 0.4226 0.0106 0.4323 0.0203

134

Tables 25 and 26 show the consumption impacts across various commodity groups
55
for
the least stringent and the most stringent emissions cap levels, respectively, of the broad-based
policy simulations. As with the changes to disposable income in Table 24, it is only the two
lowest income brackets which have a total reduction in consumption. This same pattern is
observable in Table 26, where the highest income bracket is the only one to experience increased
consumption. It is notable that consumption by the two lowest income brackets in Table 25 is not
reduced across the board. Instead, the both income brackets are increasing consumption of
Electricity (ELEC) and Water (WTER), while the $10-15k income bracket is also increasing
consumption of Other fuels (OFUE), Food (FOOD), and Household and Other Goods (EQUP,
OTHG). Indeed, it appears that price increases to Electricity (ELEC) – which result in increases
in electricity consumption across all households – is influencing reductions in consumption for
other commodities by households across all income brackets. Such price changes will affect
lower income households more because spending on necessity goods such as Electricity are likely
to be maintained in the face of price changes. All else equal, household budgets will have shrunk,
causing substitutions away from non-necessity goods. All other sectors are increasing
consumption, though Medical services (HLTH) are an interesting outlier here because total
consumption only increases marginally, while some income groups are reducing consumption.
The most favored income bracket in percent terms is the $100-150k group.

55
Services are grouped into Food (FOOD), Housing (HOUS), Gasoline (GASO), Public Transport
(LTRN), Other Transport (OTRA), Medical (HLTH), Household Goods (EQUP), Other Goods (OTHG),
Other Services (OTHS), Water (WTER), Electricity (ELEC), and Other Fuels (OFUE).

135

Table 25: Percent change in household consumption of commodity groups, broad-based policy,
emissions reductions cap at 6.7%.
Commodity
Groups
LT10k 10-15k
15-
25k
25-
35k
35-50k
50-
75k
75-
100k
100-
150k
150k+ Total
FOOD -0.40% 0.04% 0.87% 0.33% 0.32% 0.57% 0.53% 1.55% 1.27% 0.63%
HOUS -1.95% -0.19% 1.04% 0.27% 0.52% 0.17% 0.15% 0.51% 0.48% 0.23%
GASO -1.21% -0.13% 0.59% 0.70% 0.53% 0.25% 0.23% 0.68% 0.75% 0.35%
LTRN -1.30% -0.22% 0.55% 0.57% 0.49% 0.24% 0.22% 0.68% 0.65% 0.18%
OTRA -1.18% -0.07% 0.61% 0.74% 0.54% 0.25% 0.23% 0.68% 0.77% 0.42%
HLTH -3.73% -1.67% -0.09% -0.83% -0.17% 0.17% -0.04% 2.23% -1.25% -0.64%
EQUP -0.90% 0.42% 1.01% 0.60% 1.06% 0.77% 0.75% 1.94% 1.74% 1.13%
WTER -0.03% 0.19% 0.33% 0.21% 0.17% 0.11% 0.12% 0.18% 0.94% 0.26%
ELEC 0.21% 1.24% 3.13% 1.50% 1.50% 1.49% 1.60% 2.41% 4.16% 2.14%
OFUE -0.65% 0.00% 0.66% 0.23% 0.26% 0.33% 0.32% 0.88% 0.98% 0.45%
OTHG -0.99% 0.30% 0.97% 0.51% 0.97% 0.73% 0.68% 1.98% 1.48% 0.90%
OTHS -1.49% -0.31% 1.57% 0.36% 0.89% 1.42% 1.10% 6.96% 0.91% 1.06%
Total -1.85% -0.39% 1.09% 0.20% 0.66% 0.78% 0.63% 2.97% 0.70% 0.64%

Table 26 shows a much direr picture with respect to consumption. Here, reductions to
income across most income groups are reducing consumption across commodity groups, in some
cases dramatically so. For example, consumption of Medical services (HLTH) are the most
negatively impacted, which is treated as a less-necessary good. In contrast, budgets are clearly
being allocated towards more day-to-day necessity goods such as Food (FOOD), Water (WTER),
and Other fuel (OFUE). Consumption of commodities from greenhouse-gas intensive sectors
such as Electricity (ELEC) and Gasoline (GASO) is negatively impacted at rates greater than
most commodities.

136

Table 26: Percent change in household consumption of commodity groups, broad-based policy,
emissions reductions cap at 12.7%.
Commodity
Groups
LT10k 10-15k 15-25k 25-35k 35-50k 50-75k
75-
100k
100-
150k
150k+ Total
FOOD -2.59% -8.02% -9.48% -8.63% -5.30% -10.12% -9.48% -11.41% 4.70% -5.89%
HOUS -13.85% -13.75% -14.99% -12.50% -12.15% -3.76% -3.76% -3.43% -0.52% -6.64%
GASO -8.69% -13.66% -7.99% -28.10% -11.02% -4.91% -4.85% -4.66% 0.31% -8.44%
LTRN -6.58% -11.69% -7.30% -25.65% -10.42% -4.87% -4.69% -5.01% 3.42% -6.17%
OTRA -6.97% -12.07% -7.41% -26.10% -10.53% -4.85% -4.71% -4.91% 2.52% -5.89%
HLTH -45.95% -37.18% -39.07% -31.65% -21.16% -15.88% -18.41% -4.46% -46.00% -31.26%
EQUP -9.36% -17.51% -10.19% -12.30% -15.70% -12.44% -11.81% -13.56% 1.08% -8.48%
WTER -0.95% -1.99% -2.17% -2.18% -1.57% -1.39% -1.35% -1.42% 2.21% -1.10%
ELEC -1.90% -6.99% -14.77% -10.08% -10.43% -14.60% -13.06% -19.12% 9.16% -8.10%
OFUE -4.33% -6.73% -7.75% -6.84% -4.65% -6.18% -5.96% -6.34% 2.72% -4.04%
OTHG -8.46% -16.54% -9.94% -11.98% -15.48% -12.54% -11.76% -14.15% 2.86% -8.75%
OTHS -5.22% -12.21% -19.78% -16.15% -18.60% -29.66% -24.64% -54.09% 10.24% -8.43%
Total -15.34% -17.20% -20.16% -17.30% -16.27% -17.14% -16.13% -20.39% 0.06% -11.65%

Table 27 confirms the findings from Tables 21 and 24, that the broad-based policy and
narrow-based policy have very similar impacts with respect to equity. Instead, it is the cap level
that has the most significant bearing on the equity of consumption effects. In terms of total
consumption impacts, the change in the Gini coefficient is higher for consumption than total
income or disposable income. The Gini coefficient results for the base case highlight the relative
necessities of each commodity item; where Gini coefficients are smaller – and hence more
equitable – such as Food (FOOD), Gasoline (GASO), Water (WTER) and Electricity (ELEC),
poorer households are consuming similar amounts to wealthier households. Hence the
consumption changes for these commodity groups are particularly salient in equity analysis. That
distributional changes – as measured by the Gini coefficient – are particularly high as a
proportion of the base Gini for the most stringent caps is concerning.

137

Table 27: Gini coefficients for household consumption of commodity groups by income bracket.
Comparison of results for broad-based and narrow-based policy scenarios.

Broad-based policy Narrow-based policy

Emissions reduction (%) Emissions reduction (%)

6.7 9.7 12.7 6.7 9.7 12.7

Base Level Δ Level Δ Level Δ Level Δ Level Δ Level Δ
FOOD 0.236 0.239 0.002 0.245 0.009 0.251 0.015 0.238 0.001 0.244 0.008 0.251 0.015
HOUS 0.355 0.357 0.001 0.368 0.012 0.379 0.023 0.355 -0.001 0.366 0.011 0.379 0.023
GASO 0.243 0.244 0.001 0.257 0.014 0.269 0.026 0.242 -0.001 0.255 0.012 0.269 0.026
LTRN 0.289 0.291 0.002 0.301 0.012 0.311 0.022 0.289 0.001 0.299 0.011 0.311 0.022
OTRA 0.435 0.436 0.001 0.447 0.011 0.458 0.022 0.435 -0.001 0.446 0.010 0.457 0.022
HLTH 0.370 0.372 0.002 0.367 -0.003 0.364 -0.006 0.373 0.003 0.367 -0.003 0.364 -0.006
EQUP 0.456 0.459 0.002 0.469 0.013 0.481 0.025 0.457 0.001 0.468 0.012 0.481 0.024
WTER 0.222 0.223 0.001 0.225 0.003 0.227 0.005 0.223 0.001 0.225 0.003 0.227 0.005
ELEC 0.243 0.248 0.004 0.255 0.012 0.264 0.021 0.246 0.003 0.254 0.011 0.263 0.020
OFUE 0.337 0.339 0.002 0.345 0.008 0.351 0.014 0.338 0.001 0.344 0.007 0.351 0.014
OTHG 0.334 0.337 0.003 0.349 0.015 0.362 0.028 0.335 0.001 0.348 0.013 0.362 0.028
OTHS 0.470 0.472 0.002 0.495 0.025 0.522 0.052 0.469 -0.001 0.493 0.023 0.522 0.052
Total 0.411 0.413 0.002 0.429 0.018 0.446 0.035 0.411 0.000 0.427 0.016 0.446 0.035

CONCLUSION
In sum, these results confirm findings from the literature that ETP would be regressive
without any redistributive element (Burtraw, Sweeney and Walls, 2009; Grainger and Kolstad,
2009; Hassett, Metcalf, and Mathur, 2009; Metcalf et al, 2008; Metcalf, 2009; Oladosu and Rose,
2007; Parry and Williams, 2010; Rausch et al, 2010; Rose, Wei and Prager, 2012). As discussed
above, these analyses tend to focus on either consumption effects or income effects. Like papers
by Parry and Williams (2010) and Rose and Oladosu (2007), this analysis incorporates both
income and consumption effects into the equity analysis. Like the latter,) this analysis integrates
income and consumption analysis within the CGE context.
H5a: Coase Theorem: Allocation schedule will influence distributional impacts (Coase,
1960). As shown above in Chapter 3, the aggregate and sector impacts were influenced by the

138

allocation schedule, though not substantially so. Such similarities have filtered through into the
equity impact analysis. Tables 21 (total income), 24 (disposable income), and 27 (consumption)
all show that there are only subtle differences between the allocation schedules. While these
findings reject Hypothesis 5a in this case, there is nonetheless still the possibility that
distributional differences can arise with varying allocation schedules.
In particular, this analysis highlights the important relationship between incomes and
consumption. In this analysis, the hypothesis proposed in the Introduction - H5b: Income
measures of distribution will be more regressive than consumption measures (Parry & Williams,
2010; Hassett et al, 2009) – was rejected. Total incomes are more inequitable than consumption
patterns in the base case. This is explained mostly by the larger savings rates of higher income
brackets; in contrast lower income brackets borrow more and have higher transfer receipts. Hence
the equity of income distributions are softened when disposable income and consumption patterns
are considered. However, in contrast of hypothesis 5a, the policy impacts to consumption appear
to be greater than impacts to incomes. In aggregate, the impact to income of the most stringent
cap is negative 3.8% (Table 21) while the impact to consumption is negative 11.7% (Table 26).
This distinction also has implications for the equity impacts of ETP. Under the most stringent
cap, the change in the Gini coefficient for total consumption is larger than that for total income.
This highlights a broad trend across each of the measures within this study; i.e. that increasing
negative impacts in the aggregate also translate to increasingly inequitable impacts with respect to
income brackets.
There are mixed results with respect to hypothesis H5c – Relative single-factor measures
of income distribution (e.g. Gini coefficient) will highlight different concerns to measures which
examination changes to single brackets only (e.g. analysis of the relative gains to the lowest or
highest bracket) (Rose, Wei & Prager, 2012). The impacts by income brackets follow a relatively

139

consistent pattern: Impacts to the highest income bracket are small and positives, but impacts
become increasingly negatively as we move down the income bracket spectrum. The inequity of
impacts also increases as the emissions reduction cap becomes more stringent. Consistent results
such as these are captured clearly in the Gini coefficient measure. However, the Gini coefficient
does not reflect the fact that, for the 9.7% and 12.7 emissions reductions simulations, the 15-25k
income bracket is the most negatively impacted. The two lowest income brackets are less
negatively impacted, though still suffer more than the higher income brackets.
Impacts to consumption offer greater support for hypothesis 5b. This is largely because
the consumption element of the USCGE model highlights behavioral responses of households to
policy-induced changes to commodity prices; as such, households at different income levels
prioritize consumption choices between commodities. Changes to consumption for some
commodities reflect changes to the income distribution. For example, Housing consumption is
reduced most by those in the 15-25k income bracket, and lower income brackets are more
negatively impacted on the whole. However, the 15-25k income bracket reduces consumption of
transportation goods (GASO, LTRN, OTRA) to a lesser extent than the neighboring brackets of
10-15k and 25-35k. In sum, detailed analysis of consumption results across income brackets
offers greater insights than such analysis of the income distribution impacts.

140

CHAPTER FOUR:
INTER-REGIONAL EQUITY MODELING

―We have a feeling of anxiety, a feeling of uncertainty because we know that we
will be losing our homes. It is our identity. It is our whole culture at stake…We
are angry. Some people do not understand the science, but they know they are
losing their homes and they are angry they are having to pay for what other
people [in industrialized nations] have done.‖
Ursula Rakova, of the Carteret Islands, Papua New Guinea, July 29 2009
Greg Karras, senior scientist for Communities for a Better Environment, which
has filed a suit to block the cap-and-trade option [of the California AB32 Climate
Bill], called it "institutionalized environmental injustice," adding that it would
encourage "the most entrenched polluters, including oil," to continue emitting
toxics and smog-forming pollutants, which are associated with carbon emissions.
Margot Roosevelt, Los Angeles Times, November 25 2009

These quotes highlight the potential of both climate change and policy responses to stoke regional
inequality. The first quote highlights the potential costs of climate policy inaction; climate change
is disproportionately impacting communities according to their spatial location. For example, the
Carteret Islands, Papua New Guinea are predicted to be submerged by 2015, forcing its residents
to seek refugee status abroad. In this particular case it is unlikely that even substantial GHG
mitigation from the 1990s onwards (when global policy was first pursued) would have impacted

141

the dynamics of sea-level rise that will lead to the Cateret Islands being submerged. Nonetheless,
this is an indication of the stakes involved.
The second quote highlights a complexity of climate policy intervention: co-pollutants.
Industrial processes are not limited to the specific groups of pollutants, such as the GHG
pollutants that climate policy targets. Hence, when industries reduce regulated emissions, any
unregulated co-pollutants will be simultaneously reduced. The concern expressed in the quote is
that ETP allow industrial emissions to increase in some locations as long as the overall emissions
level is reduced. This creates the potential for co-pollutant ―hot-spots‖ to emerge legally (Chinn,
1998; Drury et al, 1998-99). The ―injustice‖ here is that such hot-spots could disproportionately
harm poorer and ethnic-minority communities.
Both quotes highlight the fundamental spatial tension underlying the climate change
problem. Climate change is a global phenomenon caused by pollution across international
regions, which is manifested regional and locally in highly differential impacts that in turn filter
out across the related regional and global economies. As the IPCC (2007) report asserts:
―Vulnerabilities to climate change depend considerably on specific geographic, sectoral and
social contexts…Vulnerabilities of industry, infrastructure, settlements and society to climate
change are generally greater in certain high-risk locations, particularly coastal and riverine areas,
and areas whose economies are closely linked with climate-sensitive resources, such as
agricultural and forest product industries, water demands and tourism; these vulnerabilities tend
to be localized, but are often large and growing.‖
Such spatial tensions have contributed to free-riding in global climate policy endeavors.
The problem is global, yet the sources of pollution are regulated within multitudinous
jurisdictions, both at the national and sub-national level. The Kyoto Protocol of 1997 – the

142

institutional response to the free-riding problem – appeared to have gained initial international
consensus, yet subsequent ratification failure by the U.S. congress removed the world‘s largest
polluter from the regime. The position of Mitt Romney, Republican Presidential candidate for the
2012 U.S. Election, on climate change highlights the reluctance of many U.S. citizens to reduce
emissions when other countries appear to be free-riding:
―The reality is that the problem is called global warming, not America warming.
China long ago passed America as the leading emitter of GHGs.
56
Developed-
world emissions have leveled off while developing-world emissions continue to
grow rapidly, and developing nations have no interest in accepting economic
constraints to change that dynamic. In this context, the primary effect of
unilateral action by the U.S. to impose costs on its own emissions will be to shift
industrial activity overseas to nations whose industrial processes are more
emissions-intensive and less environmentally friendly. That result may make
environmentalists feel better, but it will not better the environment.‖
There are spatial inequalities of climate change within the U.S. also, despite being
difficult to predict accurately. As such, the IPCC (2007) discussions are relatively broad with
respect to regional variation:
―All of North America is very likely to warm during this century, and the annual
mean warming is likely to exceed global mean warming in most areas. In
northern regions, warming is likely to be largest in winter, and in the southwest
USA largest in summer. Annual mean precipitation is very likely to increase in
Canada and the northeast USA, and likely to decrease in the southwest USA.
Snow season length and snow depth are very likely to decrease in most of North
America, except in the northernmost part of Canada.‖
This general language reflects uncertainty in the climate modeling literature
regarding localized results, especially when the results of numerous climate models are
combined (Knutti et al., 2010; Athanassoglou and Massetti, 2012). Such uncertainty
presents challenges for policy makers, with respect to both mitigation efforts and

56
The phrase ―long ago‖ is good rhetoric as 6years is a long time in political years, but a short time with
respect to the beginning of the industrial revolution when Western greenhouse gas emissions first increased
dramatically.

143

adaptation strategies. For example, if there were a high degree of certainty that a
particular U.S. region would be negatively impacted by climate change, then the
incentives for representatives to vote in favor of federal policy, whether to get the country
to reduce emissions collectively – and hence avoid the free-rider problem within the U.S.
– or to allocate funds towards adaptations strategies which would relatively benefit their
region.
This raises a number of important questions with respect to regional equity within the
U.S. Would industries in key states in fact be harmed by these proposals? What are the impacts of
ETP across state economies? This chapter aims to provide insights into these questions by
identifying the economic impacts of ETP across U.S. states. The political implications of these
questions are explored in the following chapter.
Like any major US federal policy, efforts to curb GHG emissions would create winners
and losers. To the extent that burdened industries are concentrated in a particular state or county,
legislators worry that climate policy will disproportionately impact their constituencies. When
applied to large scale pollution problems such as GHGs, ETPs are cheaper than command-and-
control approaches. They would also be more region-equitable if we assume that policy costs are
passed along to US regions evenly. However this assumption is clearly incorrect when we
consider that sectors across US regions pollute at different rates and would face different
regulatory burdens. As the impacts of regulatory burdens ripple throughout the economy, regions
would experience uneven welfare changes with respect to the fortunes of economic sectors within
their boundaries.

144

LITERATURE ON INTER-REGIONAL EQUITY
As with broader debates regarding the trade-offs between economic efficiency and equity
(Okun, 1974), important questions remain as to the equity impacts of ETP.
57
Okun‘s trade-off is
usually interpreted as relating to equity across social classes or races, and has been researched
extensively in recent years (Burtraw, et al., 2009; Hassett et al., 2009; Parry & Williams, 2010;
Rausch, et al., 2010; Oladosu & Rose, 2007; Rose et al., 2012). However, understanding inter-
regional equity is also important in terms of political feasibility and policy design.
Recent work on the inter-regional equity of climate policy impacts has presented
numerous insights (Burtraw and Palmer, 2008; Burtraw, Walls, and Blonz, 2009; Hassett,
Mathur, and Metcalf, 2009; Jorgenson, Slesnick, Wilcoxen, Goettle, and Ho, 2010; Paltsev et al.,
2009; Rausch et al., 2010).
58
In spite of their important contributions, these papers are all limited
with respect to the focus of this paper: the state level economic impacts of climate policy. First,
the majority of these papers explore regional groupings of states as opposed to individual states.
For example, Burtraw and Palmer (2008), Burtraw, Walls, and Blonz (2009), Hassett, et al.,
(2009), Paltsev and colleagues, and Rausch and colleagues (2010) all present results for between
9 and 13 U.S. regions while Jorgenson and colleagues (2010) analyze impacts across four U.S.
regions. Each of these papers is designed to provide more general insights into inter-regional
equity, as opposed to state impacts in particular, and many of the papers are also designed to

57
For example, beyond discussion of the direct economic costs of policy, there have been numerous
criticisms of the potential for ETP to create ―hotspots‖ of pollution concentrated in areas inhabited by
people of color and poorer communities (Chinn, 1999), however, notable recent research has suggested
otherwise (Ringquist, 2011).
58
This paper focuses on the cost impacts of climate policy, and hence reviews similar literature here.
However, it is important to note that the equity impacts of climate policy can also manifest in terms of
climate benefits to some and not others as well as ancillary clean air benefits to some and not others (Boyce
and Pastor, 2010). Fullerton (2008) also highlights the potential impacts of changes to land and house
prices, as well as transitional issues with respect to worker retraining and human capital that can impact
society as a result of major policy changes.

145

provide detailed treatment of the distributional impacts of climate policy within regions.
Moreover, most of these papers use computable general equilibrium models that explicitly
incorporate regional elements, thus providing insights into the substitutions between regions as
climate policy encourages movement towards more emissions-efficient productive activities.
The papers by Burtraw and Palmer (2008) and Burtraw, Walls and Blonz (2009) are the
exceptions here because they use outputs from regional electricity supply and demand projection
models to estimate the consumption effects of climate policy across regions. Two additional
studies examine the state-level impacts of climate policy by focusing on consumption (Boyce and
Riddle, 2009; Stanton and Ackerman, 2009). These latter two analyses present a more detailed
treatment of GHG pricing when compared to the regional electricity modeling approach. These
latter two papers also improve open other inter-regional equity of climate policy literature by
identifying impacts across states.
However, the major limitation with focusing on the direct impacts of GHG intensive
good consumption is that it does not sufficiently account for the general equilibrium effects of
climate policy as they ripple across the economy. Industries regulated by climate policy, such as
electricity generation, are expected to bear the greatest burden of policy implementation, though
some facilities within those industries – e.g. recently built power plants with efficient fuel
switching process – may even benefit from the changes. However, consequential impacts would
ripple across the economy. Prices for downstream customers of regulated industries would likely
increase – for example, all sectors using electricity, as well as households and government
institutions – while upstream customers would experience reductions in demand – e.g. coal
mining and other fuel sources.

146

The ripple effects from an ETP could counter-balance or amplify one another, because of
the centrality of GHG intensive sectors within the macro-economy. Counter-balancing could
occur when, for example, lower demand for raw fuel materials would reduce prices and offset
some of the increase in higher electricity prices to consumers. Amplification could occur when
households paying higher electricity bills could also be impacted by higher costs from goods
produced by industries experiencing higher electricity costs themselves. The computable general
equilibrium modeling approaches of Hassett, Mathur, and Metcalf (2009), Paltsev and colleagues,
Rausch and colleagues (2010), and Jorgenson and colleagues (2010) come to the fore in this
respect. Computable general equilibrium modeling is used to capture the national level impacts of
ETP and combined with a ―regionalization‖ approach (Dixon & Rimmer, 2004; Dixon, Rimmer
& Tsigas, 2007) to disaggregate national level results down to the state level.
HYPOTHESIS
H6: States with GHG-intensive sectors are likely to bear a relatively greater regulatory burden
(Burtraw et al, 2009; Hassett et al, 2009; Pizer et al, 2010).
It would stand to reason that states with the largest proportion of GHG intensive sectors
would also bear the largest regulatory burden from the policy. This would certainly be the case in
terms of the direct regulatory impacts. At the national level, the Electricity Generation, Mining,
and Fuel sectors emit over half of all U.S. industrial carbon emissions, and hence states where
these sectors contribute a large proportion of gross output are likely to bear a larger regulatory
burden, such as Wyoming (36.65%), Alaska (29.78%), Louisiana (24.30%), Oklahoma (15.53%),
and Texas (15.52%), as shown in Table 28. This hypothesis implies that indirect effects are not
addressed; these impacts will be revealed through the CGE modeling. Hence, the CGE results
will provide some level of evidence for whether the null hypothesis can be rejected.

147

Table 28: States with largest percentage of carbon intensive sectors in total state output.
State Mining
Other Fuel
Production
Fossil-
Generated
Electricity
Non-Fossil-
Generated
Electricity Total
WY 30.80 4.36 1.44 0.05 36.65
AK 27.29 1.64 0.69 0.16 29.78
LA 10.87 11.94 1.14 0.35 24.30
OK 11.83 2.02 1.56 0.12 15.53
TX 9.71 4.20 1.40 0.21 15.52
WV 8.86 1.11 1.97 0.03 11.97
NM 10.16 0.47 1.08 0.05 11.77
MT 4.82 4.68 1.18 0.62 11.31
MS 1.39 3.62 1.45 0.40 6.86
ND 3.09 1.17 1.75 0.12 6.12
CO 3.78 0.65 0.80 0.05 5.28
AL 1.78 1.31 1.40 0.59 5.08
KY 2.47 1.26 1.16 0.03 4.92
UT 2.32 1.33 0.88 0.01 4.54
KS 1.44 1.75 0.94 0.28 4.42
AR 1.63 0.88 1.04 0.64 4.19
CA 0.94 2.15 0.59 0.43 4.12
NV 2.08 0.36 1.04 0.12 3.59
AZ 1.50 0.47 0.99 0.42 3.38
IN 0.35 1.66 1.34 0.01 3.35

Figure 8: Electricity generation share of total state output.

148

Figure 9: Direct impacts to Electricity and Fuel sectors only.

REGIONALIZATION MODELING APPROACH
This regionalization approach was developed by Peter Dixon and Maureen Rimmer (2004;
see also Dixon, Rimmer & Tsigas, 2007) to provide state-level results for the USAGE model of
the United States economy.
59
The regional model takes national level results generated by the
USCGE model across a number of factors:
1. Output by sector;
2. Employment by sector;
3. Total household disposable income;
4. Household consumption by commodity;

59
The USAGE model – not to be confused with the USCGE model – is a 500 Sector CGE model of the
U.S. economy developed by Peter Dixon and colleagues at the Monash University Center of Policy Studies.

149

5. Exports by commodity;
6. Government consumption by commodity;
7. Inventory data by sector;
8. Gross domestic product and value added data;
9. Demand for commodities by sector;
10. Investment demand by sector;
11. Make matrix.
These data are combined with numerous state-level data that are sourced from various
government databases, including:
1. Output by sector for each state (referred to elsewhere as the ―Multi-Sector State
Matrix‖). This is sourced from the U.S. Bureau of Economic Analysis Regional
Economic Accounts website (A detailed write-up of the calculation procedure for
this is available upon request). This dataset is central to many of the calculations
as it provides one of the few available sources of the economic links between
states and sectors;
2. Imports and exports data by state, sourced from the USA Trade Online database
provided by the U.S. Census Bureau. This database provides both imports and
exports by state of origin and destination as well as the imports and exports
passing through ports within each state.
3. Commodity Flow Survey data (origin by destination by transport mode by
commodity) provided by the U.S. Census Bureau and the U.S. Bureau of
Transportation Statistics.

150

These factors are combined in a detailed approach that renders results for the following headline
economic indicators across each state:
1. Output by industry;
2. Gross regional product;
3. Employment;
The following state-level economic indicators are also provided:
1. Disposable income;
2. Household demand for commodities;
3. Commodity supply out of each state.
Two important features of the Dixon and Rimmer approach merit attention. First, all
results are reported in percentage change form. Second, the general form of equations used in the
regional model is:
var_R(r) = var + relevant_R(r) – Σ
g
{SHVAR(g)*relevant_R(g)}
The var_R term on the left hand side represents the regionalized variable desired, for example
intermediate demands at the state level, in percentage change form. The var term on the right
hand side is the corresponding national variable. The relevant_R(r) term is a state-level variable
identifying the difference between national var and state-level var_R(r) percentage change. To
continue the example, percentage change in output at the state level. SHVAR(g) is the coefficient
determining the share of each state (g) in the national level of var, for example state shares in
national output.
The Dixon and Rimmer (2005) approach of estimating regional impacts for a particular
variable on the basis of the combination of the equivalent national variable and relevant state

151

variables is similar to the shift-share analysis decomposition approach often used in regional
science. The classical version of the shift-share approach is to decompose a region‘s sectoral
growth into three elements: national effects, industry-mix effects, and regional-shift effects
(Coulson, 1999; Nazara and Hewings, 2004). For example, regional changes to employment are
the combination of national changes in employment (―national effects‖), the national changes in
employment to each sector (―industry-mix effects‖), and the regional specialization in each sector
(―regional shift effects‖).
The Dixon and Rimmer (2005) approach uses this fundamental logic in each of the
equations, however there are important distinctions with the shift-share framework. The first is
that the shift-share approach uses solely empirical data and is attempting to reveal changes that
have occurred to the economy over time. In contrast the Dixon and Rimmer approach is based on
CGE results outputs which aim primarily to establish counter-factual analytical estimates of
hypothetical changes to the economy. There is overlap between the two approaches in this sense,
given that CGE models are based on empirical data, and the Dixon and Rimmer approach also
includes further empirical data at the regional level. While the CGE-based approach has the
benefit of providing additional variable estimates that may not be obtainable empirically, there is
the limitation that errors within the CGE model may inappropriately influence the regional
results. Moreover, while empirical estimates of key national and sectoral variables can be
analyzed in terms any potential statistical error or co-linearity, any CGE results are inherently
inter-dependent due to the structural nature of CGE models.
A second distinction between the Dixon and Rimmer (2005) approach and the basic shift-
share framework is that the former is compiled into a structural equation model that is also
equilibriating. As such, variables within the regional model pass through a series of interactive
iterations before convergence is achieved. A third, related, distinction is that the numbers of

152

variables included in the Dixon and Rimmer approach. While more recent applications of the
shift-share approach have focused on the use of multivariate regression analysis to inform result
estimation – and hence aim to achieve a parsimonious model – the Dixon and Rimmer approach
instead combines numerous variables and data sources at both the national and regional level. For
instance, while recent shift-share analyses have focused on the potential sources of spatial inter-
relation (Nazara and Hewings, 2004), the Dixon and Rimmer approach instead focuses on the
trade relationships between states – as inspired by earlier work on ―SHIN coefficients‖ by
Wassily Leontief and others (1965) – and import and export effects. The former is identified
through analysis of inter-state trade data, while the latter is identified through a combination of
international trade data and CGE model results.
STATE-LEVEL IMPACTS OF U.S. FEDERAL EMISSIONS TRADING POLICY
State-level impacts in terms of output, disposable income, and employment are provided
in Table 29 and depicted in Figures 3-7. Due to the nature of the model, state percent change
results range around the national result rendered by the USCGE model. As shown in Table 29, for
the 6.7% emissions reduction cap, the impact to output at the national level is negative 0.49%,
and state-level results range from positive 0.11% for Kansas (KS) to negative 2.14% for
Wyoming (WY). For the more stringent emissions reductions cap of 9.7%, the national level
output impact is negative 2.97%, with state level results ranging between negative 1.96% for
Oregon and negative 4.49% for West Virginia (WV). Figure 3 depicts the spread of percent
change impacts across states for output in the 6.7% emissions reductions cap level simulation.
This Figure provides a visual example of the distribution of state-level results (from 0.11% to -
2.14%) around the national-level results (-0.49%).

153

Table 29: Percent change impacts to output, disposable income and employment
across cap level simulations, policy regulating Kerry-Lieberman sectors.

Emissions reduction: 6.7% Emissions reduction: 9.7%
State
%Δ to
output
%Δ to
disposable
income
%Δ to
employment
%Δ to
output
%Δ to
disposable
income
%Δ to
employment
AL -0.61 0.28 -1.82

-2.97 -2.06 -2.15
AK -0.99 0.13 -1.96

-3.47 -2.60 -2.69
AZ -1.00 -0.17 -2.26

-3.51 -2.69 -2.78
AR -0.41 0.55 -1.55

-2.98 -1.84 -1.94
CA -0.48 0.47 -1.62

-2.75 -2.25 -2.34
CO -0.68 0.24 -1.86

-3.10 -2.57 -2.66
CT -0.26 1.02 -1.08

-2.88 -1.87 -1.96
DE -0.24 0.99 -1.11

-3.20 -2.20 -2.29
DC -0.18 1.27 -0.83

-2.33 -2.69 -2.78
FL -0.75 0.34 -1.75

-3.70 -2.31 -2.40
GA -0.48 0.57 -1.53

-2.99 -2.26 -2.35
HI -0.73 0.83 -1.27

-3.56 -2.52 -2.62
ID -0.76 -0.08 -2.17

-2.86 -2.57 -2.66
IL -0.46 0.61 -1.49

-2.94 -2.13 -2.23
IN -0.31 0.53 -1.57

-2.23 -1.67 -1.77
IA -0.29 0.84 -1.25

-2.66 -1.91 -2.00
KS 0.11 1.11 -0.98

-2.42 -1.32 -1.41
KY -0.70 0.36 -1.74

-3.17 -1.85 -1.95
LA -0.51 0.41 -1.69

-2.88 -2.25 -2.34
ME -0.45 0.97 -1.13

-3.96 -1.33 -1.42
MD -0.58 0.64 -1.46

-3.39 -2.18 -2.27
MA -0.45 0.95 -1.15

-3.07 -1.75 -1.84
MI -0.44 0.50 -1.60

-2.90 -1.63 -1.72
MN -0.50 0.56 -1.53

-3.22 -1.77 -1.87
MS -0.60 0.49 -1.61

-3.15 -2.04 -2.14
MO -0.43 0.73 -1.36

-3.17 -1.91 -2.01
MT -1.15 -0.12 -2.22

-4.35 -2.48 -2.57
NE -0.39 0.73 -1.37

-3.11 -2.01 -2.10
NV -1.12 -0.02 -2.12

-3.51 -3.76 -3.85
NH -0.54 0.79 -1.30

-3.31 -1.79 -1.88
NJ -0.34 0.75 -1.35

-3.12 -1.88 -1.97
NM -0.94 0.10 -2.00

-3.31 -2.47 -2.57
NY -0.23 1.26 -0.84

-2.93 -1.82 -1.91
NC -0.33 0.59 -1.50

-2.51 -2.09 -2.18
ND -0.76 0.49 -1.61

-3.84 -1.74 -1.83
OH -0.35 0.71 -1.39

-2.93 -1.47 -1.57
OK -0.51 0.64 -1.46

-3.20 -1.76 -1.85
OR -0.60 0.02 -2.08

-1.96 -2.04 -2.13
PA -0.41 0.87 -1.22

-3.24 -1.63 -1.73
RI -0.58 0.98 -1.12

-3.81 -1.79 -1.88
SC -0.59 0.29 -1.81

-2.89 -2.40 -2.49
SD -0.32 0.99 -1.11

-3.42 -1.60 -1.69
TN -0.40 0.79 -1.31

-3.27 -1.54 -1.63
TX -0.55 0.24 -1.86

-2.68 -2.33 -2.42
UT -0.83 0.05 -2.05

-2.96 -2.80 -2.90

154

VT -0.64 0.96 -1.14

-3.69 -1.84 -1.93
VA -0.54 0.44 -1.65

-2.91 -2.41 -2.51
WA -0.51 0.24 -1.86

-2.87 -2.31 -2.40
WV -1.64 -0.53 -2.63

-4.49 -2.52 -2.61
WI -0.40 0.66 -1.44

-2.93 -1.59 -1.68
WY -2.14 -1.94 -4.04 -4.18 -4.98 -5.08
Total -0.49 0.58 -1.51

-2.97 -2.07 -2.16

Figure 10: Distribution of State-Level Output Results (%Δ) from Regional Model
Impacts to disposable income at the national level are positive for the least stringent cap
level (6.7%), such that state level results range between positive 1.26% for New York (NY) and
negative 1.94 for Wyoming. These results stand in contrast to the impacts to employment for the
same cap level. The national level negative 1.51% impact to employment contributes to state
level impacts ranging between negative 0.84% for DC and negative 4.04% for Wyoming. This
picture changes somewhat for the more stringent cap level (9.7%), such that disposable income is
negatively impacted by 2.07%, causing state level impacts to range between negative 1.47% for
Ohio (OH) and negative 4.98% for Wyoming. Similarly, while employment at the national level
-2.5
-2
-1.5
-1
-0.5
0
0.5
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51
% Δ to Output
States

155

is impacted by negative 2.16%, the state results range between negative 1.41% for Kansas and
negative 5.08% for Wyoming.
Clearly some states such as Wyoming and West Virginia are getting a raw deal as a result
of ETP, while states such as Ohio, New York, and Kansas are relatively less impacted. Which
factors are dominating these estimations? In the case of Wyoming and West Virginia, which are
among the most negatively impact state in terms of output, it is unsurprising that impacts to a
mining sector are contributing significantly to the negative output impacts for these states.
However, it is negative impacts to the Other mining (OMIN) sector that are particularly dominant
within the state as a whole. The other significantly negatively impacted sectors in Wyoming are
the Construction (CNSR) and Coal-generated electricity (ELCL) sectors. This is the case for both
the 6.7% and 9.7% emissions reduction cap simulations.
These two particular states are not correlated in terms of impacts across the disposable
income and employment factors. While Wyoming is also amongst the most negatively impacted
states in terms of employment (-5.08% for the 9.7% emissions reduction simulation) and
disposable income (-4.98% for the 9.7% emissions reduction simulation), West Virginia is
instead close to the national results: for the 9.7% emissions reduction simulation, -2.52% for
disposable income (compared to -2.07% at the national level) and -2.61% for employment
(compared to -2.16% at the national). For Wyoming, impacts to the Other Mining (OMIN) and
Construction (CNSR) sectors also dominate the state-sector wedge between national and state-
level employment and disposable income results. While Other Mining is also an influential sector
in determining the wedge between national and West Virginia results for employment and
disposable income, the negative impacts for that sector within West Virginia are offset by
positive impacts to the Medical services (MEDC) sector.

156

At the other end of the spectrum for output, results are more mixed; while Kansas is
positively impacted in the least stringent emissions reduction cap simulation, Oregon is the least
negatively impacted in the 9.7% emissions reduction simulation. For Kansas, positive output
impacts to sectors such as Government utilities (GVUT), Other food manufacturing (MOFD), and
Chemical manufacturing (MCHM), outweigh negative shocks to the Construction (CNSR) and
Other mining (OMIN) sectors. Meanwhile, positive shocks to Oregon‘s Electronics
manufacturing (MSEM), Other durable manufacturing (MODR), and Hydro-powered electricity
generation (ELHY) sectors offset some of the negative impacts to sectors such as Construction,
Medical services, and Retail trade (RTRD).
For disposable income and employment change under the least stringent cap (6.7%
emissions reductions), New York is relatively favorably impacted. For New York, positive
impacts to employment in the Medical services (MEDC), Education (EDUC), and Securities
(SECB) sectors counter-balance the larger negative impacts to sectors such as Other business
services (OBSV) and Construction (CNSR). For the more stringent emissions cap, Ohio is among
the relatively favorably impacted states for both disposable income and employment. While most
Ohio sectors are negatively impacted for these measures, they are mostly smaller than the same
sectors in other states.
These Ohio results are notable because they highlight an important mechanism in the
Regional Model calculations. The state-level wedge that separates state and national results is
what distinguishes states from one another. These state-level wedges are the combination of
changes to national-level sectors and the state-sector mix. In this case, Ohio has a smaller
proportion of sectors with larger negative impacts at the national level – and vice-versa for those
with smaller negative impacts.

157

Table 30: Correlations between Regional Model results, Emissions levels,
and Industry shares across all states.

Emissions reducbtion (%)

9.7 6.7

Output
Disposable
Income

Output
Disposable
Income
9.7
emissions
reduction
Output

1.000

Disposable Income

0.284 1.000

6.7
emissions
reduction
Output

0.671 0.776

1.000

Disposable Income

0.352 0.783

0.893 1.000
Vulcan
emissions
data
Commercial

0.303 0.230

0.287 0.184
Industrial

0.285 0.054

0.110 -0.069
Residential

0.288 0.290

0.340 0.282
Electricity

0.159 0.077

0.008 -0.136
Road transport

0.300 0.137

0.220 0.070
Cement

0.232 0.073

0.137 0.001
Airport

0.201 0.028

0.102 -0.017
Airborne

0.168 -0.001

0.064 -0.028
Non-road transport

0.315 0.173

0.260 0.092
Total

0.288 0.126

0.150 -0.023
Industry
share data
Mining

-0.338 -0.589

-0.660 -0.659
Other fuel production

-0.067 -0.212

-0.197 -0.286
Fossil electricity

-0.330 -0.182

-0.378 -0.312
Non-fossil electricity 0.010 0.215 0.157 0.157
Total

-0.322 -0.555

-0.625 -0.643

Why don't carbon intensive states get harmed more? This question gets to the heart of
the regional model approach. The primary reason is that carbon intensive sectors such as fossil
electricity are a small proportion of state output for most states. As shown in Table 28, only 8
states have more than 10 percent of output from Mining, Fuel, and Electricity generation.
Moreover, Electricity generation in particular is an even smaller proportion (for example, in
Table 28 the state with the highest proportion of fossil electricity output is West Virginia's at only
1.97 percent). This is important because fossil electricity is the most carbon intensive, and hence
the most negatively impacted in the CGE emissions trading policy simulations. Indeed, of the

158

carbon intensive sectors, it is instead Mining that has the most significant impact on regional
results (per the correlation matrix in Table 30). This is probably because states such as Wyoming
and Arkansas have such a large proportion of Mining (31% and 27% respectively) that they
influence overall results.
Taken together, these results suggest that Hypothesis 6 – States with GHG-intensive
sectors are likely to bear a relatively greater regulatory burden – cannot be supported. As shown
in Table 28, those states with the largest carbon-intensive industries are Wyoming, Alaska,
Louisana, Oklahoma, and Texas. In the 9.7% emissions reductions cap simulation, Wyoming is
the 3
rd
most negatively impacted state with respect to output, with Alaska 12
th
, Oklahoma 20
th
,
Louisiana 40
th
, and Texas 45
th
out of 51 states.
60
The disposable income and employment results
are less contradictory of the hypothesis, though the relationship is still not strong with Wyoming
the most negatively impacted state, and Alaska ranked 6
th
, Texas 15
th
, Louisiana 20
th
, and
Oklahoma 40
th
.
More generally, regression results highlight the generally weak relationship between the
proportion of greenhouse-gas intensive industries in a state and the results from this Regional
Model; however, some relationships are stronger than others. For total GHG intensive industry
proportions, there is some level of correlation with both output and disposable
income/employment
61
at the 6.7% emissions reduction level. However, this relationship is weaker
for disposable income/employment and especially output at the 9.7% emission reduction level. In
each of these cases, it is the proportion of Mining sectors which appears to be driving the
correlation. This is highlighted by the similar results between Mining and Total GHG Intensive

60
DC is not an official state, but is treated as one for the purposes of these analyses.
61
The state-level wedges for these factors are derived from the same data in the Regional Model.
Therefore, while the percent change results are different – because the national level results are distinct –
their rankings and relative differences between states are equivalent, causing the regression results to also
be equivalent.

159

sectors, while Other Fuel Production and Fossil-Generated Electricity both show almost no
relationship with output or disposable income/employment for either simulation.
These results reflect the above discussion with respect to the Wyoming and West
Virginia; Other mining (OMIN) within these states plays a dominant role in the results of these
particular simulations and state-level results. These correlation results suggest that Other mining
sector also plays a role in other state results. Either way, these correlation results highlight the
important distinction between direct impact analyses. As shown in Figure 2 above, in a first order
world we would expect states with high proportions of GHG sectors to be relatively burdened by
an ETP. However, when indirect impacts are estimated using more detailed sectoral data and
modeling approaches such as CGE, the results are often distinct. For example, despite the
Electricity generating sectors being negatively impacted across both simulations, these results are
not passed on to the states – as shown by the lack of correlation between fossil-generated
electricity state proportions and the simulation results in Table 30. Even upstream sectors do not
appear to be influencing the results in a clearly dominant way. Instead, the combination of
general equilibrium effects at the national level and unique sectoral mixes within each state
appear to be contributing to a much less definitive assessment of the role that national sector
results would play in state level impacts.

160

Figure 11: Change to state output, Kerry-Lieberman, 9.7% emissions reduction simulation.

Figure 12: Change to disposable income, Kerry-Lieberman, 9.7% emissions reduction simulation.

161

Figure 13: Change to state employment, Kerry-Lieberman, 9.7% emissions reduction simulation.

Figure 14: Change to state output, Kerry-Lieberman, 6.7% emissions reduction simulation.

162

Figure 15: Change to disposable income, Kerry-Lieberman, 6.7% emissions reduction simulation.

Figure 16: Change to state employment, Kerry-Lieberman, 6.7% emissions reduction simulation.

163

CHAPTER FIVE:
THE POLITICS OF CLIMATE POLICY

―I‘m an environmentalist. I want cap and trade. I just want to make sure that the
ratepayers in my state don‘t get socked hard. And that the manufacturing doesn‘t
get crippled.‖
Senator Sherrod Brown (D-Ohio), 2009
―The stark truth is that severe weather events alone will not cause global
warming to pop to the top of the national agenda – let alone revive and
strengthen the push for carbon capping legislation that surely must be one
part of America‘s (and the world‘s) fight against global warming. For that
undertaking to reemerge and triumph, fresh strategies will be needed,
based on new understandings of political obstacles and opportunities.‖
Theda Skocpol, Harvard University, 2013
The previous chapters examined the economics of climate policy in the U.S.; this chapter turns to
the politics. As the first quote highlights, federal representatives such as Senator Sherrod Brown
are concerned about the distributional implications of climate policy. Senator Sherrod Brown, a
Democrat from Ohio is not lying when he says he is an environmentalist. His voting record in
2009 on environmental matters was scored as 91% favorable by the League of Conservation
Voters; he only failed to score 100% due to absence. It is notable therefore that Senator Brown
was questioning whether a policy is sensible on the basis of the distributional impacts, and in
particular the impacts to both consumption and income.

164

Economic impact estimates clearly resonate for politicians other than Senator Brown. In a
2009 interview with The New York Times, President Obama suggested the compromises in the
Waxman-Markey bill – such as protections to the coal industry and industries vulnerable to
international competition – were ―necessary to moderate the different effects of greenhouse-gas
controls on different parts of the country‖ (Broder 2009).
In 2009, major federal climate change legislation was passed in the U.S. House of
Representatives. At the heart of this proposal was an ETP. However, similar proposals stalled in
the Senate. It was suggested at the time that the Democrats who proposed and supported the bills
in both houses had spent all their political capital in enacting federal healthcare policy, and, with
elections looming, were reluctant to push forward unilaterally with climate policy proposals that
potentially harmed industries in key states. There are states with large carbon-intensive industries
– such as Wyoming, Alaska, Louisiana, Oklahoma, and Texas - yet the federal representatives for
these states also tend to vote conservative, and hence against climate policy. In contrast, those
states with a more centrist voting record – i.e. those which are most likely to be persuadable to
vote one way or the other – tend to have relatively small proportion of carbon intensive
industries. Taken together, these results raise important questions about whether or not federal
representatives are reflecting the economic interests of all their constituents, or whether they are
instead voting for a more narrow section of special interests. Are the economic interests of states
represented in the decision making of federal representatives? Or are the economic interests of
particular groups, such as dominant economic sectors within states, being represented?
Moreover, given the increasing partisanship of congress over recent years, to what extent
are economic impacts incorporated into federal voting patterns at all? As Harvard Political
Scientist Theda Skocpol (2013) highlights, numerous political dynamics contributed to the
stalling of federal climate policy between 2008 and 2012. ―To many savvy players, prospects for

165

a legislative push for cap and trade looked excellent during and right after the presidential
campaign of 2008‖ (Skocpol, 2013). The new president, Barack Obama, supported climate
policy, and his Republican opponent in 2008, John McCain, had previously sponsored numerous
carbon-reduction proposals in the Senate. Yet during the summer recess following the passage of
Waxman-Markey, the fiscally-conservative Tea Party protest emerged as a national political
force. ―Telegenic older white protesters carrying homemade signs appeared at normally sleepy
‗town hall‘ sessions to harangue Congressional Democrats who supported health reform as well
as the Waxman-Markey bill‖ (Skocpol, 2013).
The sudden emergence of the Tea Party movement influenced electoral outcomes as well
as a rightwards shift in the policy positions of previously more centrist Republicans. John
McCain‘s previous support for climate policy quickly dissipated. Despite pulling back on Senate
climate proposals, Democrats were still ―shellacked‖ in the 2010 Congressional elections, shifting
the House in particular towards the right, and further diminishing the potential for compromise
over climate policy. By 2012, the Republican Presidential nominee, Mitt Romney, who had
previously supported the implementation of the RGGI while governor of Massachusetts, made
statements which included climate-skeptic rhetoric. As such the Tea Party movement has
influenced political decision making with more conservative candidates being elected, and
centrist candidates shifting their policy positions due to concern and fear of challenges in the
Republican primaries.
It is important to note that beyond the Tea Party there was a strong polarization of public
opinion of global warming between the years 2001 and 2010. McCright and Dunlap (2011) report
that in 2001, 67.1% of liberals surveyed in Gallup polling believed that ―the effects of global
warming have already begun to happen,‖ compared to 73.8% in 2010. In contrast, of
conservatives asked the same question, the shift was from 49.4% in 2001 down to 30.2% in 2010.

166

Both McCright and Dunlap (2011) and Guber (2013) highlight the influence that ―elite cues‖ –
i.e. communications from influential research institutions, bureaucracies, politicians and media
outlets – may have played in this process. ―Political elites on the Left largely promote mainstream
scientific knowledge regarding climate change (as reported, e.g., by the IPCC and the U.S.
National Academy of Sciences), while those on the Right regularly challenge this scientific
knowledge by promoting the views of a handful of contrarian scientists‖ (McCright and Dunlap,
2011).
Hence the role of ideology in public opinion and political voting on climate policy
appears to be dominant and increasing. Does this mean that U.S. Congressional climate policy is
unlikely to be passed in the coming years? Theda Skocpol argues in the quote at the head of the
chapter that even the extreme weather events of 2012 – such as Hurricane Sandy in the U.S.
Northeast, or the ongoing drought in Texas – are unlikely to push climate change to the top of the
national public agenda. Skocpol instead argues for a strategy of center-left politics combined with
―popular mobilization through inter-organizational alliances stretching into state and localities‖
(2013). Skocpol (2013) further argues that a policy design reflective of the 2009 Cantwell-Collins
Senate proposal, whereby government revenues generated from the auctioning of emissions
allowances are redistributed to citizens, possibly within particular states that could tip the balance
in climate policy negotiations.
62
The rationale here is that particular income groups or states
would effectively be compensated for any potential policy costs, a further discussion of which is
provided in previous chapters.
Such a proposal would respond in part to one of the major difficulties with getting large-
scale emissions reduction measures to pass Congress. As Political Scientist Doug Arnold

62
The Cantwell-Collins proposal was also notable because it was sponsored by two moderate Senate
Republicans. As is shown in the analysis below, these two Senators often vote in favor of environmental
protection laws, despite voting more conservatively on other issues.

167

observed in his seminal 1992 work, members of Congress are more likely to support policies
which concentrate benefits and diffuse costs. This is because those gaining can sell their gains to
their constituents, while any broader costs are not likely to significantly impact others. In
contrast, responses to common pool resource problems – such as climate change – diffuse
benefits and concentrate costs. Moreover, as highlighted in the introduction chapter, there is
sufficient uncertainty surrounding the future impacts of climate change that the benefits of
climate policy are also unclear. Doug Arnold (1992) sums up this tension in his main research
question: ―What are the conditions under which a legislative can muster the will to brave the
wrath of noisy constituents and enact rational measures for the broader national good?‖
Hence, underlying any potential future climate policy is likely to be a clear distributional
rationale. Given the importance of state interests in federal representative voting – especially at
the federal level where some states are over-represented with respect to population – it is
important to explore the connections between the economics and politics of federal climate
policy. This chapter relates the national and state-level general equilibrium impacts calculated in
previous chapters to federal political decision making.
LITERATURE ON U.S. CLIMATE POLITICS
Climate change politics have only recently received notable attention in the academic
literature, and sub-national policies are often the focus given the lack of a broad-based U.S.
federal policy (Bromley-Trujillo, 2011; Dierwechter, 2010; Layzer, 2007; Matisoff, 2008; Rabe,
2004; Selin & VanDeveer, 2005; Zahran, Grover, Brody & Vedlitz, 2008). Despite the lack of
enacted federal policies, there have been numerous policy proposals brought before Congress
since the Kyoto Protocol ratification process was debated. As Layzer (2007) documents, this
reflects a broader national debate over climate policy. The emergence of the climate change issue

168

on the U.S. federal political agenda in the 1980s, prompted a ―coalition of fossil-fuel-based
industries to mobilize and devote enormous sums to defusing public concern about climate
change‖ (Layzer, 2007:97). For example, in 1997, the Global Climate Coalition – a group of
more than fifty industry and trade-association representatives of the electricity utility, automotive,
and oil and gas sectors – spent $13 million on television advertising which opposed the Kyoto
ratification. While Democratic President Clinton had just won re-election and favored ratification,
Republicans had also been elected to maintain control of the House and Senate. Senator Chuck
Hagel‘s sentiments at the time reflected the Republican Party‘s mood: ―What we‘re saying is
there will be no implementation of the Kyoto Protocol and no funds expended. Boom. It‘s that
simple‖ (Fialka, 1998).
Following the failure of Kyoto ratification, there was a well-documented increase in
climate policy legislation implemented at the State and Local levels within the U.S. (Rabe, 2004;
Kraft and Kamieniecki, 2007). While federal-level climate policy debates were rare between 1998
and 2002 – partly due to the election of President George W Bush – serious proposals began to
appear from 2003 onwards. The Climate Stewardship Act, sponsored in 2003 by Republican John
McCain and Democrat Joe Lieberman, proposed the U.S.‘s first limit on GHGs. By 2005, even
Senator Hagel appeared to have changed position, introducing three bills that would provide
incentives to U.S. companies exporting or investing in carbon dioxide abatement technologies.
Such a sea-change reflected a broader shift in findings from the scientific community, and more
importantly, shifts in public opinion and among business elites (Layzer, 2007). For example: ―At
an April [2005] conference on climate change convened by the Senate Energy and Natural
Resources Committee, executives from Exelon, Duke Energy, General Electric, Shell Oil, Wal-
Mart, and other companies once again urged Congress to approve a mandatory cap on U.S. GHG

169

emissions‖ (Layzer, 2007). As the narrative at the beginning of this chapter highlighted, such an
interest in GHG reduction policies was short-lived.
Can these dynamics be explained with more general laws of political behavior? It is first
necessary to provide the caveat that, while this chapter focuses on the influence of state-level
economic factors on the policy positions of federal representatives, there are numerous avenues
through which federal policy can be enacted – i.e. federal executive agencies and federal courts –
not to mention the multitudinous state and local level jurisdictions (Kraft and Kamieniecki,
2007). As Political Scientist James Q. Wilson highlights:
―Policy making in Europe is like a prizefight: Two contenders, having earned the
right to enter the ring, square off against each other for a prescribed number of
rounds; when one fighter knocks the other one out, he is declared the winner and
the fight is over. Policy making in the United States is more like a barroom
brawl: Anybody can join, the combatants fight all comers and sometimes change
sides, no referee is in charge, and the fight lasts not for a fixed number of rounds
but indefinitely or until everybody drops from exhaustion… ‗it‘s never over‘‖
(1989:299).
Such complexity makes analysis of general trends difficult (Kraft and Kamieniecki,
2007), yet no less fertile within particular boundaries.
Focusing on Congressional policy debates, as Political Scientist John Kingdon asserts,
―the primary determinant of voting are generally considered to be party, ideology, and
constituency‖ (Anderson, 2011). In terms of ideology, the environmental politics literature has
highlighted the growing partisan nature of congressional voting on environmental issues (Shipan
& Lowry, 2001) which reflects a broader trend of U.S. federal politics (Crowe & Eberspacher,
1998; Groseclose, Levitt & Snyder, 1999; Snyder & Groseclose, 2000; Tanger, 2012; Theriault,
2006). Tanger (2012) argues that this polarization is largely attributable to a shift by Democrat

170

representatives towards pro-environmental positions.
63
In contrast, Tanger‘s (2012) adjusted
environmental voting index suggests that Republican support for pro-environment policies has
not shifted dramatically between 1970 and 2008, a finding reflected elsewhere (Brewer, Mariani,
& Stonecash, 2002).
One explanation for the cause of this increased partisanship
64
is the shift away from the
1970s approach to energy policy that focused on specific energy types, and witnessed the ―low
conflict and mutual non-interference‖ which characterizes ―distributive policies‖ (Lowry, 2008).
In its place, a ―regulatory‖ approach has come to dominate, one which is more in line with
environmental policies. This approach is more comprehensive and requires changes to traditional
behavior, and as such is likely to create ―much less stable and more conflictive‖ politics (Lowry,
2008).
Whether this partisanship is a reflection of constituency influence or core ideological
belief is a matter of debate (Bishin, 2000). This debate has relevance for the question of to what
extent the distributional economic impacts of policies influence congressional voting. If
constituency influence dominates, distributional economic impacts are important to the extent that
constituent groups are relatively harmed or benefitted by the policy. If ideological belief
dominates, then distributional economic impacts are important to the extent that representative
ideologies value relative economic impacts. For example, an ideology which promotes labor
unions and low-paid workers would likely oppose a policy which caused distributional harms to
low-income households.

63
Pro-environmental positions are those determined by the League of Conservation Voters.
64
Further explanation for the shift is provided by the Advocacy Coalition Framework (Sabatier & Weible,
2007).

171

The ideology hypothesis certainly goes some way to explaining numerous climate policy
outcomes over the past decade, and is a long-established condition in the political science
literature (Aldrich 1995; Binder, Lawrence, and Maltzman 1999; Bishin 2000, 2009; Bond and
Fleisher 1990; Cox and McCubbins 2005; Gillion 2012; Grose 2005; Grose and Middlemass
2010; McElroy and Benoit 2007; Miller and Stokes 1963; Nokken 2000; Rohde 1991; Weller
2009). However, recent studies have found that constituent interests also play a significantly
influential role. Anderson (2011) demonstrates that a greater concentration of environmental
group membership within congressional districts is a significant factor influencing Congressional
roll-call voting on environmental policy proposals. Chupp (2011) introduces an explicit policy
variable – Haiku electricity sector model
65
estimates for the impact of climate policy on
electricity prices – into the analysis of the influence of constituency interest on federal level
policy voting. Chupp (2011) thus extends the perspective provided by Anderson (2011) and finds
that both ideology and constituent economic interest are significant indicators of Congressional
roll-call voting on environmental policy issues.
If we broaden our focus briefly, the issue of distributional impacts across states has been
raised in a number of studies, although there has been limited treatment of this factors relative
importance to climate politics. In one exception, Zahran and colleagues (2008) explore
metropolitan-level climate policy making across the U.S., finding that metro areas with greater
climate impacts are less likely to adopt climate policies, while those metro areas with greater
environmental concern are more likely to do so. Regarding the benefits of climate policy, ―the
expected ecological, social, and economic impacts of climate change are geographically uneven‖
(Zahran et al., 2008). As such, regional governments would be rationally motivated to enact

65
Employed in numerous studies by scholars at the Resources for the Future (e.g. Burtraw, Walls, and
Blonz, 2009), and referenced in previous chapters.

172

climate change policies to the extent that the risk of climate change could be reduced (Goulder,
2003; Victor, 2003).
These regional results are mirrored in federal environmental policy debates. Shipan and
Lowry (2001) show that environmental causes have gained least support from federal
representatives in the South. In light of the data from Pizer et al. (2010), which shows that
carbon-intensive energy consumption is greatest in the South, these results are similar to the
Zahran et al. (2008) findings above. In Shipan and Lowry‘s (2001) findings, Southern Democrats
are less likely to support environmental policy than their party colleagues from other regions.
This suggests that in this case the constituency influence overrides the ideological aspect of
partisan politics. Cragg, Zhou, Gurney and Kahn (2012) extend this analysis and apply it directly
to the federal-level climate policy debates of 2009. They confirm the notion that ―conservative,
poor areas have higher per-capita carbon emissions than liberal, richer areas.‖ They also find that
representatives of emissions-intensive constituencies are also less likely to vote for anti-carbon
legislation, suggesting that there is a clear economic interest in federal climate policy voting.
However, as with the other studies reviewed above, Cragg and colleague‘s (2012) focus on
consumption effects may neglect the other dimensions of climate policy economic impacts, such
as the income effects studied here.

173

Figure 17: U.S. Senate Average Liberal Index across States, 2009

Figure 18: U.S. House of Representatives Average Liberal Index across States, 2009.

174

HYPOTHESES
H7a: States with relatively greater economic harms from climate policy are less likely to have
federal representatives supporting climate policy and ETP (Pizer et al., 2010; Shipan & Lowry,
2001; Zahran et al., 2008).
H7b: States with relative economic harms from climate change are more likely to have federal
representatives supporting climate policy and ETP (Goulder, 2003; Victor, 2003).
H7c: States with relatively large harms to influential GHG-intensive sectors are less likely to
have federal representatives supporting climate policy and ETP.
H7d: Direct regulatory impacts are likely to have greater influence than overall (direct plus
indirect) economic impacts on state level policy making and federal level representation of state
interests.
As shown by Cragg and colleagues (2012), there does appear to be a relationship between
GHG consumption rates and public opinions on climate policy. Assuming that politicians
represent their constituents‘ opinions in Congressional roll-call voting, we would expect for
Hypothesis 2a to be supported. After controlling for other factors such as ideology and partisan
politics, it is expected that states with relatively greater regional harms from policy are less likely
to engage in policy making, whether this is in the form of active state EPAs or equivalent, strong
climate action plans or federal representatives supporting climate policy and ETP. On the other
hand, there is evidence to suggest that the opposite may hold true. For example, Figure 3 focuses
on the Electricity Generation sector, highlighting the small proportion of output in each state
(between 0 and 2%), despite that sector contributing to nearly half of all industrial emissions.
This is notable because those states with a higher likelihood of voting against carbon reduction
policies are also those with lower proportions of Electricity Generation.

175

It is not expected that aggregate state economic impacts would dictate policy making.
Economic impacts are likely to influence policy making in a number of potentially contradictory
ways. If influential sectors within given states were likely to be negatively impacted by a federal
ETP, it is likely that these sectors would seek to influence the political process. Even though the
state as a whole may not be substantially negatively impacted, influential industries oppose ETP.
Direct regulatory burdens are also likely to play a more salient role in policy making at this level
than indirect costs. Although less accurate, direct impacts are more parsimonious and hence can
carry with them more persuasive weight to policy makers.
It is also expected that the states which benefit from climate policy – i.e. states which
would benefit relatively from the reductions to the harms resulting from climate change and
associated co-pollutants – would also be more likely to engage in state-level climate policy
making and support federal level regulation. In other words, state-level policy makers and federal
representatives would be rationally motivated to enact climate change policies to the extent that
the risk of climate change could be reduced for their state (Goulder, 2003; Victor, 2003) and
hence experience ―concentrated‖ benefits.
POLITICAL ECONOMY ANALYSIS
To examine these questions and hypotheses, this chapter undertakes a number of steps.
First, results from Chapter 5 regarding the economic impact across states of a U.S. federal ETP
are compared with the voting records of state representatives in Congress. In particular, this step
will explore the question of why, in 2009, climate policy was able to pass the House, yet stalled
in the Senate. While this question is narrow in scope, it provides useful insights into the broader
questions stimulating this analysis. Second, probit regressions analyze the influence of
partisanship, voting on environmental policy questions, and economic considerations on

176

Congressional roll-call voting on climate policy bills between the years 2007 and 2011. While
previous analyses by Anderson (2011) and Chupp (2011) have provided insightful analyses of
Congressional roll-call voting for environmental policy as a whole, there have been no analyses
of climate policy specifically.
On important question underlying the studies which explore the influence of economic
factors on political voting is: Which are the appropriate economic indicators to use? Beyond more
general indicators of economic well-being in a state – i.e. the average income – policy-relevant
economic indicators are often difficult to obtain. Cragg and colleagues (2012) use district level
emissions rates from the Vulcan project as a proxy for emissions consumption (Gurney et al,
2005). On the industry side, Anderson (2011) uses the proportions of mining and manufacturing
employment as proxies for the reliance of states on particular industries. The implication for the
indicators used by Cragg and colleagues (2012) and Anderson (2011) is that larger emissions
consumption rates or larger GHG-intensive industry rates are likely to leave that state vulnerable
to environmental or climate policy impacts. These indicators therefore only implicitly account for
policy impacts, as low-cost substitution possibilities may mean that the actual impact on states is
minimal. With this in mind, Chupp (2011) uses the HAIKU electricity price model to explicitly
incorporate policy impact estimates into the analysis. This dissertation takes a similar approach
by including the state-level general equilibrium results from the regionalized-USCGE model.
Nonetheless, the state-industry-share and emissions consumption data are also incorporated
because these are also indicators of a more direct-impact perspective on climate policy. That is to
say that state publics and representatives may well be basing their judgments on data such as
these, as opposed to the estimated policy impacts, which incorporate more complexity into the
analysis.

177

Given the substantial influence of ideology and partisanship in Congressional voting –
which, as shown in the literature review above, has increased over the past decades – the third
area of analysis in this chapter is to identify which federal representatives have voted against their
party line on climate policy bills. This focus on those representatives who ―break rank‖ is inspired
by the logic of regression-discontinuity analysis (Angrist and Pischke, 2008; Lee, 2008), whereby
variations at the margins of a natural break or threshold in a distribution provide pseudo-natural
experiment conditions. While the analysis in this chapter does not claim to emulate such
conditions, the exploration of marginal voting patterns is nonetheless instructive. Finally, this
chapter takes a broader perspective still by identifying four key trends in climate policy.
As was observed in Chapter Four, there is not a strong relationship between the
proportion of greenhouse intensive industries within a state and the regionalized CGE model
results. Specifically while there is a weak relationship between the proportion of Mining within a
state and the Regionalized-CGE model results, there is no relationship between either Other Fuel
Production or Fossil-Generated Electricity and Regionalized-CGE model results. This is
unsurprising to the extent that, while Electricity sectors are the dominant source of industrial
emissions – and hence are the focus of climate policy regulations – they represent a small
proportion the economic output in most states. Hence, in the Regionalize-CGE context, impacts
to larger downstream and upstream sectors are likely to play a greater role in state-level results.
There is also no statistical relationship between the proportion of carbon intensive sectors
and the Senate Liberal Index.
66
This is indicated in Table 31, whereby by states with the largest
proportion of carbon-intensive sectors are not predominantly conservative in their voting records.

66
The Liberal Index is produced by the National Journal and aims to rank federal representatives with
respect to their voting across all bills. The Liberal-Conservative nature of voting is determined with respect
to other voters in their party. For example, if a Democrat representative always votes in line with their party
colleagues, they are deemed 100% Liberal; any contrasting voting and they become less Liberal, such that
those at the margins (i.e. around 50%) are the least likely to vote in-step with their party colleagues.

178

While the four of the top five states in Table 31 tended to vote conservatively in 2009, four of the
next five states tended to vote liberally in 2009. Moreover, regression results for the influence of
carbon-intensive sectors on the average Senate Liberal Index for states finds no relationship
between the two, for any of the sub categories of carbon-intensive sectors presented in Table 31.
Table 31: States with largest percent of carbon intensive sectors in total output
State Mining
Other Fuel
Production
Fossil-
Generated
Electricity
Non-
Fossil-
Generated
Electricity Total
Senate
Liberal Index
(Rank)
WY 30.80 4.36 1.44 0.05 36.65 13.6 (46)
AK 27.29 1.64 0.69 0.16 29.78 50.6 (27)
LA 10.87 11.94 1.14 0.35 24.30 38.2 (35)
OK 11.83 2.02 1.56 0.12 15.53 5.6 (50)
TX 9.71 4.20 1.40 0.21 15.52 19.9 (41)
WV 8.86 1.11 1.97 0.03 11.97 67.7 (16)
NM 10.16 0.47 1.08 0.05 11.77 74.9 (12)
MT 4.82 4.68 1.18 0.62 11.31 61.1 (19)
MS 1.39 3.62 1.45 0.40 6.86 25.2 (40)
ND 3.09 1.17 1.75 0.12 6.12 57.0 (21)

Table 32: Mining, fuel and electricity sector proportions of gross output,
for states with centrist voting records, 2009*

Liberal voting
index

State-level impacts
Mining, fuel and
electricity
State Senate House
%Δ to
output** Rank
Proportion
of GSP Rank
WI 58.8 59.5 -2.92 35 1.78 41
ND 57.0 60.3 -3.84 5 6.12 10
NH 56.3 61.8 -3.31 16 1.83 38
VA 56.3 43.7 -2.91 37 1.95 37
IA 54.8 52.6 -2.66 46 1.66 43
MN 54.8 51.4 -3.22 19 2.44 26
AR 51.3 45.4 -2.98 30 3.38 19
AK 50.6 26.2 -3.47 12 29.78 2
PA 49.6 51.4 -3.42 18 3.26 21
FL 48.0 42.8 -3.70 7 2.01 35
MO 46.0 43.8 -3.17 23 1.78 40
NV 44.1 50.9 -3.51 11 3.59 18
IN 40.7 41.2 -2.23 50 3.35 20
ME 40.4 69.5 -3.96 4 1.97 36
SD 40.1 46.8 -3.42 13 1.79 39
Average
(mean) 50.4 47.3

-2.97 21.7

5.37 28.3
*States with centrist voting records are here defined as those with Liberal voting index figures
between 40 and 60 on the 100 point index for the Senate. ** For 9.7% emission reduction
simulation.

179

This finding certainly supports the case for optimism on the part of climate policy
proponents in the run up to the 2009 session that was described above. Such optimism was further
supported by the idea that a climate policy (the Waxman-Markey bill) had made it through a more
conservative-voting house (see Table 32). As also shown in Table 32, those state Senate
representatives with average centrist-voting – who are likely to have an important bearing on roll-
call outcomes – are not likely to have a large proportion of carbon-intensive sectors within their
constituency economy. This is shown by these states generally having low proportions of carbon
intensive sectors in their economy. Alaska (AK) is an obvious outlier in Table 32. Despite having
an average Senate voting index around the mean, the House average voting is far more
conservative, which possibly reflects the greater proportion of the economy in carbon intensive
industries. This last statement of course assumes the Cragg and Kahn (2009) findings regarding
the relationships between carbon consumption rates, wealth, and ideology discussed above are
also present with respect to the carbon intensity of industry.
Instead, the findings in this chapter so far suggest otherwise. In other words, there does
not appear to be a strong economic rationale for the Senate allowing the climate policy proposals
to stall in 2009. As such, it does appear that a major exogenous change is responsible for the shift
in the Senate agenda, adding further weight to Skocpol‘s assertion that the Tea Party protests in
the summer of 2009 played an important role in changing the dynamics of climate policy.
There certainly does appear to have been an increasing partisanship in federal voting on
environmental policy issues between 1970 and 2008 (Tanger, 2012), but has this pattern
continued since 2008? Table 33 suggests that, while there are indications of increasing
partisanship in the Senate, the relationship between voting on Environment bills and the Liberal
Index has remained relatively stable in the House between 2007 and 2011. There are some
important caveats to this finding. First is the definition of an environmental bill. As is common in

180

the Environmental Politics literature (Anderson, 2011; Chupp, 2011; Tanger, 2012), the League
of Conservation Voters‘ definitions of Congressional environmental bill are used here. One
limitation of comparing different years is that they may vary with respect to environmental issue
and policy design. Thus is it possible that variation between the years for the Senate and House,
despite being larger in the Senate, might be explained by the variation in bills proposed.
Table 33: Correlation matrix for probit regression data.

RGGI CC bill
Break
Rank Party Chamber
Lib
index ENV Mine
Other
fuel
Fossil
elec
Output
9.7
Dispos
9.7
Total
cnsm
CC bill 0.321 1.000
Break
Rank 0.038 -0.015 1.000

Party 0.248 0.865 0.004 1.000

Chamber 0.012 0.027 -0.003 0.020 1.000
Lib
index 0.341 0.848 -0.031 0.886 0.000 1.000

ENV 0.339 0.888 0.071 0.911 0.047 0.900 1.000

Mine -0.206 -0.197 -0.013 -0.155 0.085 -0.215 -0.206 1.000
Other
fuel -0.217 -0.154 0.005 -0.124 -0.008 -0.165 -0.163 0.605 1.000
Fossil
elec -0.276 -0.242 0.004 -0.180 0.046 -0.254 -0.258 0.484 0.314 1.000
Output
9.7 -0.142 0.010 -0.037 0.015 -0.132 -0.003 -0.009 -0.084 0.169 -0.236 1.000
Dispose
9.7 0.230 0.083 0.052 0.073 -0.066 0.101 0.098 -0.439 -0.205 -0.094 0.195 1.000
Total
cnsm -0.295 -0.100 -0.057 -0.082 -0.220 -0.092 -0.120 0.291 0.436 0.266 0.319 -0.071 1.000
Elec
cnsm -0.339 -0.218 -0.032 -0.185 -0.143 -0.244 -0.239 0.378 0.312 0.577 0.135 -0.012 0.786
RGGI = State participation in RGGI program; CC bill = climate change bill; Break rank = indicator for when a member of Congress
voted against the party line; Lib index = Liberal index; ENV = voting on non-climate environmental bills; Mine = mining share of
state output; Other fuel = fuel production share (non-mining) share of state output; Fossil elec = fossil generated electricity share of
state output; Output 97 = state output results under 9.7% emissions reduction simulation; Dispos 9.7 = state disposable income results
under 9.7% emissions reduction simulation; Total cnsm = state total emissions consumption rates; Elec cnsm = state electricity
generation emissions consumption rates.
A second caveat is that the Liberal Index is a composite of all bills voted on by
representatives, and does not exclude environmental bills. The extra level of dependency between
the variables may bias the results. That said, environmental bills represent a relatively small
proportion of total bills voted upon. During the period studied, the number of environmental bills
is as follows: 20 in the House, 15 in the Senate (2007); 13 in the House, 11 in the Senate (2008);

181

14 in the House, 11 in the Senate (2009); 10 in the House, 7 in the Senate (2010); 35 in the
House, 11 in the Senate (2011). This compares with an average of 7,021 bills per year in 2007-08,
6,841 bills per year in 2009-10, and 6,149 bills per year in 2011-12.
Voting on climate policy also appears to be strongly influenced by partisan interests,
though there is not strong evidence in Appendix G for a shift towards increasing partisanship on
climate policy voting. Tables in Appendix G show the relationship between the Senate and House
voting on bills with an explicit climate change component and climate-change-related content.
This categorization was performed by the author. A bill was determined to have an explicit
climate change component if that bill had reference to climate change or global warming in the
title (as identified either by Congress records or the League of Conservation Voters), or the
description of the bill featured explicit and prominent reference to climate change or global
warming. A bill was determined to have climate-change-related content if the purpose of the bill
was to target an environmental issue with a strong connection to climate policy, such as
renewable energy. This category is more arbitrary than the explicit-content category, as many
environmental bills have at least some connection to the climate change issue, even if indirectly.
PROBIT REGRESSION RESULTS
Probit regressions are run to assess the influence of party, the state‘s participation in the
Regional Greenhouse Gas Initiative (RGGI), ideology (the Liberal index), voting on
environmental bills, and policy-relevant economic indicators, on a pooled set of climate bills in
both houses of Congress between the years 2007 and 2011. The results in Table 34 show that the
dominant factors are the liberal index and voting on non-climate environmental bill. It is notable
that the state‘s participation in RGGI is significant. This could be viewed as those states voting in
favor of a policy which they have already invested in. By enacting a federal policy, the freeriders

182

elsewhere would be brought under similar controls already enacted under RGGI, hence reducing
the relative burden on those states. However, the RGGI covers 10 Northeast states, which are
often characterized as liberal compared to their southern counterparts in both parties. As such,
this dummy variable may be instead expressing a further nuance to the ideology indicator.
Table 34: The influence of economic factors on Congressional roll-call climate bill voting, 07-11.
Voting on Climate Policy Bills
Carbon
Industry
Share
CGE State
Output
Consumption
Party 0.289 0.192 0.222
(1.56) (1.06) (1.22)
State in RGGI 0.443** 0.578*** 0.469**
(2.60) (3.42) (2.69)
Liberal Index 4.038*** 4.100*** 4.039***
(8.02) (8.21) (7.99)
Voting on Non-Climate Environmental Bills 2.540*** 2.625*** 2.640***
(10.04) (10.47) (10.49)
State Mining Share -0.020
(-0.92)
State Other Fuel Production Share -0.004
(-0.11)
State Fossil Electricity Generation Share -0.254
(-1.49)
State Output Impact, 9.7% Emissions Reduction 0.134
(1.05)
State Income Impact, 9.7% Emissions Reduction -0.137
(-0.99)
State Total Emissions Consumption 0.005*
(1.76)
State Electricity Emissions Consumption -0.014
(-2.04)
Intercept -3.337 -3.624***
(-13.44) (-17.01)
N 2,374 2,374
Pseudo R-sq 0.786 0.785
z statistics in parentheses
* p<0.1, ** p<0.01, *** p<0.001

It is also notable that the economic indicators are generally insignificant. This is with
respect to all three types of economic indicators used here – industry shares, regionalized-CGE
results, and emissions consumption. The one exception is the total emissions consumption rates,

183

which are weakly significant. However, the positive coefficient here very much conflicts with the
intuition that lower emissions rates would imply a less vulnerability to climate policy impacts,
and hence a greater likelihood to vote in favor of climate policy. That said, it is important to note
that these regressions include voting in both the Senate and House. While the former are state
representatives, the latter are district representatives, and hence state-level economic indicators
are less likely to be of relevance, especially in large states such as California. Ultimately this
could lead to omitted variable bias within the analysis, however, district level emissions data are
available from the Vulcan project (Gurney et al, 2005) for more in-depth analysis of House voting
and the influence of emissions consumption.
BREAKING RANK RESULTS
The significant influence of ideology on climate policy voting in these results, and in on
environmental policy more generally in the literature, raises the question: Who is ―breaking rank‖
with their party whip when voting on climate policy issues? In terms of party, neither is
significantly more likely to break rank. As shown in Table 35, while larger numbers of
Republicans did so in both chambers in 2007 and 2008 – for both ―explicit‖ and ―related‖ climate
bills, the trend shifted towards more Democrats breaking rank in the Senate in years 2009-11, and
a mixed picture for the House during the same period.
Table 35: Congressional representatives ―breaking rank‖
on climate bill roll-call voting, 07-11.

Explicit Related

Senate House Senate House
Year D R D R D R D R
2011 4 1 29 5 13 33 5 10
2010 0 0 0 0 30 16 5 6
2009 4 2 43 8 0 2 6 0
2008 4 7 0 0 5 21 29 62
2007 14 35 3 55 0 17 54 72

184

Table 36: The influence of economic factors on Congressional representatives ―breaking rank‖ on
climate bill roll-call voting, 07-11.
Breaking Rank on
Climate Policy Bill
Voting

Democrat Republican

Total Senate House Total Senate House
State in RGGI -0.513 -0.568 0.458* -0.193 0.438*
(-1.48) (-1.56) (2.15) (-0.20) (1.93)
Liberal Index -6.356*** -0.529 -7.056*** 1.534* 3.889 1.386*
(-8.10) (-0.25) (-7.93) (2.14) (0.78) (1.88)
Voting on Non-Climate
Environmental Bills
-1.282*** -6.142** -1.056** 3.295*** 7.718* 3.209***
(-3.36) (-2.97) (-2.66) (9.86) (2.01) (9.30)
State Mining Share -0.000 -0.058 0.012 -0.174* -0.140 -0.195*
(0.00) (-0.50) (0.33) (-1.82) (-0.25) (-1.88)
State Other Fuel
Production Share
0.040 0.166 -0.002 0.150 -5.421 0.178*
(0.81) (1.30) (-0.03) (1.51) (-1.72) (1.69)
State Fossil Electricity
Generation Share
0.320 0.802 0.244 -0.106 1.160 -0.067
(1.39) (1.07) (0.94) (-0.36) (0.74) (-0.21)
Intercept 3.341*** 3.392* 3.687*** -2.940*** -2.870 -2.92***
(6.36) (1.95) (6.21) (-7.95) (1.976) (-7.44)
N 1,247 113 1093 1,188 135 1053
adj. R-sq 0.373 0.405 0.385 0.467 0.786 0.442
t statistics in parentheses
* p<0.1, ** p<0.01, *** p<0.001

It is notable in Table 36 – which presents results for the ―break rank‖ probit regressions –
that Democrats voting against climate policy are likely to be centrist with respect to voting on
non-climate environmental bill and voting on bills in general (i.e. the liberal index). This is also
the case for House Democrats, though for Senate Democrats only the former indicator is
significant. Again, as with voting on climate bills, the economic factors are again insignificant. It
is perhaps surprising that economic factors are not significant for Democrats breaking rank on
climate policy voting. For example, Democrat Senator Mary Landrieu of Louisiana consistently
votes against both climate and environmental bills, and her state contains a large oil industry.
Clearly for Landrieu and other Democrats breaking rank, it is centrist voting records more
generally, as opposed to economic factors distinguishing the states they represent, which
predominantly influences climate bill voting.

185

On the Republican side, voting on non-climate environmental bills is significantly
correlated with breaking rank for Republicans in both chambers. Interestingly, for Senate
Republicans, as with their Democrat colleagues, general voting (i.e. the liberal index) is not
significantly correlated with breaking rank on climate policy voting. Instead, whether or not the
representative‘s state participates in RGGI is significant. As with the general climate bill voting
analyzed above, the RGGI dummy may instead reflect a more liberal stance of northeastern
Republicans compared to their southern colleagues. For example, Republican Senator Susan
Collins of Maine – which participates in RGGI – often votes pro-climate and pro-environment,
and has a relatively centrist voting record. Of the economic indicators, the Mining sector share
and Other fuel production share are significant for the House regression. While the Mining
production sector share result follows intuition – that is, representatives from states with a larger
sector are less likely to break rank – the results for Other fuel production are counter-intuitive.
RESULTS SUMMARY
In sum, results here suggest that hypothesis 7a – States with relatively greater regional
harms from policy are less likely to have federal representatives supporting climate policy and
ETP (Pizer et al., 2010; Shipan & Lowry, 2001; Zahran et al., 2008) – is not supported. Voting on
climate bills instead appears to be dominated by ideology. While it is possibly that ideology is
instead turn influenced by constituent interests, economic or otherwise, these results suggest that
link is not strong. There is no statistical relationship between the proportion of carbon intensive
sectors or Regionalized-CGE results and the Liberal Index score of representatives. As such, the
data also do not support hypothesis 7c: States with relatively large harms to influential GHG-
intensive sectors are less likely to have federal representatives supporting climate policy and
ETP. Similarly, hypothesis 7d – Direct regulatory impacts are likely to have greater influence
than overall (direct plus indirect) economic impacts on state level policy making and federal level

186

representation of state interests – is also unsupported by the results. This is not to say that such
forces do not play some role in the decision making of federal roll call voting. Instead, the
analyses here suggest that ideology is dominant and that neither the carbon intensity of state
industries (i.e. direct impacts) nor the regionalize-CGE results (i.e. indirect impacts) are strongly
linked to either ideology or climate bill voting.
If we look at the policy benefits side of the ledger, while hypothesis 7b – States with
relative regional harms from climate change are more likely to have federal representatives
supporting climate policy and ETP (Goulder, 2003; Victor, 2003) – makes sense in theory, the
high degree of uncertainty surrounding climate change impacts to U.S. regions makes such
analysis difficult. It is notable that representatives from particular states are more inclined to
―break rank‖ from their party whip and vote on climate bills. Interestingly, this does not appear to
be strongly linked to breaking rank on environmental policy more general. As such, we cannot
rule out other explanations as to why, for instance, Republicans from states such as Maine and
New Jersey have often broken rank on climate bill voting. Both states were members of the
RGGI, so experience of state level policy – and hence regulated state entities having a first-mover
advantage – may be a strong explanation with respect to benefits. However, the possible threat of
climate change – and the extent to which those representatives believe the climate science –
cannot be ruled out either.
TRENDS IN CLIMATE POLICY WORLDWIDE
To develop a ground-up analysis of the role of knowledge in climate change policy,
ideally one would identify policy decisions and work backwards to identify explanations and
theoretical implications. However there have been a large number of policy decisions in this area
over the past three decades. Discussion of climate change policy tends to focus on mitigation

187

efforts; such policies in five selected OECD countries are presented in Table 40. Yet there are a
number of different policy outcomes. As the Congressional Budget Office highlights (Table 41),
the U.S., which has no explicit GHG mitigation policy, nonetheless still spend budget dollars in
the following areas of climate change policy:
Technology programs that develop, demonstrate, and deploy new products or
processes to reduce GHG emissions;
Scientific research directed toward explaining the processes of climate change and
monitoring the global climate;
Assistance to other countries as they work to reduce GHG emissions; and
Tax incentives that encourage businesses and households to adopt technologies
that curtail the use of fossil fuels and reduce GHG emissions. (Webre, 2010)
67

In the interests of parsimony, instead of discussing each possible climate change policy,
this paper presents a number of key trends that highlight the influence of knowledge on climate
change policy decisions. Trends allow for temporal elements to be incorporated into the analysis,
and it is important to consider what is changing and what is constant over time. Trends also allow
for inter-temporal influences to be identified. This is particularly important when considering the
role of knowledge in policy making, as knowledge is constantly evolving as a dialectic between
the cognitive pursuits of analysts and the social interactions of the policy process.
AN INCREASING NUMBER OF CLIMATE CHANGE POLICIES
As shown in Table 39, there have been an increasing number of climate change policies
worldwide since the 1980s. This trend began with the gradual development of climate science in
academia and government institutions, areas dominated by the cognitive elements of Wildavsky‘s

67
The figures for these U.S. climate change policy budgetary expenditures are presented in Table V below.
As is indicated, while funding for these policies was relatively stable between 1998 and 2008, technology
programs and tax incentives both received increases in funding during that period. The most dramatic
change has come with the American Recovery and Reinvestment Act funding of green technology projects,
where $35.7 billion were appropriated, more than the previous four years combined.

188

policy analysis framework. Yet as the climate science has progressed and begun to influence
policy debates and decisions, the social interactions between institutions and jurisdictions have
become more prominent. This process has been discussed most extensively through the literature
on epistemic communities and boundary organizations (Chilvers, 2008; Haas, 1992).
The cognitive realm of climate science had been developing since the late 19
th
century
and even by the late 1970s opinion among climatologists was split as to whether the planet was
warming or cooling (Rahm, 2010: pp.11-13). As the 1980s progressed, the case for warming
strengthened, and governments around the world began to take note. The first major action took
place at the international level with the creation of the Intergovernmental Panel on Climate
Change in 1988.
Climate change policy has been dominated by the United Nations Framework
Convention on Climate Change, which was first open for signatures in 1992, and entered into
force in 1994. This Convention produced a secretariat to support operations, and parties to the
convention have met annually since to develop and negotiate climate change policy. The Kyoto
Protocol of 1997 was agreed upon at the third Convention of the Parties (COP 3), establishing
legally binding emissions reductions obligations for developed countries. COP meetings have
since explored various issues – including sustainable development (COP 7), climate change ethics
and adaptation for developing countries (COP 10) – and have agreed upon extensions to Kyoto
such as international emissions trading (UN‘s Joint Implementation and Clean Development
Mechanism, COP 6), as well as technology transfers to developing countries (COP 8).
68

68
As highlighted by the Institutional Analysis and Development Framework (Ostrom, 2007), in a non-
hierarchical world (such as the international system), mutual and credible agreements are necessary to
avoid the problem of free riders.

189

Table 37: Major Climate Change Policy Landmarks, Worldwide 1988-2010.
Year International National Subnational
1988 Intergovernmental Panel on
Climate Change established

1989
1990
1991
1992 UN Framework Convention on
Climate Change (UNFCCC)

1993
1994 UNFCCC enters into force
1995
1996
1997 UN Kyoto Protocol (COP 3)
1998 BP ETS
1999
2000 Denmark ETS
2001 COP 6: International emissions
trading mechanisms agreed upon
UK ETS; U.S. rejects Kyoto
Protocol

2002
2003 Chicago Climate
Exchange; New South
Wales, Australia ETS
2004 COP 10, Buenos Aires Plan of
Action promotes adaption of
developing countries to climate
change

2005 EU ETS first phase begins; COP
11, Montreal Action Plan to
extend Kyoto Protocol
Norway ETS; Japan Voluntary
ETS
RGGI ETS
2006 California AB32
2007 Kyoto Protocol goes into affect Massachusetts v. EPA, Supreme
Court finds GHG air pollutants
covered by Clean Air Act

2008 EU ETS second phase begins New Zealand ETS
2009 COP15 Kyoto renegotiations
postponed
US Waxman-Markey House bill
passes, but Kerry-Lieberman
Senate bill stalls; Australia
Senate rejects ETS bill.

2010 Tokyo ETS

190

Table 38: Climate Change Mitigation Policy in Selected OECD countries.
Australia Canada European
Union
Japan United States
% of global
emissions*
1.45% 1.94%
13.37%
(EU 27)
3.56% 18.44%
Emissions per
capita*
26.9 tons 22.6 tons
10.3 tons
(EU 27)
10.5 tons 23.5 tons
Kyoto Protocol
target
108% 94% 92% (EU 15) 84% N/A
Change in
emissions
1990-2007
+30.0% +26.2%
-4.3% (EU 15)
-9.3% (EU 27)
+8.2% -16.8%
Domestic
emission
targets
5-25% below
2000 levels by
2020; 60%
below 2000 by
2050
20% below
2006 levels by
2020; 60-70%
below 2006
levels by 2050
20-30% below
1990 levels by
2020
25% below
1990 levels by
2020
Return to 1990
levels by 2020
Primary policy
instrument
No overriding
mechanism
No overriding
mechanism
Emissions
trading
International
offsets
No overriding
mechanism
Source: Torney and Gueye, 2009; * = 2005.
Table 39: U.S. Federal Climate Change Funding ($billions)
‗98 ‗99 ‗00 ‗01 ‗02 ‗03 ‗04 ‗05 ‗06 ‗07 ‗08 ‗09 ARRA
Technology
1.6 2.2 2.2 2.0 2.0 3.0 3.3 3.1 3.0 3.7 4.3 5.2 35.2
Climate Science
2.2 2.1 2.1 2.1 2.0 2.4 2.3 2.1 1.8 1.9 1.9 2.0 0.5
International Aid
0.2 0.4 0.2 0.3 0.3 0.3 0.3 0.3 0.3 0.2 0.2 0.3 0.0
Total
4.0 4.5 4.4 4.4 4.2 5.7 5.8 5.4 5.1 5.8 6.4 7.5 35.7
Forgone
Revenues from
Tax Preferences n.a. n.a. n.a. n.a. n.a. 0.7 0.6 0.4 1.2 1.6 1.9 2.2 n.a.
Source: CBO (2010)
These international agreements have influenced national policy in numerous ways.
Signatories to the Kyoto Protocol then had to decide how to meet their agreed upon targets. The
European Union acted collectively to achieve their goals; which was made possible by the
UNFCCC COP allowing emissions trading between countries under the Kyoto protocol. As
shown in Table 39, a number of European states preempted the EU emissions trading scheme,
essentially gaining practice in the process. The U.S. and Australia took different approaches, both
failing to ratify the Kyoto protocol at the federal level. In both instances, this sparked policy
initiative at the sub-national level (Abate, 2005-6). New South Wales in Australia, the RGGI in

191

North America, and AB32 in California
69
are the three examples currently in operation, yet
numerous more at both the state (New Mexico, Illinois, Florida, Massachusetts, Oregon) and
regional (Western Climate Initiative, Midwestern Greenhouse Gas Regional Accord [Rose et al.,
2009]) levels have been debated.
It is important to note that the growth in climate change policy is not solely an elitist
phenomenon. This increase appears to have run in tandem with increased awareness of climate
change science within both policy elites and public opinion. What has caused this increase, and
why are there so many different policy responses adopted worldwide? Knowledge of climate
change science appears to play an important role. Numerous studies have highlighted the
divergence of awareness and opinion regarding climate change between European and US
citizens (Bord, Fisher & O‘Connor, 1998; Brewer, 2004; Leiserowitz, 2005; Lorenzoni,
Nicholson-Cole & Whitmarsh, 2007,) which helps to explain these regions‘ different policy
approaches. The role of mass media (Antilla, 2005; Boykoff, 2007; Carvalho, 2007; Carvalho and
Burgess, 2005; Weingart, Engels, & Pansegrau, 2000) and non-governmental organizations
(Carpenter, 2001; Gough and Shackley, 2001) in influencing public opinion has been
documented. However, public opinion is only indirectly related to policy making. NGO and
epistemic community influence on policy makers and policy negotiations is more direct, and has
also been documented (Agrawala, 1998).
Institutional mechanisms are an important filter in the role of knowledge in policy
decision making. Australia and the U.S. both failed to ratify the Kyoto protocol, and have
subsequently deliberated yet failed to pass federal level climate change legislation. These
representative bodies stand in contrast to the U.S. state level EPAs which have passed climate
action plans, and the EU Commission which passed the EU Emissions Trading Scheme. Public

69
The Regional Greenhouse Gas Initiative is a cooperation of 10 states in North East US and East Canada.

192

opinion can play an important role when decision making is representative. California, whose
citizens are more supportive of climate change policy, has passed major climate change policy
through the parliamentary process, and a public referendum has since upheld the policy.
A MOVE TOWARDS ADAPTATION
In recent years, policy debate has moved towards discussion of adaptation. While
mitigation efforts aim to reduce the likelihood of catastrophic climate change in the future,
adaptation policy aims to address the consequences of climate change, whether preemptively or
retroactively. Much discussion in this area centers on the protection of the most vulnerable
segments of society. It has been argued that climate change has cause the displacement of
communities in flooded areas of Bangladesh and inhabitants of some low-lying islands in Papua
New Guinea among many others. Indeed, the potential for climate change to induce migration is
evident in the numerous examples where water scarcity, floods, deforestation, land degradation,
salination, dust storms, and extreme weather – all of which can be caused by climate change –
have caused displacements (Reuveny, 2007: pp663-7). However, even wealthy regions such as
California have begun to turn their attention to the question of adaptation, as highlighted by the
California Climate Adaptation Task Force (Pacific Council, 2010).
As with climate change policy more generally, the initiative to move towards adaptation
policy began in academia and spread to international organizations – the UNFCCC first discussed
adaptation at COP 1 in 1995, and began address the issue in policy terms in the Marrakesh
Accords of 2001. Two prominent areas of policy have emerged from COP conferences since; the
study of vulnerability and adaptation to climate change, and the development of aid funds that
target adaptation. Similar efforts have subsequently developed at the national level in the U.S.
and Europe.

193

A SHIFT IN POLICY TOOLS: FROM COMMAND AND CONTROL TO TAX AND
EMISSIONS TRADING SCHEMES
Another important aspect of policy interactions are those between the stages of the policy
process, and the feedback loops that occur therein. Once a climate change policy has the political
traction to be passed, decision makers must choose a policy tool with which to achieve desired
outcomes. In line with broader air quality policy, carbon taxes and emissions trading schemes
have become increasingly popular policy instruments. In comparison to the so-called ―command-
and-control‖ policies of the 1970s – the first, and costly, attempts at environmental regulation –
various relatively low-cost regulations began to emerge. Emissions taxes, which added costs to
productive activities relative to their emissions output, are easiest applied to a single emissions
source, such as petroleum. Emissions taxes became particularly popular in Europe during 1990s,
yet failed to gain traction in the U.S. Instead, U.S. policy makers initially turned to emissions
trading schemes (ETS), which cap emissions at a given level and allow regulated companies to
trade their emissions allowances in a market. In contrast to emissions taxes, ETS are relatively
effective at dealing with multiple source emissions.
What factors are driving these decisions and what role does knowledge play in policy tool
choice? Academic treatment of policy tools comparison has influenced the policy agenda over
time. Environmental economists writing about pollution regulation more generally (Baumol &
Oates, 1971; Dales, 1968) were able to show the cost effectiveness of emissions trading schemes
when compared with the previously-used command and control approach of regulation. In other
words, equivalent emissions reductions could be achieved at lower cost by using market forces to
determine the facilities at which emissions could be reduced at lowest cost to society as a whole.

194

Yet ideas within the ivory tower do not easily pass into policy decisions. Emissions
trading schemes were discussed in the environmental economics literature of the 1970s, yet the
first examples only appeared in practice in the 1990s. Some of this delay is due to knowledge
taking some time to pass through from academics to their students and the relevant decision
makers. In some cases this is due to environmental economists gaining positions of power
themselves. For example, the appointment of Jos Delbeke, an economist, into a leading position
in the EU Commission‘s climate change unit played an important role in that organization‘s
change in stance – from opposed to emissions trading in 1997 to proposing the largest program of
its kind in 2000 (Skjaerseth and Wettestad, 2010).
But once that knowledge transfer has occurred, there also remains a legitimacy gap. How
would academic ideas gain legitimacy in the decision making if there are no examples in
practice? Moreover, because of the ―informal evaluation‖ process discussed above – whereby
decision makers use particular decision rules to identify specific policy alternatives to be
compared – once policy approaches have gained legitimacy, they would more easily adopted by
policy makers. This is the logic behind the innovations and diffusions model (Berry and Berry,
2003), and appears to be at play with the spread of emissions trading schemes. The US Sulfur
Dioxide and Los Angeles RECLAIM programs were touted among environmental economists as
a success
70
and subsequently the EU and numerous other countries and sub-national jurisdictions
have adopted the policy approach. It is important to note that this model relies on there being a
large enough pool of jurisdictions between which policy approaches can diffuse, which is clearly
the case with a broad number of jurisdictions at international, national, and sub-national levels.

70
It is important to note however that those in the environmental justice field saw the outcomes of the Los
Angeles RECLAIM program quite differently (Drury et al, 1998; Chinn, 1999)

195

A GROWING INFLUENCE OF POLICY ANALYSIS NETWORKS
Non-governmental actors have played an increasingly important role in climate change
policy analysis. The following examples highlight the ongoing dialectic between the cognitive
elements of academic and analysis organization modeling capabilities and the interactive
elements of policy making authorities.
Only since 2005 has the US EPA produced a consistent stream of economic analyses
regarding climate change (EPA, 2010). In these analyses EPA has performed over the past few
years, which have largely been commissioned by Congress, have relied on a combination of
analytical models. These models include ADAGE (a computable general equilibrium (CGE)
model developed and run by RTI International), IGEM (an intertemporal general equilibrium
model developed and run by Dale W. Associates, which is directed by Harvard economist Dale
W. Jorgenson), FASOMGHG (a forestry and agricultural sector model developed by Dr McCarl
at Texas A&M University), GTM (a global timber model developed by Dr. Sohngen at Ohio
State University), and MiniCAM (a climate assessment model developed an run at the Joint
Global Change Research Institute, University of Maryland).
Non-governmental organizations were also important in developing knowledge around
emissions trading for the EU policy. Skjaerseth and Wettestad (2010) describe how the EU
Commission‘s climate change unit ―built up its independent knowledge base on emissions trading
by contracting the external consultants Foundation for International Law and Development
(FIELD), together with the Center for Clean Air Policy (CCAP) in Washington DC, which
produced seven scoping papers and two summary reports on design issues.‖
A similar trend can be seen at the state and regional level of the U.S. Private companies
such as McKinsey and ICF International, as well as the non-profit Center for Climate Strategies,

196

have worked with state governments on their Climate Action Plans (see Table42 and Figure 23),
as well as with regional climate initiatives, particularly in the areas of economic impacts
modeling and emissions inventories. The internet connections between these groups and
government agencies are highlighted in the appendix below. As Adam Rose, a consultant with
Center for Climate Strategies suggests, this process of knowledge influence here is also rife with
disagreements. In many cases, business lobby groups hire consultants such as Charles Rivers
Associates and Beacon Hill International to provide alternative analyses. These groups tend to
produce more pessimistic analyses, with large negative impacts to the economy from climate
change policies. Interestingly, the majority of states have adopted climate change action plans, in
spite of these more pessimistic analyses. Adam Rose suggests decision makers‘ thinking may be
influenced by the critique and rebuttals of each side‘s work that takes place during the decision
making process.

Figure 19: U.S. State Climate Action Plans, 2010 (CCS, 2010).

197

Table 40: U.S. State Climate Action Plans and Regional Climate Initiatives, 2010.
State/Region Year Completed Consultant
Alabama 1997
Alaska 2009 CCS
Arizona 2006 CCS
Arkansas 2008 CCS
California 2006 ICF,CCS
Colorado 2007 CCS
Connecticut 2005 CCS
Delaware 2000
Florida 2008 CCS
Hawaii 1998/In progress CCS
Idaho In progress
Illinois 1994/2007 CCS
Iowa 1996/2008 CCS
Kansas 2010 CCS
Kentucky 2008 CCS
Maine 2004 CCS
Maryland 2008 RESI, CCS
Massachusetts 2004 CCS
Michigan 2009 CCS
Minnesota 2008 CCS
Missouri 2002
Montana 2007
Nevada 2008 CCS
New Hampshire 2001/2009 CCS
New Jersey 2008/In progress CCS
New Mexico 2006 CCS
New York 2003 CCS
North Carolina 2007 CCS
Oregon 2004/2008 CCS
Pennsylvania 2009 CCS
Rhode Island 2002 CCS
South Carolina 2008 CCS
Tennessee 1999/2007
Utah 2007 CCS
Vermont 2007 CCS
Virginia 2008 CCS
Washington 2008 CCS
Wisconsin 2008 CCS
RGGI 2008 ICF International
Western Climate Initiative In progress ICF International
Midwestern Greenhouse Gas
Regional Accord
In progress ICF International

198

A BROADENING OF ANALYTICAL CRITERIA: AN INCREASING ROLE FOR
ECONOMIC SECTORS, REGIONS, AND EQUITY.
As highlighted above, climate change policy analysis has both influenced and responded
to policy makers and policy decisions. A further example of this dialectic between the cognitive
and interactive elements of the policy process is the broadening of analytical criteria in climate
change policy analysis. Economic efficiency is the standard focus of economic analysis. Yet
policy makers and citizens are often less concerned by the overall social welfare impacts, and
more concerned with the distribution of costs and benefits between groups within society. In
particular, there has been increasing attention paid in the environmental economics literature to
the role of economic sectors, regions, and distributional equity. Economic sectors are commonly
modeled in economic analyses because they are the foundations of input-output analysis, which is
central to most macro-economic models (Rose 1999). However, the modeling of multiple regions
has entered into analyses only recently (Burtraw et al, 2008). This analytical criteria is clearly
influenced by growth in climate policy that crosses traditional economic regions, such as the
climate initiatives in the U.S.
The distributional impacts of climate change policy have been a concern of academics for
longer. the distributional implications of emissions trading schemes
71
for climate change policy
(Boyce & Riddle, 2007; Burtraw et al, 2009; Hassett et al, 2009; Metcalf, 2009; Rausch et al,
2010; Shammin & Bullard, 2009), a strand which examines the costs of policy across household
income brackets. Climate policies can impact households if they cause prices of GHG-intensive
products (and these products‘ downstream customers) to increase. Low income households tend

71
The relative health costs and benefits of ETS are important to consider, but beyond the scope of this brief.
Also, not all of these studies have explicitly modeled emissions trading schemes. However, carbon taxes
and emissions trading schemes are usually modeled in a similar if not identical way (i.e. by increasing
production taxes by a level relative to the emissions intensity of that economic sector).

199

to be more vulnerable to price changes of GHG-intensive products, as they commonly spend a
greater proportion of their income on these goods, especially heating, transport, and foods. The
focus on emissions trading scheme-induced price changes (Burtraw et al, 2009; Hassett et al,
2009; Metcalf et al, 2008; Metcalf, 2009; Parry et al, 2007) only portrays one side of equity
impacts.
Policy costs can be passed on to households through other means (Oladosu & Rose,
2007; Fullerton, 2008), such as changes to labor and capital income of households. Emissions
trading schemes would be regressive if lower income households are more dependent on
greenhouse-gas intensive industries for income, or are less able to transition to emerging green
industries. Those studies with a broader definition of income distributional impacts often use
computable general equilibrium models which incorporate both consumption and income
schedule impacts (Fullerton & Heutal, 2007; Oladosu & Rose, 2007; Rausch et al, 2009, 2010).
Most such models have been modeled as both static and dynamic (Rausch et al, 2010), the
difference here being whether relevant exogenous and endogenous factors are updated over time.
Using this comprehensive approach, Oladosu and Rose (2007) find that a regional-level policy
would be relatively progressive, imposing greater costs on higher income brackets than lower
income brackets.
The mechanism for compensating disproportionately impacted households is notable
when allowance auctions can generate revenue, and comparisons are common in the literature
(Burtraw & Palmer, 2008; Metcalf et al, 2008; Metcalf et al 2009; Oladosu & Rose, 2007).
Burtraw et al (2009) compare revenue recycling alternatives, including reductions in payroll tax
and income tax, lump sum payments, and expansions in the Earned Income Tax Credit program,
finding the latter two to be the least regressive. It is likely that other options for revenue
recycling, such as investment in research and development for alternative energy technology, and

200

reductions in corporate taxes, would be relatively more regressive, especially in the short term.
However, such options might have different outcomes in the long run, especially when
considering alternative criteria such as the environmental impact of the given policy.
Despite the controversy over knowledge in the climate change policy process, numerous
government efforts to mitigate GHGs are under way, and more continue to be proposed. What
then is the role of knowledge in climate change policy?
To explore this question, this paper began by arguing that policy processes are too
cumbersome and complex to provide parsimonious theory, while theories of the process do not
provide sufficient traction to fully answer the question. In particular, the critical element of the
dialectic between the cognitive and interactive elements policy has not been sufficiently
articulated in theoretical development. In response this paper aims to answer the question from
the ground up and highlight the importance of this cognitive-interactive dialectic.
Hence, five trends of climate change policy over the past three decades are identified: 1)
An increasing number of climate change policies, and the role of knowledge within these trends;
2) A shift in policy tools, from command and control to tax and emissions trading schemes, with
a focus on the later; 3) A move toward adaptation; 4) A growing influence of policy analysis
networks; 5) A broadening of analytical criteria, an increasingly role for sectors, regions, and
equity.
Analysis of these trends confirms that Wildavsky‘s cognitive-interactive framework is
apparent. Moreover, there is a clear dialectic between the two elements in the case of climate
change policy. Climate change policy was born out of problems identified in the cognitive sphere,
the academic field of climate science, and has influenced policy making since the 1980s. Once
policy makers have determined to mitigate GHG emissions, numerous policy options are at their

201

disposal, which have been the subject of comparative-analytical studies. This highlights one
feedback element of the cognitive-interactive dialectic. Further examples include the funding of
climate science and adaptation research by governments. Policy analysis networks are one
prominent embodiment of the cognitive-interactive dialectic. A broad array of policy knowledge
producers, including academics, private companies, government institutions, and non-profit
organizations can be seen to be interacting around the area of climate change policy.
By emphasizing the cognitive-interactive dialectic, this analysis highlights the importance
of institutional decision rules in determining what types of knowledge are used and how these
types of knowledge might influence policy making. This is especially the case for the wicked
problem of climate change where policy knowledge is contested at all levels. As such,
representative policy making institutions can result in limited policy action when a decision rule
is not reached, while agency-based policy making can produce different outcomes with similar
levels of contested knowledge.

202

CONCLUSION: SEEKING BALANCE

―Using the results from formal economic models, the Review estimates that if we
don‘t act, the overall costs and risks of climate change will be equivalent to
losing at least 5% of global GDP each year, now and forever. If a wider range of
risks and impacts is taken into account, the estimates of damage could rise to
20% of GDP or more. In contrast, the costs of action – reducing GHG emissions
to avoid the worst impacts of climate change – can be limited to around 1% of
global GDP each year.‖
The Stern Review: The Economics of Climate Change, 2006

―Looking back, I underestimated the risks. The planet and the atmosphere seem
to be absorbing less carbon than we expected, and emissions are rising pretty
strongly. Some of the effects are coming through more quickly than we thought
then. This is potentially so dangerous that we have to act strongly. Do we want to
play Russian roulette with two bullets or one? These risks for many people are
existential.‖
Lord Nicholas Stern, World Economic Forum, January 2013

These quotes from Lord Nicholas Stern follow a different trajectory to those of James
Lovelock at the beginning of the introduction chapter. While James Lovelock appears to have
become less certain about the Earth‘s future, Lord Stern appears to have strengthened his stance.
Indeed, this very distinction between two of the eminent scholars in the fields of climate science

203

and environmental economics respectively, highlights the potential for debate and disagreement
within the fields. Seeking balance between contrasting positions is one of the aims of this
dissertation. This work analyzes the economic changes resulting from a U.S. federal ETP, a
market-based policy approach used by governments worldwide to reduce GHG emissions.
Computable general equilibrium (CGE) modeling is used to identify the economic equity of these
policy impacts, specifically the distributional economic impacts across U.S. household income
groups and regions. CGE results are also used to explore the influence of economic impacts on
federal and state level climate policy making.
A number of themes can be traced throughout this dissertation: Change, inequality,
uncertainty, and balance. Change. Lord Stern‘s quotes highlight the potentially dramatic changes
which societies face. Global climate change is predicted to substantially alter social life across the
planet (IPCC, 2007). Increasing temperatures, rising sea levels, desertification of arable lands,
and increasingly volatile weather patterns are predicted to displace populations, damage eco-
systems, and harm global economic productivity. If climate models are accurate, significant
modification of our lifestyles, business practices, technologies, and industrial processes can avoid
the worst climatic changes. Whether or not such mitigation of GHG emissions is undertaken,
there is likely to be substantial adaptation of social life worldwide (IPCC, 2007; Pacific Council,
2010).
Inequality. Lord Stern‘s second quote highlights the potential existential threat faced by
some peoples and countries across the globe. As such, climate change will impact social life with
great inequality. Although climate change is a global phenomenon, the impacts will be localized
to particular ecologies around the world, each with their own characteristics (IPCC, 2007). Poor
communities across the world inhabit locations vulnerable to climate change, and have few
resources with which to adapt and protect themselves against localized climatic changes. While

204

developing countries are expected to be particularly vulnerable to climate change (IPCC, 2007),
vulnerable groups within developed countries such as the US are also likely to be
disproportionately impacted (Morello-Frosch, Pastor, Sadd, & Shonkoff, 2009).
Balance. This dissertation analyzes the outcomes of ETP, a policy approach that is
capable of balancing multiple policy goals, most notably the three ―Es‖ of sustainability:
environmental protection, economic efficiency, and equity (Nugent & Sarma, 2002). This work
identifies and evaluates the trade-offs between these important criteria. CGE modeling balances
production functions of interconnected economic sectors, and calculates the maximum possible
economic output within given policy constraints. This work balances the competing economic
modeling goals of parsimony and detail. Moreover, this dissertation captures both income and
consumption effects, calculates both relative and absolute equity, and compares both inter-
household and inter-regional distributional impacts of policy intervention.
A fourth theme of balance runs throughout this dissertation. This dissertation presents a
balanced account of climate change policy debates, exploring the arguments for all climate policy
standpoints. This work also explores the question of policy impacts while drawing upon two of
the major fields in the public policy field, namely economics and political science. This is not to
suggest that other perspectives on the impact of climate change on social life are not relevant or
important to policy studies. Improved understanding of the ecological, sociological and
psychological impacts of climate change, for example, should lead to better definitions of the
problem and enhance policy making. The field of environmental economics provides both a rich
tradition of scholarship with relevance to climate change policy, as well as a metric – money –
that is broadly known, easily measured, often surveyed and recorded, and arguably a good
indicator of social welfare. On the other hand, the environmental economics field is limited by the
questions that it is able to answer. For example, the field has tended to focus on the costs of

205

climate policy intervention, rather than the benefits (costs avoided) because of the substantial
uncertainty surrounding the causal theory of climate change. The field of political science also
has a burgeoning literature that explores climate change policy (Brewer, 2004; Dierwechter,
2010; Lutsey & Sperling, 2008; McCright & Dunlap, 2003; Rabe, 2004; Selin & VanDeveer,
2005; Zahran et al., 2008), as well as a long tradition of literature regarding the role of economic
interests and knowledge in the political and policy making processes. (Haas, 1992; Hart & Victor,
1993; Smith, 2000; Weible, 2008).
This dissertation focuses on mitigation, and explores the outcomes of a policy tool – ETP
– that is capable of balancing multiple policy goals, most notably the three ―Es‖ of sustainability;
environmental protection, economic efficiency, and equity (Nugent and Sarma, 2002). This work
does focus on the costs rather than the benefits (costs avoided) of policy intervention, due to the
substantial uncertainty that exists in that area.
72
I conduct an economic impact analysis using
CGE modeling, which balances the production functions of interconnected economic sectors and
maximizes output given policy constraints. In doing so, this work aims to balance the competing
goals of parsimony and detail in modeling. Moreover, this dissertation captures both income and
consumption effects, calculates both relative and absolute equity, and compares both inter-
household and inter-regional distributional impacts of policy intervention.
CHAPTER RESULTS SUMMARY
Chapter Two: ―Emissions Trading Policy Modeling‖ used the USCGE model detailed in
Chapter One to calculate the impacts of an ETP on the U.S. economy. The USCGE model
featured innovations, such as the electricity sector being disaggregated into 10 sectors by fuel

72
One famous attempt to estimate the economic costs of climate change, and hence the benefits of
(perfectly implemented) government intervention, is the Stern Review on the Economics of Climate Change
(2006). While the report made a significant contribution to the debate on climate change policy, numerous
scholars have criticized elements of it, especially those concerned with future uncertainties such as the
discount rate, adaptation potentials, the weighting of future damages, and the costs of future mitigation.

206

source: 4 Fossil (Coal, Oil, Gas, and Other Fossil), 1 Nuclear, and 5 Renewable (Biomass,
Geothermal, Hydro, Solar, and Wind). Carbon emissions were obtained from the EPA emissions
inventory, and the emissions constraint was linked to fuel combustion and industrial processes of
regulated sectors. Other notable assumptions included AEEI (in the base case and policy
simulations) and manufacturing electricity-use efficiency improvements in the policy simulations
only. Two allocation schedules were compared: 1) A narrow-based policy regulating electricity
sectors only, similar to that used in the RGGI; and 2) a broad-based policy regulating electricity
sectors along with other polluting sectors such as manufacturing and mining. This latter allocation
schedule reflects climate bills such as the Kerry-Lieberman Senate bill of 2009 and the European
Union Emissions Trading Scheme.
This dissertation makes a number of contributions to the literature with respect to
emissions trading modeling. First, I have a disaggregated electricity sector; many prior studies
have a single electricity sector, or fossil vs. non-fossil, e.g. I have also integrated equity elements
into the CGE model. In particular the multi-sector income distribution matrix, combines with a
household consumption element, to feed household substitutions back into the model. I have also
included EPA emissions data – for both combustion and industrial processes – in the emissions
constraint function. This aims to account for emissions in sectors such as mining.
Results in Chapter Two provided insights into Hypotheses 1-4 that were laid out in the
Introduction Chapter. The first hypotheses – H1: Aggregate results will be less than negative 1%
impact on output (Rose & Dormady, 2011; Rose, Wei & Prager, 2012) – is not supported by the
results. Instead, the impacts are larger for more stringent emissions reductions levels, thus
supporting the second hypothesis – H2: Coase Theorem: Aggregate results will be largely
unaffected by the allocation schedule (Coase, 1960). However, it is notable that the aggregate
economic impacts with respect to emissions allowance allocation – i.e. the broad-based vs.

207

narrow-based policy results – were not substantially different. This reflects the claims of the
Coase Theorem, that aggregate social welfare should be unaffected by emissions rights
allocations. The results were also shown to be somewhat sensitive to key assumptions regarding
electricity-use efficiency improvements in the manufacturing sectors. While the macro-economic
impacts were not substantially altered by changes to these assumptions, the impacts with respect
to particular sectors were. This highlights the importance of policy design being sensitive to
modeling assumptions.
The third set of hypotheses related to sectoral impacts: H3a: GHG-intensive sectors are
likely to bear the greatest regulatory burden; H3b: Downstream and upstream sectors would
absorb some of this burden. Hypothesis H3a was shown to be correct: regulated sectors in both
policy designs (broad-based and narrow-based policies) were the most negatively impacted
sectors in percentage terms. Non-regulated sectors have not only absorbed some of the burden,
per hypothesis H3b, but experience large negative shocks to output also. These results highlight
the interconnectedness of economies which CGE aims to represent. Shocks to one area of the
economy, especially one as fundamental as the electricity generation sector, will have potentially
substantial impacts on other sectors.
Hypotheses H4a and H4b aimed to highlight the potential benefits to particular sectors:
H4a: Green technology and government sectors are likely to benefit relatively more from ETP;
H4b: Sectors importing and exporting green technology and GHG-intensive substitutes are likely
to benefit relatively from ETP. Of the renewable electricity generation sectors, only the hydro-
generated electricity sector increased output. Government sectors fared relatively well in the
simulations.

208

Chapter Three provided two notable modeling features. While the majority of the
literature focuses on consumption effects, and some others provide income effects, there are few
that explore income and consumption effects simultaneously. Only one paper (Rose and Oladosu,
2007) provides both effects in the CGE framework. This analysis extends that approach by
incorporating the Multi-Sector Income Distribution Matrix within the USCGE modeling
framework. This enables detailed, integrated analysis of the impact of climate policy on the
income distribution. In addition, detailed consumption effects are provided by the USCGE model.
Results in Chapter Three confirm findings from the literature that ETP would be
regressive without any redistributive element (Burtraw, Sweeney and Walls, 2009; Grainger and
Kolstad, 2009; Hassett et al., 2009; Metcalf et al, 2008; Metcalf, 2009; Oladosu and Rose, 2007;
Parry and Williams, 2010; Rausch et al, 2010; Rose, Wei and Prager, 2012). As discussed above,
these analyses tend to focus on either consumption effects (Burtraw et al., 2009; Hassett et al.,
2009; Metcalf et al., 2008; Metcalf, 2009; Parry, Sigman, Walls, & Williams, 2005) or income
effects (Rose, Wei and Prager, 2012). Like papers by Parry and Williams (2010) and Rose and
Oladosu (2007), this analysis incorporates both income and consumption effects into the equity
analysis. Like the latter (Rose and Oladosu, 2007) this analysis integrates income and
consumption analysis within the CGE context.
In particular, this analysis highlights the important relationship between incomes and
consumption. In this analysis, the hypothesis proposed in the Introduction - H5a: Income
measures of distribution will be more regressive than consumption measures (Parry & Williams,
2010; Hassett et al, 2009) – was rejected. Total incomes are more inequitable than consumption
patterns in the base case. This is explained mostly by the larger savings rates of higher income
brackets; in contrast lower income brackets borrow more and have higher transfer receipts. Hence
the equity of income distributions are softened when disposable income and consumption patterns

209

are considered. However, in contrast of hypothesis 5a, the policy impacts to consumption appear
to be greater than impacts to incomes. In aggregate, the impact to income of the most stringent
cap is negative 3.8% while the impact to consumption is negative 11.7%. This distinction also has
implications for the equity impacts of ETP. Under the most stringent cap, the change in the Gini
coefficient for total consumption is larger than that for total income. This highlights a broad trend
across each of the measures within this study; i.e. that increasing negative impacts in the
aggregate also translate to increasingly inequitable impacts with respect to income brackets.
There are mixed results with respect to hypothesis 5b – Relative single-factor measures
of income distribution (e.g. Gini coefficient) will highlight different concerns to measures which
examination changes to single brackets only (e.g. analysis of the relative gains to the lowest or
highest bracket) (Rose, Wei & Prager, 2012). The impacts by income brackets follow a relatively
consistent pattern: Impacts to the highest income bracket are small and positives, but impacts
become increasingly negatively as we move down the income bracket spectrum. The inequity of
impacts also increases as the emissions reduction cap becomes more stringent. Consistent results
such as these are captured clearly in the Gini coefficient measure. However, the Gini coefficient
does not reflect the fact that, for the 9.7% and 12.7 emissions reductions simulations, the 15-25k
income bracket is the most negatively impacted. The two lowest income brackets are less
negatively impacted, though still suffer more than the higher income brackets.
Impacts to consumption offer greater support for hypothesis 5b. This is largely because
the consumption element of the USCGE model highlights behavioral responses of households to
policy-induced changes to commodity prices; as such, households at different income levels
prioritize consumption choices between commodities. Changes to consumption for some
commodities reflect changes to the income distribution. For example, Housing consumption is
reduced most by those in the 15-25k income bracket, and lower income brackets are more

210

negatively impacted on the whole. However, the 15-25k income bracket reduces consumption of
transportation goods (GASO, LTRN, OTRA) to a lesser extent than the neighboring brackets of
10-15k and 25-35k. In sum, detailed analysis of consumption results across income brackets
offers greater insights than such analysis of the income distribution impacts.
Chapter Four regionalizes results from the national-level USCGE model down to the state
level. This analytical approach has been undertaken in the regional science literature generally –
such as the ―shift-share‖ approach – and with respect to CGE modeling. Indeed, this
regionalization approach is taken largely from Dixon and Rimmer (2004), who developed an
approach to take national level data from their USAGE model of the national U.S. economy and
apportion it to U.S. states. The analysis in Chapter Four adds data sources to the Dixon and
Rimmer approach, especially in the area of international and inter-state trade, which have been
updated and improved upon since Dixon and Rimmer first wrote their model. Chapter Four
contributes to the literature, as the regionalized-CGE approach has not been undertaken in the
climate policy literature and few papers take results to the state-level. Also, while some papers
explore the state-level impacts of climate policy with respect to consumption, this paper explores
the income and production impacts.
Results in Chapter Four suggest that Hypothesis 6 – States with GHG-intensive sectors
are likely to bear a relatively greater regulatory burden – cannot be supported. Those states with
the largest carbon-intensive industries are Wyoming, Alaska, Louisiana, Oklahoma, and Texas. In
the 9.7% emissions reductions cap simulation, Wyoming is the 3
rd
most negatively impacted state
with respect to output, with Alaska 12
th
, Oklahoma 20
th
, Louisiana 40
th
, and Texas 45
th
out of 51
states.
73
The disposable income and employment results are less contradictory of the hypothesis,

73
DC is not an official state, but is treated as one for the purposes of these analyses.

211

though the relationship is still not strong with Wyoming the most negatively impacted state, and
Alaska ranked 6
th
, Texas 15
th
, Louisiana 20
th
, and Oklahoma 40
th
.
More generally, regression results highlight the generally weak relationship between the
proportion of greenhouse-gas intensive industries in a state and the results from this Regional
Model; however, some relationships are stronger than others. For total GHG intensive industry
proportions, there is some level of correlation with both output and disposable
income/employment
74
at the 6.7% emissions reduction level. However, this relationship is weaker
for disposable income/employment and especially output at the 9.7% emission reduction level. In
each of these cases, it is the proportion of Mining sectors which appears to be driving the
correlation. This is highlighted by the similar results between Mining and Total GHG Intensive
sectors, while Other Fuel Production and Fossil-Generated Electricity both show almost no
relationship with output or disposable income/employment for either simulation.
Chapter Five undertakes a number of analytical steps. First, results from Chapter Four
regarding the economic impact across states of a U.S. federal ETP are compared with the voting
records of state representatives in Congress. In particular, this step will explore the question of
why, in 2009, climate policy was able to pass the House, yet stalled in the Senate. While this
question is narrow in scope, it provides useful insights into the broader questions stimulating this
analysis. Second, Congressional roll-call voting on climate policy bills is explored to identify the
impact of ideology on climate and environmental policy between the years 2007 and 2011.
Given the substantial influence of ideology and partisanship in Congressional voting –
which, as shown in the literature review above, has increased over the past decades – the third

74
The state-level wedges for these factors are derived from the same data in the Regional Model.
Therefore, while the percent change results are different – because the national level results are distinct –
their rankings and relative differences between states are equivalent, causing the regression results to also
be equivalent.

212

area of analysis in this chapter is to identify which federal representatives have voted against their
party line on climate policy bills. This focus on those representatives who ―break rank‖ is inspired
by the logic of regression-discontinuity analysis (Angrist and Pischke, 2008; Lee, 2008), whereby
variations at the margins of a natural break or threshold in a distribution provide pseudo-natural
experiment conditions. While the analysis in this chapter does not claim to emulate such
conditions, the exploration of marginal voting patterns is nonetheless instructive. Finally, this
chapter takes a broader perspective still by identifying four key trends in climate policy within the
US and beyond.
This chapter makes contributions to the literature such that very few papers have
explicitly linked economic policy impacts to political outcomes, and to my knowledge none have
done so for climate policy. While previous analyses by Anderson (2011) and Chupp (2011) have
provided insightful analyses of Congressional roll-call voting for environmental policy as a
whole, there have been no analyses of climate policy specifically. The analysis of breaking rank
representatives also provides useful insights into the motivations behind these key
representatives‘ decision making. There are also analyses of direct and indirect economic impacts
– I have presented the direct impact indicators here, and the indirect (i.e. regionalized-CGE
impacts) are also generally not significant.
Results in Chapter Five suggest that hypothesis 7a – States with relatively greater
regional harms from policy are less likely to have federal representatives supporting climate
policy and ETP (Pizer et al., 2010; Shipan & Lowry, 2001; Zahran et al., 2008) – is not
supported. Voting on climate bills instead appears to be dominated by ideology. While it is
possible that ideology is instead turn influenced by constituent interests, economic or otherwise,
these results suggest that link is not strong. There is no statistical relationship between the
proportion of carbon intensive sectors or Regionalized-CGE results and the Liberal Index score of

213

representatives. As such, the data also do not support hypothesis 7c: States with relatively large
harms to influential GHG-intensive sectors are less likely to have federal representatives
supporting climate policy and ETP. Similarly, hypothesis 7d – Direct regulatory impacts are
likely to have greater influence than overall (direct plus indirect) economic impacts on state level
policy making and federal level representation of state interests – also unsupported by the results.
This is not to say that such forces do not play some role in the decision making of federal roll call
voting. Instead, the analyses here suggest that ideology is dominant and that neither the carbon
intensity of state industries (i.e. direct impacts) nor the regionalize-CGE results (i.e. indirect
impacts) are strongly linked to either ideology or climate bill voting.
If we look at the policy benefits side of the ledger, while hypothesis 7b – States with
relative regional harms from climate change are more likely to have federal representatives
supporting climate policy and ETP (Goulder, 2003; Victor, 2003) – makes sense in theory, the
high degree of uncertainty surrounding climate change impacts to U.S. regions makes such
analysis difficult. It is notable that representatives from particular states are more inclined to
―break rank‖ from their party whip and vote on climate bills. Interestingly, this does not appear to
be strongly linked to breaking rank on environmental policy more general. As such, we cannot
rule out other explanations as to why, for instance, Republicans from states such as Maine and
New Jersey have often broken rank on climate bill voting. Both stats were members of the RGGI,
so experience of state level policy – and hence regulated state entities having a first-mover
advantage – may be a strong explanation with respect to benefits. However, the possible threat of
climate change – and the extent to which those representatives believe the climate science –
cannot be ruled out either.
What does all this mean for U.S. climate policy? It is first important to note that, given
the results identified in the national, sectoral, and equity analyses, climate policy should be

214

designed carefully to avoid exacerbating inequity within the U.S. As argued in the introduction,
there is justification for government intervention under uncertainty – i.e. the precautionary
principle – when the potential harms are particularly catastrophic, where poorer communities are
likely to bear a greater burden, or both. There is a strong case to be made for climate change to fit
within this bracket. However, if government intervention is justified on equity grounds, climate
policy should therefore also be designed to avoid regressive outcomes. While this analysis does
not highlight the potential for redistribution under ETP, numerous studies have. Moreover, some
of these studies have found improved equity can be achieved without sacrificing efficiency.
FURTHER RESEARCH
With respect to emissions trading policy CGE modeling, I aim to model the recycling of
revenue in the auction context and compare them with the grandfathering approach used here.
These comparisons will provide useful insights in terms of specific policy designs. I also aim to
include further GHG emissions into the analysis. This is not likely to have a major impact on the
results present here, though agricultural methane emissions are likely to impact that sector.
Finally, I am very interested by the role that cheap natural gas might play in US emissions
moving forward, and how much this may influence the call for government intervention.
Further research will examine the revenue recycling impacts discussed earlier, as well
further equity measures such as the Kolm-Pollack index. I am also interested to explore a more
general equity analysis approach within the CGE framework. The Miyazawa inter-relational
multiplier has been examined in the input-output context, and provided interesting contributions
to the debate over trickle-down economics.
I aim to make the regional model sensitive to state-level policy path dependency issues,
as well as incorporating state-level consumption impacts. Furthermore, top-down regionalization

215

approaches like this have advanced in recent years to include more bottom-up detail, such as
inter-state substitutions, and I aim to explore elements like this. I plan to extend this analysis
moving forward by deepening the predictive capacity of the regression model, especially with
respect to the current Congress and whether climate policy is likely to pass.

216

BILBLIOGRAPHY
Abate, R.S. (2005-6). Kyoto or not, here we come: The promise and perils of the piecemeal
approach to climate change regulation in the United States. Cornell Journal of Law and
Public Policy, 15, 369-401.
Adger, W.N., Paavola, J., Huq, S., & Mace, M.J. (2006). Fairness in adaptation to climate
change. Cambridge, MA: The MIT Press.
Agrawala, S. (1998). Context and early origins of the Intergovernmental Panel on Climate
Change. Climatic Change, 39(4), 605-620.
Akhurst, M., Morgheim, J., & Lewis, R. (2003). Greenhouse gas emissions trading in BP. Energy
Policy, 31(7), 657-663.
Aldrich, J. (1995). Why parties? Chicago: University of Chicago Press.
Anderson, S.E. (2011). Complex constituencies: Intense environmentalists and representation.
Environmental Politics, 20(4), 547-565.
Angrist, J.D., & Pischke, J.S. (2008). Mostly harmless econometrics: An empiricist’s companion.
Princeton, NJ: Princeton University Press.
Antilla, L. (2005). Climate of skepticism: U.S. newspaper coverage of the science of climate
change. Global Environmental Change, 15(4), 338-352.
Araar, A., Dissou, Y., & Duclos, J. Y. (2008). Household incidence of pollution control policies:
A robust welfare analysis using general equilibrium effects. Centre Interuniversitaire sur
le Risque les Politiques Economiques et l’Emploi. Working Paper 08-09. Retrieved from
http://www.cirpee.org/fileadmin/documents/Cahiers_2008/CIRPEE08-09.pdf
Athanassoglou, S., & Massetti, A. (2012). Climate change impacts and ambiguity. Presented at
Western Economic Association International Conference, San Francisco, June 12.
Baumol, W.J., & Oates, W.E. (1971). The use of standards and prices for protection of the
environment. The Swedish Journal of Economics, 73(1), 42-54.

217

Baumol, W.J. & Oates, W.E. (1975). The theory of environmental policy. Cambridge, UK:
Cambridge University Press.
Bergman, L. (1990). Energy and environmental constraints on growth: A CGE modeling
approach. Journal of Policy Modeling, 12(4):671-691.
Bergman, L. (2005). CGE modeling of environmental policy and resource
management. Handbook of Environmental Economics, Elsevier. Chp 24. vol. 3, pp. 1273-
1306.
Bergman, L., & Hill, M. (2000). Productivity and growth effects of acidification: A dynamic
CGE modeling study of Sweden, Discussion paper, Stockholm School of Economics,
Stockholm
Bergman, L., & Henrekson, M. (2005). CGE modeling of environmental policy and resource
management. In K-G Maler & J.R. Vincent (Eds.) Handbook of Environmental
Economics (Volume 3), (1273-1306). New York, NY: North-Holland.
Berry, F.S. & Berry, W.D. (2003). Innovation and diffusion models in policy research. In. P.A.
Sabatier (Ed.) Theories of the policy process. (223-260). Cambridge, MA: Westview
Press.
Betsill, M., & Hoffman, M.J. (2011). The contours of ―cap and trade‖: The evolution of emissions
trading systems for greenhouse gases. Review of Policy Research, 28(10), 83-106.
Binder, S.A., Lawrence, E.D., & Maltzman, F. (1999). Uncovering the hidden effect of party.
Journal of Politics, 61(3), 815-31.
Bisschop, J., & Meeraus, A. (1982). On the development of General Algebraic Modeling System
in a strategic planning environment. Mathematical Programming Study, 20, 1-29.
Bishin, B. G. (2000). Constituency influences in Congress: Does subconstituency matter?
Legislative Studies Quarterly, 25(3), 389-415.
Bishin, B. (2009). Tyranny of the minority: The subconstituency politics theory of representation.
Philadelphia: Temple University Press.

218

Blanchflower, D.G., & Oswald, A.J. (2004). Well-being over time in Britain and the U.S.A.
Journal of Public Economics, 88(7-8), 1359-1386.
Bond, J.R. & Fleisher, R. (1990). The President in the legislative arena. Chicago: University of
Chicago Press.
Bord, R.J., Fisher, A., & O‘Connor, R.E. (1998). Public perceptions of global warming: United
States and international perspectives. Climate Research, 11, 75-84.
Boyce, J. K., & Riddle, M. (2009). Cap and dividend: A state-by-state analysis. Political
Economy Research Institute. Retrieved from http://www.progressive-
economics.org/papers/CAP_DIVIDEND_states.pdf
Boykoff, M.T. (2007). Flogging a dead norm? Newspaper coverage of anthropogenic climate
change in the United States and United Kingdom from 2003 to 2006. Area, 39(4). 470-
481.
Brewer, M.D., Mariani, M.D.& Stonecraft, J.T. (2002). Northern Democrats and party
polarization in the U.S. House. Legislative Studies Quarterly, 27(3), 423-444.
Brewer, T.L. (2004). U.S. public opinion on climate change issues: Implications for consensus-
building and policymaking. Climate Policy, 4(4), 359-376.
Broder, J.M. (2009). With something for everyone, climate bill passed. The New York Times.
Retrieved from
http://www.nytimes.com/2009/07/01/us/politics/01climate.html?pagewanted=all&_r=0
Brook, A., Kendrick, D., & Meeraus, A. (1988). GAMS: A user’s guide. Redwood City,
California: Scientific Press.
Bromley-Trujillo, R. (2011). Environmental policy support in the American states. (Doctoral
dissertation) Retrieved from ProQuest.
Brownstein, R., & Szymanski, C. (2009). The climate dividing lines. National Journal. Retrieved
from http://www.nationaljournal.com/magazine/the-climate-divinding-lines-20090627
Bullard, R. (1983). Solid waste sites and the Black Houston community, Social Inquiry, 53, 82-
95.

219

Burfisher, M.E., Robinson, S. and Thierfelder, K. (2001). The impact of NAFTA on the United
States. Journal of Economic Perspectives, 15(1), 125–144.
Burniaux, J-M, Martin, J.P. and Martins, J.O. (1992). GREEN: A global model for quantifying
the cost of policies to curb CO
2
emissions. OECD Economic Studies, No. 19.
Burtraw, D., & Palmer, K. (2008). Compensation rule for climate policy in the electricity sector.
Journal of Policy Analysis and Management, 28(4), 819-847.
Burtraw, D., Sweeney, R., & Walls, M. (2009). The incidence of US climate policy: Alternative
uses of revenues from a cap-and-trade auction. RFF Discussion Paper no. 09-17.
Burtraw, D., Walls, M., and Blonz, J.A. (2009). Distributional impacts of carbon pricing policies
in the electricity sector. RFF Discussion Paper no.09-43.
Burtraw, D. (2011). Regulation and modeling in the electricity utility sector. RFF Discussion
Paper. Retrieved from
http://www.rff.org/Publications/Pages/PublicationDetails.aspx?PublicationID=21782
Bushnell, J., & Chen, Y. (2009). Allocation and leakage in regional cap-and-trade markers for
CO
2
. Resource and Energy Economics, 34(4), 647-688.
Busch, C. (2008). Showing the world how it’s done. Retrieved from
http://www.ucsusa.org/assets/documents/global_warming/Showing_the_World.pdf
Carpenter, C. (2001). Businesses, green groups and the media: The role of non-governmental
organizations in the climate change debate. International Affairs, 77(2), 313-328.
Carvalho, A. (2007). Ideological cultures and media discourses on scientific knowledge: Re-
reading news on climate change. Public Understanding of Science, 16(2), 223-243.
Carvalho, A., & Burgess, J. (2005). Cultural circuits of climate change in U.K. broadsheet
newspapers 1985-2003. Risk Analysis, 25(6), 1457-1469.
Chen & Liu (2005). Dynamic incentive contracts under no-commitment to periodic auditing and a
non-retrospective penalty system. Journal of Economics, 85(2), 107-139.

220

Chilvers, J. (2008). Deliberating competence theoretical and practitioner perspectives on effective
participatory appraisal practice. Science, Technology & Human Values, 33(2), 155-185.
Chinn, L.N. (1999). Can the market be fair and efficient? An environmental justice critique of
emissions trading. Ecology Law Quarterly, 26(1) 80-125.
Chupp, B.A. (2011). Environmental constituent interest, green electricity policies, and legislative
voting. Journal of Environmental Economics and Management, 62(2), 254-266.
Climate Change Strategies (2010). State climate action plans. Retrieved from
http://www.climatestrategies.us/library
Climate Research Unit, University of East Anglia (2012). Average annual global temperature
data. Retrieved from http://www.cru.uea.ac.uk/cru/data/temperature/hadcrut3vgl.txt.
Coase, R. (1960). The problem of social cost. Journal of Law Economics, 3, 1-44.
Cole, D. H. & Grossman, P.Z. (1999). When is command-and-control efficient? Institutions,
technology and the comparative efficiency of alternative regulatory regimes for
environmental protection. Wisconsin Law Review, 887-938.
Collier, P., Conway, G., & Venables, T. (2008). Climate change and Africa. Oxford Review of
Economic Policy, 24(2), 337-353.
Conrad, K., & Schroder, M. (1993). Choosing environmental policy instruments using general
equilibrium models. Journal of Policy Modeling, 15(5-6), 521-543.
Conrad, K., & Schmidt, T.F.N. (1998). Economic effects of an uncoordinated versus a
coordinated carbon dioxide policy in the European Union: An applied general
equilibrium analysis. Economic Systems Research, 10(2), 161-182.
Coulson, N.E. (1999). Sectoral sources of metropolitan growth. Regional Science and Urban
Economics, 29(6), 723-743.
Cox, G.W. & McCubbins, M.D. (2005). Setting the agenda. New York: Cambridge University
Press.

221

Cragg, M.I., Zhou, Y., Gurney, K., & Kahn, M.E. (2009). Carbon geography: The political
economy of congressional support for legislation intended to mitigate greenhouse gas
production. Economic Inquiry, 51(2), 1640-1650.
Crowe, G., & Eberspacher E. A. (1998). Partisanship and constituency influences on
Congressional rollcall voting behavior in the U.S. House. Res publica .3 Eds. J. Bona &
A. Burnette. Bloomington, IL: Illinois Wesleyan University, 29-34.
Dales, J.H. (1968). Pollution, property & prices: An essay in policy-making and economics.
Toronto, Canada: University of Toronto Press.
Devarajan, S., & Robinson, S. (2002). The impact of computable general equilibrium models on
policy. Paper presented at Frontiers in Applied General Equilibrium Modeling
Conference, New Haven, CT.
Dervis, K., De Melo, J., & Robinson, S. (1982). General equilibrium models for development
policy, Cambridge, UK: Cambridge University Press.
Dierwechter, Y. (2010). Metropolitan geographies of US climate action: Cities, suburbs, and the
local divide in global responsibilities, Journal of Environmental Policy & Planning, 12:
1, 59-82
Dixon, P.B., Parmenter, B., Sutton, J., & Vincent, D.P. (1982). ORANI: A multisectoral model of
the Australian economy. Amsterdam, NL: North-Holland.
Dixon, P.B. (2006). Evidence-based trade policy decision making in Australia and the
development of equilibrium modelling. CoPS/IMPACT Working Paper Number G-163.
Retrieved from http://www.monash.edu.au/policy/elecpapr/g-163.htm
Dixon, P.B. & Rimmer, M.T. (2002). Dynamic general equilibrium modeling for forecasting and
policy: A practical guide and documentation of MONASH. Amsterdam, Netherlands:
Elsevier Science.
Dixon, P.B. & Rimmer, M.T. (2004). Disaggregation of results from a detailed general
equilibrium model of the U.S. to state level. Monash University, Centre of Policy
Studies/IMPACT, Centre Study g-145. Retrieved from
http://www.monash.edu.au/policy/ftp/workpapr/g-145.pdf June 13th 2011.

222

Dixon, P.B. & Rimmer, M.T. (2010). Validating a detailed, dynamic CGE model of the USA.
Economic Record, 86(s1), 22-34.
Dixon, P.B., Rimmer, M.T., and Tsigas, M.E. (2007). Regionalising results from a detailed CGE
model: Macro, industry and state effects in the U.S. of removing major tariffs and quotas.
Papers in Regional Science. 86(1), 31-55.
Dormady, N.C. (2012). Emissions markets, power markets, and market power: A study of
interactions between contemporary emissions markets and deregulated energy markets.
(Doctoral dissertation). Retrieved from ProQuest.
Drury, R.T., Belliveau, M.E., Kuhn, J.S., & Bansal, S. (1998-1999). Pollution trading and
environmental injustice: Los Angeles‘ failed experiment in air quality policy. Duke
Environmental Law and Policy Forum, 231, 251-55.
Dudek, D. J., & Palmisano, J. (1988). Emissions trading: Why is this thoroughbred hobbled?
Columbia Journal of Environmental Law, 13(2), 217-56.
Evans, D., & Kruger, J. (2006). Taking up the slack: Lessons from a cap-and-trade program in
Chicago. RFF Discussion Paper 06-36. Retrieved from
http://www.rff.org/RFF/documents/RFF-DP-06-36.pdf
Fialka, J. (1998). Global-warming debate gets no consensus in industry. Wall Street Journal,
April 16.
Fiddaman, T. (2010). Climate science, climate policy & Montana. Retrieved from
http://blog.metasd.com/2010/05/climate-science-climate-policy-and-montana/
Fisher-Vanden, K., Schu, K., Sue Wing, I., & Calvin, K. (2011). Decomposing the impact of
alternative technology sets on future carbon emissions growth. Unpublished manuscript.
Retrieved from http://people.bu.edu/isw/papers/Fisher-
Vanden,%20Schu,%20Sue%20Wing,%20and%20Calvin,%20June%2027,%202011.pdf
Fliessbach, K., Weber, B., Trautner, P., Dohmen, T., Sunde, U., Elger, C. E., & Falk, A. (2007).
Social comparison affects reward-related brain activity in the human ventral striatum.
Science, 318(5854), 1305.

223

Fullerton, D. (2008). Distributional effects of environmental and energy policy: An introduction.
NBER Working Paper, No. 14241. Retrieved online at:
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1231690
Fullerton, D., & Heutel, G. (2007). The general equilibrium incidence of environmental taxes.
Journal of Public Economics, 91(3-4), 571-591.
Gillion, D.Q. (2012). Protest and Congressional behavior: Assessing racial and ethnic minority
protests in the district. Journal of Politics, 74(4), 950-62.
Goettle, R.J., & Fawcett, A.A. (2009). The structural effects of cap and trade climate policy.
Energy Economics, 31(2) S244-253.
Goldenberg, S. (2013). How Donors Trust distributed millions to anti-climate groups: The
secretive funding network distributed $118m to 102 groups including some of the best-
known thinktanks on the right. The Guardian, London. Retrieved online at:
http://www.guardian.co.uk/environment/2013/feb/14/donors-trust-funding-climate-
denial-networks
Goulder, L.H. (1995). Environmental taxation and the double dividend: A reader‘s guide.
International Tax and Public Finance, 2(2), 157-183.
Goulder, L. H. (2003). Benefit-cost analysis and climate-change policy. In Griffin, J.M. (Ed.)
Global climate change: The science, economics, and politics, (67-91). Northampton,
MA: Edward Elgar.
Goulder, L. H., Parry, I. W. H., Williams III, R. C., & Burtraw, D. (1999). The cost-effectiveness
of alternative instruments for environmental protection in a second-best setting. Journal
of Public Economics, 72(3), 329-360.
Goulder, L.H. and S.H. Schneider (1999). Induced tecnical change and the attractiveness of CO
2

abatement policies, Resource and Energy Economics, 21(2-3), 211-253.
Grose, C.R. (2005). Disentangling Constituency and Legislator Effects in Legislative
Representation. Social Science Quarterly, 86(2), 427-43.
Grose, C.R. & Middlemass, K.M. (2010). Listen to what i say, not how i vote: Congressional
support for the President in Washington and at home.‖ Social Science Quarterly, 91(1),
143-67.

224

Groseclose, T, Levitt, S.D. and Snyder, J.M. Jr. (1999). Comparing interest group scores across t
ime and chambers: Adjusted ADA scores for the U.S. Congress. American Political
Science Review 93 (1), 33-50.
Guber, D.L. (2013). A cooling climate for change? Party polarization and the politics of global
warming. American Behavioral Scientist, 57(1), 93-115.
Gurney, K. R., Y.-H. Chen, T. Maki, S. Randy Kawa, A. Andrews, and Z. Zhu. (2005).
Sensitivity of atmospheric CO2 inversions to seasonal and interannual variations in fossil
fuel emissions. Journal of Geophysical Research, 110(D10), 308.
Haas, P.M. (1992) Introduction: Epistemic communities and international policy coordination.
International Organization, 46(1), 1–35.
Hahn, R. W. (1989). Economic prescriptions for environmental problems: How the patient
followed the doctor's orders. The Journal of Economic Perspectives 3(2), 95-114.
Hahn, R. W. and R. L. Axtell (1995). Reevaluating the relationship between transferable property
rights and command-and-control regulation. Journal of Regulatory Economics, 8(2), 125-
48.
Hardin, G. (1998). The tragedy of the commons. Science, 280(5364), 682-683.
Hassett, K. A., Mathur, A., & Metcalf, G. E. (2009). The incidence of a US carbon tax: A lifetime
and regional analysis. The Energy Journal, 30(2), 155-177.
Hazilla, M. and R.J. Kopp (1990). Social cost of environmental quality regulations: A general
equilibrium analysis, Journal of Political Economy, 98(4), 853-873.
Hill, M. (2001). Essays on environmental policy analysis: Computable general equilibrium
approaches applied to Sweden, (Doctoral dissertation). Retrieved from ProQuest.
Hochman, H. M. (1974). Rule change and transitional equity. In H. Hochman & G. E. Peterson
(Eds.), Redistribution through public choice (321–341). New York: Columbia University
Press for the Urban Institute.
Horridge M, Madden J, Wittwer G (2005). The impact of the 2002–2003 drought on Australia.
Journal of Policy Modeling, 27, 285–308.

225

IMPLAN. (2010). Input-output data. Retrieved from www.implan.com
Intergovernmental Panel for Climate Change (IPCC, 2007). Summary for Policymakers. In: M.L.
Parry, O.F. Canziani, J.P. Palutikof, P.J. van der Linden and C.E. Hanson, (Eds.), Climate
Change 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II
to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, ,
Cambridge University Press, Cambridge, UK, 7-22. Retrieved from
http://www.ipcc.ch/pdf/assessment-report/ar4/wg2/ar4-wg2-spm.pdf.
Jacobson, M. (2008). On the causal link between carbon dioxide and air pollution mortality.
Geophysical Research Letters. 35(L03809).
Jaffe, J., Ranson, M., & Stavins, R.N., (2009). Linking tradable permit system: A key element of
emerging international climate policy architecture. Ecology Law Quarterly, 36, 789-808.
James, T. E., & Jorgensen, P. D. (2009). Policy knowledge, policy formulation, and change:
Revisiting a foundational question. Policy Studies Journal, 37(1), 141-162.
Johansen, L. (1960). A Multisectoral Study of Economic Growth, North-Holland, Amsterdam.
Johnson, K. (2009). MIT to Republicans: Lay of the scaremongering on climate costs. The Wall
Street Journal, Environmental Capital Blog. Retrieved from
http://blogs.wsj.com/environmentalcapital/2009/04/02/mit-to-republicans-lay-off-the-
scaremongering-on-climate-costs/
Jorgenson, D. W., & Wilcoxen, P. J. (1990). Environmental regulation and US economic growth.
The Rand Journal of Economics, 21(2), 314-340.
Jorgenson, D.W., Goettle, R., Ho, M.S., Slesnick, D.T., & Wilcoxen, P.J. (2010). The
distributional impact of climate policy. The Berkeley Electronic Journal of Economic
Analysis & Policy, 10(2), 17.
Kehoe, T.J. (2003). An evaluation of the performance of applied general equilibrium models of
the impact of NAFTA. Federal Reserve Bank of Minneapolis, Research Department Staff
Report 320.
Knutti, R. (2010). The end of model democracy? Climatic Change, 102(3-4), 395-404.

226

Konishi, H. (2005). Intergovernmental versus intersource emissions trading firms are
noncompliant. Journal of Environmental Economics and Management, 49(2), 235-261.
Kraft, M.E., & Kamieniecki, S. (2007). Business and environment policy: Corporate interests in
the American political system. Cambridge, MA: The MIT Press.
Kunreuther, H.C., & Michel-Kerjan, E.O. (2009). At war with the weather: Managing large-scale
risks in a new era of catastrophes. Cambridge, MA: The MIT Press.
Layzer, J. (2007). Deep Freeze: How Business Has Shaped the Global Warming Debate in
Congress In Kraft, M.E., & Kamieniecki, S. (Eds.). Business and environment policy:
Corporate interests in the American political system. Cambridge, MA: The MIT Press.
League of Conservation Voters (2013). Retrieved from http://www.lcv.org/scorecard/scorecard-
archives/.
Lee, D.S. (2008). Randomized experiments from non-random selection in U.S. House elections.
Journal of Econometrics, 142(2), 675–97.
Lee, H., & Roland-Holst, D. (1997). The environment and welfare implications of trade and tax
policy. Journal of Development Economics, 52(1), 65-82.
Leiserowitz, A.A., (2005). American risk perceptions: Is climate change dangerous? Risk
Analysis, 25(6), 1433-1442.
Leontief, W., Morgan, A., Polenske, K., Simpson, D., & Tower, E. (1965). The economic impact
– industrial and regional – of an arms cut. Review of Economics and Statistics, XLVII,
217–241.
Liew LH (1984). ‗Tops-down‘ versus ‗bottoms-up‘ approaches to regional modeling. Journal of
Policy Modeling 6, 351–367.
Lorenzoni, I., Nicholson-Cole, & Whitmarsh, L. (2007). Barriers perceived to engaging with
climate change among UK public and their policy implications. Global Environmental
Change, 17(3-4), 445-459.
Lovelock, J. (2006). The revenge of Gaia: Why the Earth is fighting back and how we can still
save humanity. London, UK: Penguin.

227

Lowry, W.R. (2008). Disentangling energy policy from environmental policy. Social Science
Quarterly, 89(5), 1195-1211.
Lutsey, N., & Sperling, D. (2008). America‘s bottom-up climate change mitigation policy.
Energy Policy, 36(2), 673-685.
Manne A.S. (1963). Key sectors of the Mexican economy 1960-1970. In A.S. Manne and H.M.
Markowitz, (Eds), Studies in Process Analysis (379-400). New York: Wiley.
Marriot, J. (2009). An Electricity-focused economic input-output model: Life-cycle assessment
and policy implications of future electricity generation scenarios. (Doctoral dissertation).
Retrieved from ProQuest.
Mas-Collel, A., Whinston M.D. and Green J.R. (1995). Microeconomic theory. New York:
Oxford University Press.
Matisoff, D.C. (2008). The adoption of state climate change policies and renewable portfolio
standards: Regional diffusion or internal determinants? Review of Policy Research 25,
527–546.
McCright, A.M., & Dunlap, R.E. (2003). Defeating Kyoto: The conservative movement‘s impact
on U.S. climate change policy. Social Problems, 50(3), 348-373.
McCright, A.M., & Dunlap, R.E. (2011). The politicization of climate change and polarization in
the American public‘s views of global warming, 2001-2010. Sociological Quarterly,
52(2), 155-194.
McElroy, G. & Benoit, K. (2007). Party groups and policy positions in the European parliament.
Party Politics, 13(1), 5-28.
Meagher, GA, Adams, PD, and Horridge, JM. (2000). Applied general equilibrium modelling and
labour market forecasting Center of Policy Studies, Monash University. Retrieved from
http://www.monash.edu.au/policy/conf/17Meagher.pdf
Meeraus, A. (1983). An algebraic approach to modeling. Journal of Economic Dynamics and
Control, 5, 81-108.
Metcalf, G. E. (2009). Designing a carbon tax to reduce US greenhouse gas emissions. Review of
Environmental Economics and Policy, 3(1), 63-83.

228

Miller, R.E. & Blair, P.D. (2007). Input-output analysis: Foundations and extensions. Upper
Saddle River, NJ: Prentice-Hall.
Miller, W.E. & Stokes, D.E. (1963). Constituency influence in Congress. American Political
Science Review, 57(1), 45-56.
Montgomery, W.D. (1972). Markets in licenses and efficient pollution control programs. Journal
of Economic Theory, 5, 395-418.
Montgomery, D., Baron, R., Bernstein, P., Bloomberg, S., Ditzel, K., Lane, L., et al. (2009).
Impact on the economy of the american clean energy and security act of 2009 (HR 2454).
Unpublished manuscript.
Morello-Frosch, R., Pastor, M. and Sadd, J. (2002). Integrating environmental justice and the
precautionary principle in research and policy making: The case of ambient air toxics
exposures and health risks among schoolchildren in Los Angeles. Annals of the American
Academy of Political and Social Science, 584, (Health and the Environment), 47-68.
Morello-Frosch, R., Pastor, M., Sadd, J., & Shonkoff, S. (2009). The climate gap: Inequalities in
how climate change hurts americans and how to close the gap. Program for
Environmental and Regional Equity. Retrieved from
http://dornsife.usc.edu/pere/publications/
Mrozek, J.R., & Keeler, A.G. (2004). Pooling of uncertainty: Enforcing tradable permits
regulation when emissions are stochastic. Environmental and Resource Economics,
29(4), 459-481.
Murphy, J.J., & Stranlund, J.K. (2006). Direct and market effects of enforcing emissions trading
programs: An experimental analysis. Journal of Economic Behavior & Organization,
61(2), 217-233.
Murray, B., & Ross, M. T. (2008). Economic modeling of federal cap-and-trade legislation:
Variations on the Lieberman-Warner bill. Presented at the RFF Workshop on Federal
Policy to Reduce U.S. Greenhouse Gas Emissions Workshop. Resources for the Future,
Washington, DC. June 18, 2008.
NASA (2012b). GISS temperature analysis. Retrieved from
http://data.giss.nasa.gov/gistemp/tabledata_v3/GLB.Ts+dSST.txt.

229

NASA (2012b). Uncertainties: Unresolved questions about Earth‘s climate. Retrieved from
http://climate.nasa.gov/uncertainties/
Nazara, S., & Hewings, G.J.D. (2004). Spatial structure and taxonomy of decomposition in shift-
share analysis. Growth and Change, 35(4), 476-490.
NOAA (2012). Average temperatures. Retrieved from
ftp://ftp.ncdc.noaa.gov/pub/data/anomalies/annual.land_ocean.90S.90N.df_1901-
2000mean.dat
Nokken, T. (2000). Dynamics of Congressional loyalty: Party defection and roll-call behavior,
1974-97. Legislative Studies Quarterly, 25(3), 417-44.
Nordhaus, W.D. (1994). Managing the Global Commons: The Economics of Climate Change.
Cambridge: The MIT Press.
Nugent, J.B., & Sharma, C.V.S.K. (2002). The three e‘s – efficiency, equity, and environmental
protection – in search of ―win-win-win‖ policies: A CGE analysis of India. Journal of
Policy Modeling, 24(1), 19-50.
Oladosu, G. (2000). A non-market computable general equilibrium model for economic analysis
of climate change in the Susquehanna river basin. (Doctoral dissertation). Retrieved from
ProQuest.
Oladosu, G., & Rose, A. (2007). Income distribution impacts of climate change mitigation policy
in the susquehanna river basin economy. Energy Economics, 29(3), 520-544.
Okun, A. (1974). Efficiency and Equity: The Big Trade-Off. Washington, DC: Brookings
Oreskes, N. (2004). The scientific consensus on climate change. Science, 306(5702), 1686.
Pacific Council on International Policy (2010). Preparing for the effects of climate change: A
strategy for California. Retrieved from http://www.pacificcouncil.org
Paltsev, S., Reilly, J.M., Gurgel, A.C., & Metcalf, G.E. (2007). Assessment of U.S. cap-and-trade
proposals. NBER Working Paper. Retrieved from http://www.nber.org/papers/w13176

230

Paltsev, S., Reilly, J. M., Jacoby, H. D., & Morris, J. F. (2009). The cost of climate policy in the
United States. Energy Economics, 31, S235-S243.
Parry, I. W. H., Sigman, H., Walls, M., & Williams III, R. C. (2005). The incidence of pollution
control policies. in H. Folmer and T. Tietenberg, eds., International Yearbook of
Environmental and Resource Economics 2006/2007, Cheltenham, UK: Edward Elgar
Publishers
Parry, I. W. H., & Williams III, R. C. (2010). What are the costs of meeting distributional
objectives for climate policy? The BE Journal of Economic Analysis & Policy, 10(2).
Partridge, M.D., and Rickman, D.S. (2010). Computable general equilibrium (CGE) modeling for
regional economic development analysis. Regional Studies, 44(10),1311-1328.
Paul, A., Burtraw, D., & Palmer, K. (2008).Compensation for electricity consumers under a U.S.
CO
2
emissions cap. RFF Discussion Paper 08-25 Retrieved from
http://www.rff.org/RFF/documents/RFF-DP-08-25.pdf.
Pearson, K.R. (1988). Automating the computation of solutions of large economic models.
Economic Modelling, 7, 385-395.
Percival, R.V. (2006). Who‘s afraid of the precautionary principle? Pace Environmental Law
Review, 23(1), 21-81.
Pigou, A. (1920). The economics of welfare. London, UK: McMillan & Co.
Pizer, W., Sanchirico, J.N., & Batz, M. (2010). Regional patterns of U.S. household carbon
emissions. Climatic Change, 99(1-2), 47-63.
Prager, F., Beeler Asay, G.R., Von Winterfeldt, D., & Lee, B. (2011). Exploring reductions in
london underground passengers following the 2005 bombings. Risk Analysis, 31(5) 773-
786.
Rabe, B. G. (2004). Statehouse and greenhouse: The emerging politics of American climate
change policy. Washington, DC: Brookings Institution Press.
Rahm, D. (2010). Climate change policy in the United States: The science, the politics, and the
prospects for change. Jefferson, NC: McFarland Publisher.

231

Rausch, S., Metcalf, G. E., Reilly, J. M., & Paltsev, S. (2010). Distributional implications of
alternative US greenhouse gas control measures. The BE Journal of Economic Analysis &
Policy, 10(2), Symposium, Article 1.
Reuters (2012). Global CO
2
emissions rise 3 percent in 2011: Report. Retrieved from
http://www.reuters.com/article/2012/07/19/us-emissions-global-
idUSBRE86I0DG20120719
Reuveny, R. (2007). Climate change-induced migration and violent conflict. Political Geography,
26(6), 656-673.
Ringquist, E.J. (2005). Assessing the evidence regarding environmental inequities: A meta-
analysis. Journal of Policy Analysis and Management, 24(2), 223-247.
Ringquist, E.J. (2011). Trading equity for efficiency in environmental protection? Environmental
justice effects from the SO
2
allowance trading program. Social Science Quarterly, 92(2),
297-323.
Rittel, H. W. J., & Webber, M. M. (1973). Dilemmas in a general theory of planning. Policy
Sciences, 4(2), 155-169.
Rohde, D. (1991). Parties and leaders in the postreform House. Chicago: University of Chicago
Press.
Rose, A. (1995). Input-output economics and CGE models. Structural Change and Economic
Dynamics, 6(3), 295-304.
Rose, A. (2009). Introduction to computable general equilibrium analysis. Lecture at USC Price
School of Public Policy. September 2009.
Rose, A. & Dormady, N. (2011). A meta-analysis of the economic impacts of climate change
policy in the United States. The Energy Journal. 32(2), 143-165.
Rose, A., Stevens, B., & Davis, G. (1988). Natural Resource Policy and Income Distribution.
Johns Hopkins University Press, Baltimore, MD.
Rose, A., & Miernyk W. (1989). Input-output analysis: the first fifty years. Economic Systems
Research 1, 229–271.

232

Rose, A., Stevens, B., Edmonds, J., & Wise, M. (1998). International equity and differentiation in
global warming policy. Environmental and Resource Economics. 12(1), 25-51.
Rose, A. & Oladosu, G. (2002). Greenhouse gas reduction policy in the United States: Identifying
winners and losers in an expanded permit trading system. The Energy Journal, 23(1), 1-
18.
Rose, A., Oladosu, G., Lee, B., & Beeler Asay, G. (2009). The economic impacts of the
September 11 attacks: A computable general equilibrium analysis. Peace Economics,
Peace Science, and Public Policy, 15(2).
Rose, A., Stevens, B., & Davis, G. (1982). Natural resource policy and income distribution.
Baltimore, MD: John Hopkins University Press.
Rose, A., Wei, D., and Prager, F. (2012). Distributional impacts of greenhouse gas emissions
trading: Alternative allocation and recycling strategies in California. Contemporary
Economic Policy, 30(4), 603-17.
Rose, A., Wei, D., Wennberg, J., & Peterson, T. (2009). Climate change policy formation in
Michigan. International Regional Science Review, 32(4), 445.
Rutherford, T. (1985). Operating instructions for CASGEN. Department of Operations Research,
Stanford University.
Sabatier, P. A., & Weible, C. M. (2007). The advocacy coalition framework: Innovations and
clarifications. In P. A. Sabatier (Ed.), Theories of the policy process (189-222). Boulder:
Westview Press.
Sandee, J. (1960). A long-term planning model for India. New York: Asia Publishing House.
Sandor, R., Walsh, M., & Marques, R. (2002). Greenhouse gas trading markets. Philosophical
Transactions of The Royal Society A: Mathematical, Physical & Engineering Sciences.
360(1797) 1889-1900.
Schultze, C.L. (1977). The public use of private interest. Washington, DC: The Brookings
Institution.

233

Schweitzer, L., & Stephenson, M. (2007). Right answers, wrong questions: Environmental justice
as urban research. Urban Studies, 44(2), 319-337.
Selin, H., & VanDeveer, S. D. (2005). Canadian-U.S. environmental cooperation: Climate change
networks and regional action. American Review of Canadian Studies, 35(2), 353–378.
Shammin, M. R., & Bullard, C. W. (2009). Impact of cap-and-trade policies for reducing
greenhouse gas emissions on US households. Ecological Economics, 68(8-9), 2432-2438.
Shipan, C.R. & Lowry, W.R. (2001). Environmental policy and party divergence in Congress.
Political Research Quarterly. 54(2), 245-263.
Shoven, J.B., & Whalley, J. (1984). Applied general-equilibrium models of taxation and
international trade: An introduction and survey. Journal of Economic Literature, 22(3)
1007-1051.
Shoven, J.B., & Whalley, J. (1992). Applying general equilibrium. Cambridge, UK: Cambridge
University Press.
Shulock, N. (1999). The paradox of policy analysis: If it is not used, why do we produce so much
of it? Journal of Policy Analysis and Management, 18(2), 226-244.
Skjaerseth, J.B., & Wettestad, J. (2010). The origin, evolution, and consequences of the EU
emissions trading system. Global Environmental Politics, 9(2), 101-122.
Skocpol, T. (2013). What it will take to counter extremism and engage Americans in the fight
against global warming? Retrieved from
http://www.scholarsstrategynetwork.org/sites/default/files/skocpol_captrade_report_janu
ary_2013y.pdf
Smith, M.A. (2000). American business and political power. Chicago: University of Chicago
Press.
Snyder, J.M. Jr., & Groseclose. T. (2000). Estimating party influence in congressional roll-call
voting. American Journal of Political Science, 44,193–211.

234

Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K., et al. (2007). IPCC, 2007:
Climate change 2007: The physical science basis. Contribution of working group I to the
fourth assessment report of the intergovernmental panel on climate change.
Stranlund, J.K., Costello, C., & Chavez, C.A. (2005). Enforcing emissions trading when
emissions permits are bankable. Journal of Regulatory Economics, 28(2),181-204.
Sue Wing, I. (2004). Computable general equilibrium models and their use in economy-wide
policy analysis. Technical Note No.6. MIT Joint Program on the Science and Policy of
Global Change. Retrieved from
http://web.mit.edu/globalchange/www/MITJPSPGC_TechNote6.pdf
Sunstein, C.R. (2005). Laws of fear: Beyond the precautionary principle. Cambridge University
Press, Cambridge, UK.
Sunstein, C.R. & Thaler, R. (2008). Nudge: improving decisions about health, wealth, and
happiness. Yale University Press, New Haven, CT.
The Economist (2006). Big questions and big numbers: We cannot live without big and ambitious
economic models. But neither can we entirely trust them. Retrieved from
http://www.economist.com/node/7159491
Tanger, S. (2012). Party polarization: Congressional divergence environmental policy 1970-
2008. (Master‘s Thesis). Retrieved from: http://etd.auburn.edu/etd/handle/10415/3026
Tietenberg, T. (2008). The evolution of emissions trading. Retrieved from
http://www.aeaweb.org/annual_mtg_papers/2008/2008_90.pdf
Theriault, S.M. (2006). Party polarization in the U.S. Congress. Party politics, 12(4), 483-503.
U.S. Bureau of Labor Statistics (BLS, 2011). Division of occupational employment statistics.
Retrieved from http://www.bls.gov/oes/.
U.S. Bureau of Labor Statistics (BLS, 2011). Consumer expenditure survey. Retrieved from
http://www.bls.gov/cex/

235

U.S. Congressional Budget Office (2010). Federal climate change funding. Retrieved from
http://www.cbo.gov/sites/default/files/cbofiles/ftpdocs/112xx/doc11224/03-26-
climatechange.pdf
U.S. Energy Information Agency. (2006). Manufacturing energy consumption survey (MECS)
Retrieved from http://www.eia.gov/emeu/mecs/
U.S. Environmental Protection Agency, Office of Air and Radiation (2001). The benefits and
costs of the clean air act from 1990 to 2020. Retrieved from
http://www.epa.gov/oar/sect812/feb11/fullreport.pdf
U.S. Environmental Protection Agency (2010). Economics of climate change. Retrieved from
http://yosemite.epa.gov/ee/epa/eed.nsf/pages/ClimateEconomics.html
U.S. Environmental Protection Agency (2012). eGRID data. Retrieved from
http://www.epa.gov/cleanenergy/energy-resources/egrid/index.html
U.S. Internal Revenue Service (IRS, 2011). Tax statistics. Retrieved from
http://www.irs.gov/taxstats/
Varian, H.R. (1992). Macroeconomic Analysis, New York: W.W. Norton & Company.
Victor, D. G. (2003). International agreements and the struggle to tame carbon. In Global climate
change: The science, economics, and politics, edited by J. M. Griffin, 204-29.
Northampton, MA: Edward Elgar.
Webre, P. (2010). Federal climate change programs: Funding history and policy issues.
Collingdate, PA: Diane Publishing.
Weible, C. M. (2008). Expert-based information and policy subsystems: A review and synthesis.
Policy Studies Journal, 36(4), 615-635.
Weller, N. (2009). Trading policy: Constituents and party in U.S. trade policy. Public Choice,
141(1-2), 87-101.
Weingart, P., Engels, A., & Pansegrau, P. (2000). Risks of communication: Discourses on climate
change in science, politics, and the mass media. Public Understanding of Science, 9(3),
261-283.

236

Wettestad, J. (2005). The making of the 2003 EU emissions trading directive: An ultra-quick
process due to entrepreneurial proficiency? Global Environmental Politics, 5(1), 1-23.
Wiene, J.B. (2006-7). Think globally, act globally: The limits of local climate policies. University
of Pennsylvania Law Review.155, 101-119.
Wildavsky, A. (1988). Searching for Safety. Piscataway, NJ: Transaction Publishers.
Wilson, J.Q. (1989). Bureaucracy: What government agencies do and why they do it. New York,
NY: Basic Books.
Winchester, N., Caron, J. and Rausch , S. (2012). Leakage from sub-national climate initiatives,
Presented at Western Economic Association International Conference, San Francisco,
June 12.
Wildavsky, A. & Wildavsky, A. (2008). Risk and Safety. The Concise Encyclopedia of
Economics. Library of Economics and Liberty. Retrieved from
http://www.econlib.org/library/Enc/RiskandSafety.html
Zahran, S, Grover, H., Brody S.D., &Vedlitz, A. (2008). Risk, stress, and capacity: Explaining
metropolitan commitment to climate protection. Urban Affairs Review, 43(4): 447-474.

237

APPENDIX A: USCGE MODEL SECTORS, PARAMETERS, VARIABLES, AND
OTHER KEY ELEMENTS
Table A1: USCGE Code Identifiers.
Identifier Descriptions
i Sectors
gi Commodities
HH Households – 9 income brackets
gv Government institutions
FG_D Federal Government Defense
FG_ND Other Federal Government
SG State and Local Government
ENT Enterprises
IV Investment
STK, SKT Stock change
ROW Rest of World
L Labor
K Capital
E, EP Exports
M, MP Imports
TAX Tax
TE Export Tax
TM Import Tax
TK Capital Tax
TX Indirect Tax
SL Sales
DS Domestic Sales
DD Domestic Demand
VA Value Added
ro Cost function exponent
sh Cost function parameter
tr Tax rate
re Retained earnings rate
dpr Depreciation rate
mps Marginal propensity to save
mpb Marginal propensity to borrow
invs Investment demand shares
cac Capital consumption matrix
stc Inventory change coefficient
fis Factor income distribution shares
msidm Multi-Sector Income Distribution Matrix coefficients
wdflb Wage distortion factor for labor
wdfk Wage distortion factor for capital
SAL Sales
DCQ Domestic goods in consumption
DSL Domestic sales

238

PRD Production
FCPS Factor income distribution
DY Disposable income
SY Household saving
HHPUR Household purchases
UTIL Utility
TRNRC Transfers
REAN Retained earning
HHSAV Household savings
HHBOR Household borrowings
INVD Investment demand
BOP Balance of Payments
PC Price of domestic goods
PX Output price
PD Domestic price
PM Import price
PE Export price
PQ Composite goods supply price
PL, WGRLB Price of labor
PK, WGRK Price of capital
PV Net price
PDMD Demand price
COST Nested CES cost function in calibrated share form
FCU Factor use
FCS Factor supply
DEPR Depreciation
TOTEMS Aggregate emissions across USCGE sectors

Table A2: USCGE Sector Codes and Descriptions.
USCGE
Sector
Code Sector Description
ABEEF Beef Cattle Ranching and Farming
ADARY Dairy Cattle Ranching and Farming
AOLVS Other Livestock
APOUL Poultry and Egg Production
AFISH Fishing and Aquaculture
AOTH Other Agriculture
COAL Coal
CRUD Crude Oil and Natural Gas
OMIN Other Mining
CNSR Construction
MFML Fluid Milk Manufacturing
MOML Non-Milk Dairy Product Manufacturing
MANM Animal Slaughtering and Meat Processing

239

MPTY Poultry Processing
MFSH Seafood Product Preparation and Packaging
MOFD Other Food Processing
MCHM Chemicals
MPET Petroleum Refining
MOND Other Non-Durable Mfg
MPRM Primary Metals
MORD Ordnance
MSEM Electronics
MODR Other Durable Mfg
TAIR Air Transport
TRUK Truck Transport
TWAT Water Transport
TRAL Rail Transport
TOTH Other Transport
TLTP Private Transit
COMC Communications
INFO Information
ELCL Electric Utilities Coal
ELGS Electric Utilities Gas
ELOL Electric Utilities Oil
ELOF Electric Utilities Other Fossil
ELNU Electric Utilities Nuclear
ELBM Electric Utilities Biomass
ELGT Electric Utilities Geothermal
ELHY Electric Utilities Hydro
ELSL Electric Utilities Solar
ELWD Electric Utilities Wind
GASU Gas Utilities
PWAT Private Water Utilities (Retail)
SANT Sanitary Services
WTRD Wholesale Trade
RTRD Retail Trade
REST Real Estate
BANK Finance Banking & Credit
SECB Security Brokers
INSR Insurance
OODW Owner Occupied Dwellings
HOTR Hotel and Restaurants
PSRV Personal Services
VSRV Veterinary Services
WAST Waste Manage Remediation
OBSV Other Business Services
ENTR Entertainment
EDUC Education
MEDC Medical Services
OSOC Other Health & Social Services

240

TLTG Local Public Transportation
GVUT Federal State Local Water Utilities (Wholesale)
FGML Federal Military
OGOV Other Government
SGGV State-Local Government
NCMP Noncomparable Imports

Table A3: Household purchases groups by commodity.
Household
purchases group Commodities produced by sector
Food ABEEF, ADARY, AOLVS, APOUL, AFISH, AOTH, MFML, MOML,
MANM, MPTY, MFSH, MOFD
Housing CNSR, REST, OODW
Gasoline MPET
Public Transport TLTG, TLTP
Other Transport TAIR, TWAT, TRAL
Medical MEDC
Household Goods MPRM, MSEM, MODR, MORD
Other Household
Goods
OMIN, MCHM, MOND, NCMP
Other Household
Services
TRUK, TOTH, COMC, INFO, WTRD, RTRD, BANK, SECB, INSR, HOTR,
PSRV, VSRV, WAST, OBSV, ENTR, EDUC, OSOC, FGML, OGOV, SGGV
Water PWAT, GVUT, SANT
Electricity ELCL, ELGS, ELOL, ELNU, ELBM, ELGT, ELHY, ELSL, ELWD
Other Fuel CRUD, COAL, MPET, GASU

241

APPENDIX B: U.S. 2007 SECTOR-BASED CARBON EMISSIONS INVENTORY
This appendix presents a U.S. 2007 sector-based carbon dioxide emissions inventory, as
well as the steps taken and assumptions made during this process.
Carbon dioxide emissions are presented for 74 sectors of the U.S. economy. These
sectors are listed in the first column of Table B1 below, and are based upon the list of economic
sectors used in the IMPLAN U.S. Input-Output tables. These IMPLAN sectors are presented with
high resolution where the impact of GHG policy would be most variable within sectors, such as
the utilities sector, and aggregated where variance of policy impact between sectors would be
minimal, such as the construction sectors.
The 2007 EPA emissions inventory (EPA, 2009) provides the base data for this sector-
based carbon dioxide emissions inventory. The 2007 EPA emissions inventory details aggregate
GHG emissions for the U.S., along with the calculation process for the following areas: ―Energy,‖
―Industrial Processes,‖ ―Solvent and Other Product Use,‖ ―Agriculture,‖ ―Land Use, Land-Use
Change, and Forestry,‖ ―Waste,‖ and ―Other.‖
Translating the 2007 EPA emissions inventory to the 74 sectors is the primary task. This
process is relatively straightforward for the ―Electricity Generation‖ (EPA inventory, Chapter 3-
3) and ―Transportation‖ (EPA inventory, Chapter 3-13) components of fossil fuel combustion as
the relevant sectors are well matched. However, as shown in the assumptions column of Table B2
below, the transportation sector requires additional data on the proportion of residential
transportation use (which is subtracted from the EPA total), as well as additional data on the
courier and postal service sectors. Translation of the ―Industrial Process‖ (EPA inventory,
Chapter 4) and ―Non-Energy Use Fossil Fuel Consumption‖ (EPA inventory, Chapter 3-30)
elements to their relative sectors is also straightforward. However assumptions are made about

242

the proportion of emissions specific sectors are responsible for, which are also detailed by sector
in Table B2.
The 2007 EPA inventory presents only aggregate figures for fossil fuel combustion in the
industrial and commercial sectors. The following approach is adopted to disaggregate fossil fuel
combustion for these sectors (data for all steps are provided in Table B4 below). Data is taken
from the IMPLAN US Input-Output transactions table to identify the coal, oil, and gas purchases
(using the sum of purchases from ―Oil and Gas Extraction,‖ ―Coal Mining,‖ ―Natural Gas
Distribution,‖ ―Petroleum Refineries,‖ ―Petrol and Coal Products Mftg,‖ and ―Pipeline
Transportation‖) for each industrial and commercial sector. These IMPLAN values are converted
from dollar (purchase) to unit (consumption) values, and then multiplied by the respective carbon
dioxide emission coefficients in CO
2
per MMbtu (EPA Emissions Inventory 2006, Annex 2) for
each fuel type.
Not all coal, oil, and gas purchased by the coal, oil and gas producing sectors will be
consumed by that sector and cause emissions. Some sectors are producing or marginal
companies, purchasing the fuel to trade with other sectors. Hence to avoid double counting,
purchases in these sectors‘ purchases are weighted (reduced) by an estimated ratio of fuel
combustion to total fuel use. Table B3 details the estimation process to identify weights for the
―Oil and Gas Extraction,‖ ―Coal Mining,‖ ―Natural Gas Distribution,‖ ―Petroleum Refineries,‖
―Petrol and Coal Products Mftg,‖ and ―Pipeline Transportation‖ sectors.
Moreover, numerous industrial sectors use fuel as feedstock in their industrial processes.
As industrial process emissions are accounted for in the EPA 2007 emissions inventory, most
industrial sector purchases are also weighted according to the proportion of feedstock use in fuel

243

consumption.
75
For the manufacturing sectors (MECS, 2007), proportions of total manufacturing
carbon emissions per sector are used to improve accuracy (see Table B4, ―% of total mftg CO2
emissions‖). MECS data is presented by energy type and converted to CO2 emissions using the
emissions factors from EPA.
76

The resulting data are then adjusted to match the control totals provided by the EPA
emissions inventory for the aggregated ―Industrial‖ and ―Commercial‖ sectors. Finally, the
emissions inventory in 74 sectors is translated to the USCGE 40 sector scheme. The additional
resolution of the 74 sector plan allows for future improvements in terms of resolution to the
USCGE scheme to be easily implemented.

75
For the commercial sector it is assumed that all fuel purchased would be combusted to provide heat or
electricity.
76
Energy use for electricity, renewable, and shipments are included in the MECS data, though only carbon
emitting fuel types are included in the calculations.
244

Table B1: US 2007 carbon emissions (Tg CO
2
) by relevant sector and NAICS code.

Sector NAICS code
Industrial
Process
Fossil Fuel
Combustion Total

Sector
Agriculture 111000-112000 -13.8 26.9 13.1 I
Forestry and Logging 113000-115000 -910.1 1.4 -908.7 I
Oil and Gas Extraction 211000 7.9 8.1 16.0 I
Coal Mining 212100 0.0 12.3 12.3 I
Other Mining 212200-213000 0.0 7.0 7.0 I
Electric Power Generation, Transmission and Distribution (Coal) 221100 0.0 1561.1 1561.1 I
Electric Power Generation, Transmission and Distribution (Oil) 221100 0.0 43.9 43.9 I
Electric Power Generation, Transmission and Distribution (Gas) 221100 0.0 296.6 296.6 I
Natural Gas Distribution 221200 30.2 1.3 31.5 I
Water, Sewage and Other Systems 221300 0.0 0.2 0.2 I
Construction and Engineering 236000-238000 0.0 90.8 90.8 I
Food, Beverage, Tobacco, Textiles 311000-316000 0.0 101.7 101.7 I
Wood Product Manufacturing 321000 0.0 9.7 9.7 I
Paper and Printing 322000-323000 0.0 87.7 87.7 I
Petroleum and Coal Products Manufacturing 324000 102.8 89.9 192.7 I
Chemical Manufacturing (Petrochemical) 325000 4.1 9.8 13.9 I
Chemical Manufacturing (Fertilizer) 325000 11.5 11.1 22.6 I
Chemical Manufacturing (Other) 325000 7.4 112.8 120.2 I
Plastics and Rubber Products Manufacturing 326000 9.9 12.3 22.2 I
Clay Product and Refractory Manufacturing 327100 0.0 17.3 17.3 I
Glass and Glass Product Manufacturing 327200 5.2 13.2 18.4 I
Cement and Concrete Product Manufacturing 327300 44.5 37.6 82.1 I
Lime and Gypsum Product Manufacturing 327400 14.6 19.2 33.8 I
Other Nonmetallic Mineral Product Manufacturing 327900 3.2 3.3 6.5 I
Iron and Steel Mills and Ferroalloy Manufacturing 331100 70.5 84.6 155.1 I
Steel Product Manufacturing from Purchased Steel 331200 0.0 2.3 2.3 I
Alumina and Aluminum Production and Processing 331300 4.3 9.8 14.1 I
Nonferrous Metal (except Aluminum) Production and Processing 331400 0.8 4.8 5.6 I
Foundries 331500 6.7 8.0 14.7 I
Fabricated Metal Product Manufacturing 332000 0.0 20.3 20.3 I
Machinery Manufacturing 333000 0.0 11.4 11.4 I

245

Computer and Electronic Product Manufacturing (Semiconductors) 334400 0.1 0.0 0.1 I
Computer and Electronic Product Manufacturing (Others) 334100-334600 0.0 0.0 0.0 I
Electrical Equipment, Appliance, and Component Manufacturing 335000 0.0 3.5 3.5 I
Transportation Equipment Manufacturing 336000 0.0 22.7 22.7 I
Furniture and Related Product Manufacturing 337000 0.0 2.0 2.0 I
Miscellaneous Manufacturing 339000 0.0 2.2 2.2 I
Merchant Wholesalers, Durable Goods 423000-424000 0.0 0.0 0.0 C
Wholesale Electronic Markets and Agents and Brokers 425000 0.0 92.8 92.8 C
Store and Other Retailers 440000-450000 0.0 37.6 37.6 C
Air Transportation 481000 2.1 155.7 157.8 T
Rail Transportation 482000 0.7 50.8 51.5 T
Water Transportation 483000 0.7 50.8 51.5 T
Truck Transportation 484000 5.5 407.4 412.9 T
Transit and Ground Passenger Transportation 485000 0.2 16.9 17.1 T
Pipeline Transportation 486000 7.1 34.6 41.7 T
Scenic and Sightseeing Transportation 487000 0.0 3.6 3.6 T
Support Activities for Transportation 488000 0.1 10.2 10.3 T
Postal Service 491000 0.0 1.8 1.8 T
Couriers and Messengers 492000 0.4 30.4 30.8 T
Warehousing and Storage 493000 0.0 0.2 0.2 T
Publishing Industries (except Internet) 511000 0.0 0.5 0.5 C
Motion Picture and Sound Recording Industries, Broadcasting 512000-515000 0.0 1.4 1.4 C
Internet, Telecommunications, and Other Information Services 516000-519000 0.0 7.0 7.0 C
Finance 520000 0.0 6.6 6.6 C
Real Estate, Rental and Leasing Service, Lessors 530000 0.0 3.6 3.6 C
Professional, Scientific, and Technical Services 541000 0.0 8.1 8.1 C
Management of Companies and Enterprises 551000 0.0 0.3 0.3 C
Administrative and Support Services 561000 0.0 10.5 10.5 C
Waste Management and Remediation Services 562000 20.8 1.3 22.1 C
Educational Services 611000 0.0 13.0 13.0 C
Health Care and Social Assistance 620000 0.0 5.2 5.2 C
Entertainment, Recreation, and Tourism 710000-720000 0.0 19.0 19.0 C
Repair and Maintenance 811000 0.0 2.1 2.1 C
Personal and Laundry Services 812000 0.0 1.8 1.8 C
Religious, Grantmaking, Civic, Professional, and Similar Organizations 813000 0.0 3.5 3.5 C

246

Federal, State, and Local Government (OES Designation) 999000 0.0 0.0 0.0 G
Federal Executive Branch (OES designation) 999100 0.0 0.0 0.0 G
…of which utilities 0.0 165.1 165.1 G
State Government (OES designation) 999200 0.0 0.0 0.0 G
…of which electric 0.0 165.1 165.1 G
Local Government (OES designation) 999300 0.0 0.0 0.0 G
…of which electric 0.0 165.1 165.1 G
…of which gas 0.0 2.2 2.2 G
Households Fossil Fuel Combustion* 340.6*
Household Transport* 1124.9*
Others in EPA inventory:
International Bunker Fuels 108.8
US Territories 50.8
Biomass ---Wood, Ethanol 247.8
US Territories 6.7 50.8 57.5
Total (without households, other EPA) 3,658.5
Total (without households, forestry and agro sinks, other EPA) 4,554.0
Total (with households, without forestry and agro sinks, other EPA) 6,019.5
*Households included for accounting purposes only. Sector codes: I = Industrial; C = Commercial; T = Transport; G = Government.

247

Table B2: Carbon emissions data sources and assumptions by sector.

Sector NAICS code Data source Assumptions
Agriculture
111000-
112000
EPA EI 7-1: ―Land converted to cropland‖
(5.9), ―Cropland remaining cropland‖ (-
19.7).
No CO
2
in industrial process. Agriculture crops and
livestock grouped into one sector to ease calculation
of EPA data.
Forestry and Logging
113000-
115000
EPA EI 7-1: Forest acting as carbon sink
(910.0).
―Wetlands remaining wetlands‖ (1) and ―Grassland
remaining or converted to grassland‖ (31.4) not
included. All forestry and logging grouped because
industries integrated.
Oil and Gas Extraction 211000
EPA EI 3-40: Gas ―Field production‖ (7.4),
Petroleum ―Field production‖ (0.3), ―Tank
venting‖ (0.2).

Coal Mining 212100 EPA EI 3-33 Minimal CO
2
released in industrial process.
Other Mining
212200-
213000
EPA EI (no mention) Minimal CO
2
released in industrial process.
Electric Power Generation,
Transmission and Distribution
(Coal)
221100
EPA EI 3-3;
EIA EPIO**
Disaggregated further into private and government
ownership - ratio from EPIO data.**
Electric Power Generation,
Transmission and Distribution
(Oil)
221100
Electric Power Generation,
Transmission and Distribution
(Gas)
221100
Natural Gas Distribution 221200
EPA EI 3-40: Distribution (minimal). Plus
EPA EI 3-40 portion of ―Natural Gas to
Chemical Plants.‖ US Economic Census
2002.
Assume Natural Gas to Chemical Plants is accounted
for in Natural Gas distribution and Pipeline
transportation. Proportion between these sectors
estimated with Economic Census revenue data.
Water, Sewage and Other
Systems
221300 Minimal CO
2
released in process.
Construction and Engineering
236000-
238000

Minimal CO
2
released in process. Assume no land
use related emissions.
Food, Beverage, Tobacco,
Textiles
311000-
316000
Minimal CO
2
released in process.

248

Wood Product Manufacturing 321000 Minimal CO
2
released in process.
Paper and Printing
322000-
323000
Minimal CO
2
released in process.
Petroleum and Coal Products
Manufacturing
324000
EPA EI 4-1: ―Metallurgical Coke‖ (3.8).
Plus EPA EI 3-30: ―Non-Energy Use
Fossil Fuel Consumption‖ (127.2), minus
―Natural Gas‖ to ―Chemical Plants‖ (8.1),
―Lubricants‖ (21), ―Miscellaneous‖ (9.9).
Assume included emissions are the result of this
sector and not downstream in production process.
Chemical Manufacturing
(Petrochemical)
325110 EPA EI 4-2: ―Petrochemical‖ (4.1)
Disaggregation because Petrochemical and Fertilizer
(Ammonia) are major CO
2
producers in chemical
manufacturing sector. Fertilizer is 83 percent of
Ammonia industry.
Chemical Manufacturing
(Fertilizer)
325300 EPA EI 4-1: 83% of ―Ammonia‖ (11.5)
Chemical Manufacturing
(Other)
325000 (rest)
EPA EI 4-1: ―Titanium dioxide‖ (1.9),
17% of ―Ammonia‖ (2.3), half ―Soda Ash‖
(2.05), ―Phosphoric Acid‖ (1.2).
Plastics and Rubber Products
Manufacturing
326000
EPA EI 3-22: ―Miscellaneous products‖
(9.9).
Assume plastics and rubber products are equal to
―Miscellaneous products.‖
Clay Product and Refractory
Manufacturing
327100 Minimal CO
2
released in process.
Glass and Glass Product
Manufacturing
327200
EPA EI 4-1: Half ―Soda ash‖ (2.05), Half
―Limestone and dolomite‖ (3.1).
Assume all limestone and dolomite used in glass and
glass product manufacturing.
Cement and Concrete Product
Manufacturing
327300 EPA EI 4-1: ―Cement‖ (44.5).
Lime and Gypsum Product
Manufacturing
327400 EPA EI 4-1: ―Lime‖ (14.6).
Other Nonmetallic Mineral
Product Manufacturing
327900
EPA EI 4-1, 4-2: Some ―Silicon Carbide‖
(0.1), half ―Lime and dolomite‖ (3.1).

Iron and Steel Mills and
Ferroalloy Manufacturing
331100
EPA EI 4-1: 92% of ―Iron and Steel
Production‖ (73.6), ―Ferroalloy‖ (1.6),
―Flux stone‖ (2).
Split with Foundries using fossil fuel combustion
calculation to estimate ratio.
Steel Product Manufacturing
from Purchased Steel
331200 Minimal CO
2
released in process.
Alumina and Aluminum
Production and Processing
331300 EPA EI 4-2: ―Aluminum‖ (4.3).
Nonferrous Metal (except
Aluminum) Production and
331400 EPA EI 4-2: ―Zinc‖ (0.5), ―Lead‖ (0.3).

249

Processing
Foundries 331500
EPA EI 4-1: 8% of of ―Iron and Steel
Production‖ (73.6), ―Ferroalloy‖ (1.6),
―Flux stone‖ (2).
Split with Iron and Steel Mills and Ferroalloy
Manufacturing using fossil fuel combustion
calculation to estimate ratio.
Fabricated Metal Product
Manufacturing
332000 Only forging, welding likely to emit. Minimal CO
2
released in process.
Machinery Manufacturing 333000 Only forging, welding likely to emit. Minimal CO
2
released in process.
Computer and Electronic
Product Manufacturing
(Semiconductors)
334400 EPA EI 4-2: Some ―Silicon carbide‖ (0.1)
Computer and Electronic
Product Manufacturing (Others)
334100-
334600

Minimal CO
2
released in process; Sectors aggregated
to avoid redundancy.
Electrical Equipment,
Appliance, and Component
Manufacturing
335000 Minimal CO
2
released in process.
Transportation Equipment
Manufacturing
336000 Only forging, welding likely to emit. Minimal CO
2
released in process.
Furniture and Related Product
Manufacturing
337000 Minimal CO
2
released in process.
Miscellaneous Manufacturing 339000 Only forging, welding likely to emit. Minimal CO
2
released in process.
Merchant Wholesalers, Durable
Goods
423000-
424000
No process emissions
Wholesale Electronic Markets
and Agents and Brokers
425000 No process emissions
Store and Other Retailers
440000-
450000
No process emissions
Air Transportation 481000
EPA EI 3-13: ―Commercial Aircraft‖
(153.9), 83% of ―Aviation Gasoline‖ (1.8).
No international bunker fuels (EPA EI 3-13: ―not
accounted for in national emissions totals‖). Assume
proportion of commercial aircraft using aviation
gasoline is same as that using jet fuel.
Rail Transportation 482000 EPA EI 3-13: Diesel Rail (46.0)
Water Transportation 483000 EPA EI 3-13: Various ―Water‖ (50.8)
Truck Transportation 484000
EPA EI 3-13: ―Medium- and Heavy-Duty
Trucks‖ (407.4), 8.5% of ―Light-Duty
Trucks‖ (44.4).

Transit and Ground Passenger 485000 EPA EI 3-13: ―Buses‖ all fuel types (12.1),

250

Transportation Electric ―Rail‖ (4.8)
Pipeline Transportation 486000
EPA EI 3-13: ―Pipeline‖ (34.6). EPA EI 3-
40: some ―Transmission and Storage‖ (0.1)

Scenic and Sightseeing
Transportation
487000
EPA EI 3-13; 2007 Economic Census data.
(3.6)
Proportion of total non-residential transportation,
calculated with Economic Census data.
Support Activities for
Transportation
488000
USPS Annual Emissions Report 2007;
2007 Economic Census data.
Data extrapolated from Courier sector as similar
revenue size.
Postal Service 491000 USPS Annual Emissions Report 2007
Couriers and Messengers 492000
Fedex emissions report, 2007 Economic
Census data
Fedex is typical of the courier industry sector; 39.5
CO
2
tons per $100,000 revenue multiplied by total
sector revenue.
Warehousing and Storage 493000 Minimal CO
2
released in process.
Publishing Industries (except
Internet)
511000 No process emissions
Motion Picture and Sound
Recording Industries,
Broadcasting
512000-
515000
No process emissions
Internet, Telecommunications,
and Other Information Services
516000-
519000
No process emissions
Finance 520000 No process emissions
Real Estate, Rental and Leasing
Service, Lessors
530000 No process emissions
Professional, Scientific, and
Technical Services
541000 No process emissions
Management of Companies and
Enterprises
551000 No process emissions
Administrative and Support
Services
561000 No process emissions
Waste Management and
Remediation Services
562000
EPA EI 3-47: ―Incineration‖ of waste
(20.8).

Educational Services 611000 No process emissions
Health Care and Social
Assistance
620000 No process emissions
Entertainment, Recreation, and
Tourism
710000-
720000
No process emissions
Repair and Maintenance 811000 No process emissions

251

Personal and Laundry Services 812000 No process emissions
Religious, Grantmaking, Civic,
Professional, and Similar
Organizations
813000 No process emissions
Federal, State, and Local
Government (OES Designation)
999000
Federal Executive Branch (OES
designation)
999100
…of which utilities EPA EI 3-3, EIA EPIO**
Total electricity emissions disaggregated into private
and government ownership.**
State Government (OES
designation)
999200
…of which electric EPA EI 3-3, EIA EPIO**
Total electricity emissions disaggregated into private
and government ownership.**
Local Government (OES
designation)
999300
…of which electric EPA EI 3-3, EIA EPIO**
Total electricity emissions disaggregated into private
and government ownership.**
…of which gas EPA EI 3-3, EIA EPIO**
Proportion of local government electric relative to
national gas to electricity ratio. **
Households Fossil Fuel
Combustion*
EPA EI Energy
Household Transport*
Total transportation emissions minus transportation
sector emissions.
Others in EPA inventory:
International Bunker Fuels Not included.
US Territories
Biomass ---Wood, Ethanol In the long run, net emissions do not increase.
*Households included for accounting purposes only. **Utilities: For the electricity sectors, total emissions are retrieved from the EPA EI 3-3 and then
disaggregated on the basis of EIA EPIO, which details the ratio of private to public electricity generation, as well as the ratio of federal state and local
ownership generation; For gas, it is assumed that the ratio of public to private ownership in the gas sector is the same as the electricity sector. Therefore
public gas emissions are taken as the relative proportion of the Natural Gas Distribution sector.
Definitions:
EPA EI - Environmental Protection Agency Emissions Inventory 2007. ―3-3‖ e.g. refers to Page 3-3.
EIA EPIO – Energy Information Administration Electricity Power Industry Overview 2007

252

Table B3: Estimation approach for ratio of fossil fuel combustion to industrial process input energy use.

Purchasing sector
Producing sector
20 – Oil and gas
extraction
21 – Coal mining
32 – Natural gas
distribution
115 –
Petroleum
refineries
119 – Petrol
and coal
products
manufacturin
g
(1) (2) (3) (1) (2) (3) (1) (2) (3)
20 – Oil and gas extraction 0.49 106.1m 52.0m 1 0.2m 0.2m 0.026 517.0m 13.4m
Ratio
estimated
using EIA
MECS 2006
data.
Ratio
estimated
using EIA
MECS 2006
data.
21 – Coal mining 1 13.4m 13.4m 0.49 62.6m 30.7m 1 0 0
32 – Natural gas distribution 0.39 34.6m 13.5m 1 5.1m 5.1m 0.49 2.1m 1.0m
115 – Petroleum refineries 0.39 98.4m 38.4m 1 7.1m 7.1m 0.026 2.5m 0.06m
119 – Petrol and coal
products manufacturing
0 0.4m 0 0 0.1m 0 0 0.01m 0
337 – Pipeline transportation 0.39 30.2m 11.8m 1 0.2m 0.2m 0.026 253.9m 6.6m
Total fossil fuel combustion 118.4m 43.3m 21.1m
Total fuel input use 283.2m 75.4m 775.6m
Combustion as proportion of
total fuel input use
0.46 0.57 0.03 0.49 0.31

(1) Coefficient weighting the ratio of input fossil fuel combustion to industrial process. 1 is assigned where sector does not produce input. 0 is assigned to
Petrol and coal products manufacturing as the fuel outputs are unlikely to be used for fossil fuel combustion in other heavy industrial sectors. 0.026 is
assigned to where Natural gas distribution purchases fuel inputs from Oil and gas extraction, Petroleum refineries, and Pipeline transportation, as 2.6
percent of natural gas is used by the pipeline and distribution sector, according to the EIA ―Natural Gas consumption by end use.‖ 0.39 is assigned
where Oil and gas extraction is the purchasing sector as well as extracting the initial fuel source; e.g. where gas is extracted by Oil and gas extraction
and then transported by Natural gas distribution to an Oil and gas extraction industry site. 0.49 is assigned where fuel input is produced in that sector.
These last two coefficients (0.39 and 0.49) are estimated using data from the EIA MECS 2006 and assuming that the Petroleum refineries sector is
similar to the Oil and gas extraction and Coal mining sectors.
(2) Tons of carbon dioxide input from producing sectors, estimated using process detailed above.
(3) Fossil fuel combustion estimate, calculated by multiplying (1) and (2) for each purchasing sector.

253

Table B4: Fuel combustion CO
2
emissions for industrial sectors.

Sector

Sector code

CO2 emissions
from fuel
combustion
Weights

Industry CO2
emissions
% of
industry
CO2
combustion
emissions
Estimated
CO2
combustion
emissions
(EPA/
IMPLAN/
MECS)
% of fuel
purchases
used for
combustion
in CO2
(2)
CO2 per
mftg
sector
(MECS)
% of total
mftg CO2
emissions
(MECS)
Agriculture 111000-112000 467,607,725 1.0000 467,607,725 0.032 26,922,538
Forestry and Logging 113000-115000 25,159,718 1.0000 25,159,718 0.002 1,448,572
Oil and Gas Extraction 211000 306,778,070 0.4600
(1)
141,117,912 0.010 8,124,871
Coal Mining 212100 374,452,760 0.5700
(1)
213,438,073 0.015 12,288,708
Other Mining 212200-213000 122,021,022 1.0000 122,021,022 0.008 7,025,366
Natural Gas Distribution 221200 775,613,248 0.0300
(1)
23,268,397 0.002 1,339,679
Water, Sewage, Etc 221300 2,865,641 1.0000 2,865,641 0.000 164,989
Construction 236000-238000 1,577,530,628 1.0000 1,577,530,628 0.107 90,826,404
Food, Beverage,
Tobacco, Textiles 311000-316000 271,908,433 1.0000 66,543 0.146 1,766,079,265 0.120 101,682,101
Wood Product Mftg. 321000 32,282,507 0.9369 6,323 0.014 167,812,092 0.011 9,661,789
Paper and Printing 322000-323000 243,951,559 1.0000 57,368 0.126 1,522,571,686 0.104 87,662,140
Petroleum and Coal
Products Mftg. 324000 7,969,437,288 0.1118 58,853 0.129 1,561,975,437 0.106 89,930,813
Chemical Mftg.
(Petrochemical) 325000 2,817,507,603 0.1466 6,393 0.014 169,674,068 0.012 9,768,993
Chem Mftg. (Fertilizer) 325000 0 0.4241 7,268 0.016 192,899,747 0.013 11,106,212
Chem. Mftg. (Other) 325000 0 0.4678 73,800 0.162 1,958,682,294 0.133 112,771,229
Plastics and Rubber
Products Mftg.. 326000 78,789,708 0.9934 8,043 0.018 213,458,150 0.015 12,289,864
Clay Product and
Refractory Mftg. 327100 15,869,534 0.9991 11,337 0.025 300,900,534 0.020 17,324,363
Glass and Glass Product
Mftg. 327200 29,786,534 1.0000 8,655 0.019 229,712,979 0.016 13,225,736
Cement and Concrete
Product Mftg. 327300 43,322,519 1.0000 24,618 0.054 653,371,025 0.044 37,617,869

254

Lime and Gypsum
Product Mftg. 327400 30,373,821 1.0000 12,570 0.028 333,616,323 0.023 19,207,976
Other Nonmetallic
Mineral Prod. Mftg 327900 14,737,146 0.9761 2,191 0.005 58,147,372 0.004 3,347,838
Iron and Steel Mills &
Ferroalloy Mftg 331100 156,600,032 0.6952 55,334 0.121 1,468,594,233 0.100 84,554,385
Steel Prod. Mftg from
Purchased Steel 331200 25,665,468 0.9680 1,510 0.003 40,080,657 0.003 2,307,646
Alum. Production and
Processing 331300 29,315,565 0.9303 6,443 0.014 171,009,527 0.012 9,845,882
Nonferrous Metal
(except Aluminium) 331400 14,408,273 0.8429 3,167 0.007 84,046,223 0.006 4,838,965
Foundries 331500 13,282,037 0.9092 5,235 0.011 138,948,693 0.009 7,999,978
Fabricated Metal Product
Mftg 332000 100,926,591 0.9953 13,305 0.029 353,121,114 0.024 20,330,966
Machinery Mftg. 333000 77,800,746 1.0000 4,934 0.011 198,785,422 0.014 11,445,080
Elec. Equip Appl., &
Component Mftg 334000-335000 62,799,147 0.9793 4,816 0.005 60,016,369 0.004 3,455,445
Transportation
Equipment Mftg. 336000 61,122,050 0.9958 14,851 0.033 394,140,054 0.027 22,692,633
Furniture and Related
Product Mftg. 337000 9,417,423 1.0000 1,303 0.003 34,588,037 0.002 1,991,408
Miscellaneous Mftg. 339000 11,130,711 1.0000 1,439 0.003 38,203,394 0.003 2,199,562
Total Manufacturing 12,110,434,696 456,302 1.000
Total Industrial

14,683,443,813 1.000 845,400,000
(1) Weight calculated in Table B3
(2) Shaded manufacturing sectors in ―% of fuel purchases used for combustion in CO2‖ are provided for comparison purposes, and are not used to
calculate the final ―Industry CO
2
emissions,‖ as described in the text above.

255

Table B5: Industrial Emissions Coefficient calculation.
Sector
Pre-policy
Production
Pre-policy
Emissions
(EPA
Inventory)
Industrial
Emissions
Coefficient
CRUD 260.6125 15.00 0.057557
MCHM 962.2119 23.00 0.023903
MPET 595.2209 95.80 0.160949
MOND 827.6794 16.90 0.020419
MPRM 249.7142 82.30 0.329577
MSEM 670.4251 0.10 0.000149
MODR 2421.333 67.50 0.027877
TAIR 114.8464 2.10 0.018285
TRUK 253.185 5.50 0.021723
TWAT 37.28706 0.70 0.018773
TRAL 54.27394 0.70 0.012898
TOTH 284.057 0.60 0.002112
TLTP 32.23794 0.10 0.003102
GASU 97.29354 30.20 0.310401
WAST 68.01618 20.80 0.305810
TLTG 25.27695 0.10 0.003956

Table B6: Fuel Emissions Constraint Coefficients.
USCGE
Sector
Code
Fuel Inputs
COAL CRUD OMIN MPET GASU
ABEEF 2.29 2.36 1.76 1.59 1.37
ADARY 2.29 2.36 1.76 1.59 1.37
AOLVS 2.29 2.36 1.76 1.59 1.37
APOUL 2.29 2.36 1.76 1.59 1.37
AFISH 2.29 2.36 1.76 1.59 1.37
AOTH 2.29 2.36 1.76 1.59 1.37
COAL 1.30 1.35 1.00 0.91 0.78
CRUD 1.05 1.09 0.81 0.73 0.63
OMIN 1.18 1.22 0.91 0.82 0.71
CNSR 2.29 2.36 1.76 1.59 1.37
MFML 2.29 2.36 1.76 1.59 1.37
MOML 2.29 2.36 1.76 1.59 1.37
MANM 2.29 2.36 1.76 1.59 1.37
MPTY 2.29 2.36 1.76 1.59 1.37
MFSH 2.29 2.36 1.76 1.59 1.37
MOFD 2.29 2.36 1.76 1.59 1.37
MCHM 0.71 0.73 0.55 0.49 0.43
MPET 1.12 1.16 0.86 0.78 0.67
MOND 2.29 2.36 1.76 1.59 1.37

256

MPRM 2.29 2.36 1.76 1.59 1.37
MORD 2.29 2.36 1.76 1.59 1.37
MSEM 2.29 2.36 1.76 1.59 1.37
MODR 2.29 2.36 1.76 1.59 1.37
TAIR 8.47 8.77 6.53 5.91 5.09
TRUK 17.39 17.98 13.40 12.13 10.44
TWAT 42.83 44.30 33.02 29.88 25.72
TRAL 25.04 25.90 19.31 17.47 15.04
TOTH 2.53 2.62 1.95 1.77 1.52
TLTP 3.71 3.84 2.86 2.59 2.23
COMC 4.71 4.88 3.63 3.29 2.83
INFO 4.71 4.88 3.63 3.29 2.83
ELCL 116.05 120.05 89.47 80.96 69.69
ELGS 18.88 19.53 14.56 13.17 11.34
ELOL 14.20 14.69 10.95 9.91 8.53
ELOF 97.90 101.28 75.48 68.30 58.79
ELNU 0.00 0.00 0.00 0.00 0.00
ELBM 0.00 0.00 0.00 0.00 0.00
ELGT 0.00 0.00 0.00 0.00 0.00
ELHY 0.00 0.00 0.00 0.00 0.00
ELSL 0.00 0.00 0.00 0.00 0.00
ELWD 0.00 0.00 0.00 0.00 0.00
GASU 0.07 0.07 0.05 0.05 0.04
PWAT 2.29 2.36 1.76 1.59 1.37
SANT 2.29 2.36 1.76 1.59 1.37
WTRD 4.71 4.88 3.63 3.29 2.83
RTRD 4.71 4.88 3.63 3.29 2.83
REST 4.71 4.88 3.63 3.29 2.83
BANK 4.71 4.88 3.63 3.29 2.83
SECB 4.71 4.88 3.63 3.29 2.83
INSR 4.71 4.88 3.63 3.29 2.83
OODW 4.71 4.88 3.63 3.29 2.83
HOTR 4.71 4.88 3.63 3.29 2.83
PSRV 4.71 4.88 3.63 3.29 2.83
VSRV 4.71 4.88 3.63 3.29 2.83
WAST 2.29 2.36 1.76 1.59 1.37
OBSV 4.71 4.88 3.63 3.29 2.83
ENTR 4.71 4.88 3.63 3.29 2.83
EDUC 4.71 4.88 3.63 3.29 2.83
MEDC 4.71 4.88 3.63 3.29 2.83
OSOC 4.71 4.88 3.63 3.29 2.83
TLTG 2.36 2.44 1.82 1.64 1.41
GVUT 0.00 0.00 0.00 0.00 0.00
FGML 0.00 0.00 0.00 0.00 0.00
OGOV 4.71 4.88 3.63 3.29 2.83
SGGV 0.00 0.00 0.00 0.00 0.00
NCMP 0.00 0.00 0.00 0.00 0.00

257

APPENDIX C: ELECTRICITY SECTOR DISAGGREGATION
This document describes the process for disaggregating data for the Private and Public
Electricity Generation (PELE and GELE) sectors in the USCGE model. A new set of sectors
reflecting electricity generated by different energy types, including Coal, Oil, Gas, Other Fossil,
Nuclear, Biomass, Hydro, Solar, and Wind (see Table C1).
Input-output data for the US often includes electricity generation, transmissions and
distribution activities as one single sector. This is the case for the USCGE model base data, as
well as the IMPLAN 491 sectoring scheme (sector 31). I was not able to find any examples of I-O
tables with this sector disaggregated with any greater resolution. This makes disaggregation of the
Electricity Generation sectors (PELE and GELE – private and public electricity generation
respectively) difficult.
Two approaches for disaggregation are found in the literature. Joe Marriott – a former
Ph.D. student at Carnegie Mellon – wrote his dissertation on disaggregating the electricity sector,
using a top-down estimation procedure based on state level eGRID data from the US
Environmental Protection Agency and employment per sector data from the US Bureau of
Economic Analysis. Ashley Winston – formerly working with Peter Dixon at Monash CoPS –
also wrote a useful discussion of how to disaggregate agricultural and energy sectors. Winston
focused on bio-fuels, and used a largely bottom-up approach, compiling detailed data on the
costs, production, and consumption of each bio-fuel. While Winston‘s approach is more detailed,
Marriott‘s top-down approach seems more appropriate for this exercise because the US EPA
eGRID data is sufficiently detailed, and because the time requirements are less. It is also more
appropriate because of the data requirements in the USCGE model, as described below.
The disaggregation approach employed in this dissertation makes the following
assumptions:

258

1) The Electricity Generation, Transmission, and Distribution sector is disaggregated only by
energy source and not by function. Hence, transmissions and distribution transactions are
assumed to be equivalent to the generation transactions for each energy type (with respect to
the other sectors and institutions in the economy). However, this is a reasonable assumption
given that the two activities are mutually dependent;
2) Public and private Electricity Generation, Transmission, and Distribution are treated as
equivalent. Following deregulation of the industry, the distinctions between public and
private provision are less pronounce. Moreover, over the long run – which is the time
framework for this study – the distinctions between the two in terms of the capacity to invest
in GHG emissions reduction technologies is marginal;
3) The inputs for electricity generation would largely come from the state in which a facility is
located in. Hence the sectoral mix of a state can be used as a proxy with which to
disaggregate the electricity generation inputs for that state. Similarly, the outputs of
electricity generation are purchased largely by sectors producing in a given state, so the same
proxy can be applied (this time adjusting the electricity generation to account for electricity
transfers between states).
77

This approach combines two sets of data:
1) A state-sector matrix representing industry output within a state as a share of industry output
for the whole US. Gross output data per sector and state is readily available at the NAICS 2-
digit level (around 80 sectors) from the Bureau of Economic Analysis‘ Regional Economic
Account website. Converting the 80 sectors in this data set to the 35-50 sectors used in the

77
Marriott provides an example: ―If we know that 60% of all widgets are manufactured in Idaho, and 40%
are produced in Kentucky, the generation mix of Idaho – expressed as a 6 element array where each
element is a percentage of a particular generation type – shown in Figure 8, is multiplied by 0.6 and the
generation mix of Kentucky is multiplied by 0.4. This produces two new arrays, which are added to
produce a [third] single array. This [third array] is the new sector consumption mix for widgets.‖

259

CGE model is a straightforward process. (Marriott uses employment by state and industry,
which seems to be a less than ideal indicator);
2) A state-energy type generation/consumption matrix (generation for outputs, consumption for
inputs) representing the proportion of each type of energy (i.e. coal, oil, gas, nuclear,
renewables…) generated or consumed within a given state. I also already have this data,
obtained from the US EPA eGRID dataset. The generation matrix is presented in Table C6
below – the consumption matrix is similar but accounts for flows of electricity between
states.
Further adjustments have to be made for the columns of the input-output data, that is, the
inputs to electricity generation, transmission, and distribution. These adjustments are summarized
in Table C10 below. Fuel inputs to electricity generation need to be allocated to the sector
generation power with that particular energy type. Coal inputs are allocated to coal-powered
electricity generation and the same logic applies for CRUD (crude oil) and MPET (petroleum
refining), both of which are allocated to the oil-powered electricity generation sector. It is
important to note that the input-output tables include zero values for the GASU (gas utilities)
sector, so that is not allocated here. However, this allocation would be necessary in the base data
for a different year including GASU inputs to electricity generation. The inputs from AOTH
(other agriculture) are allocated to biomass-generated electricity generation.
Adjustments are also required for three further sectors: MCHM (chemical
manufacturing), WAST (waste management and remediation), and the water utility sectors,
GVUT (public) and PWAT (private).

260

CHEMICAL MANUFACTURING
Chemical manufacturing sector inputs to electricity generation are numerous. Products
produced by this industry can range from fuel inputs – including various petrochemicals and
chemical gases – to industrial gases used in the fuel combustion process, to chemicals used in the
nuclear power generation process, to lubricants and plastics used in turbines, and to paints and
adhesives. Disaggregating the chemical inputs to electricity generation into the relevant energy
types is not straightforward. Tables C7 and C8 below provide details of an estimation procedure
which is based on 2007 U.S. Economic Census data for the ―Value of sales, shipments, receipts,
revenue or business done‖ for various chemical sub-sectors. These data are presented in Table
C7, and separated into five categories:
1) Those subsectors ―Predominantly used by fossil fuel electricity generation‖ (i.e. 32511
Petrochemical manufacturing and 32512 Industrial gas manufacturing) are allocated as inputs
to the fossil fuel generation according to the shares determined in the state-energy
disaggregation approach above;
2) The subsector ―Predominantly used by nuclear electricity generation‖ (32518 Other basic
inorganic chemical) is allocated to the nuclear electricity generation sector;
3) Those subsectors ―Not significantly used‖ by fossil fuel electricity generation (e.g. 32519
Other basic organic chemical manufacturing) are first weighted by a factor of 0.1,
78
and then
allocated as inputs to all electricity generation sectors according to the shares determined in
the state-energy disaggregation approach above;
4) ―No weighting across energy types‖ refers to subsectors that are likely to be used as inputs to
all electricity generation sectors (such as 32521 Resin and synthetic rubber manufacturing

78
This weight is not based on empirical data, but aims to reflect the relatively small use of these sectors
inputs in the electricity generation sectors.

261

and 32531 Artificial and synthetic fibers and filaments manufacturing). These are allocated to
all electricity generation sectors according to the shares determined in the state-energy
disaggregation approach above. However, first, these values are weighted by a factor of
0.25
79
to account for the fact that inputs to electricity generation from these subsectors are
relatively less than inputs to all sectors.
5) Subsectors ―Not used‖ as inputs to electricity generation are discounted from estimations.
Tables C8 and C9 present the calculations for the allocation of chemical inputs to electricity
generation across electricity generation sectors, as based on the categories above. Chemical
manufacturing inputs to electricity generation sectors ($175.68m, from the input-output table)
are separated into the five categories described above according to the relevant proportions
(Table C8) and then allocated across energy types, as shown in Table C9, according to the
shares described above.
WASTE MANAGEMENT AND REMEDIATION
Waste management and remediation are provided to the electricity generation sectors for
fossil fuels, nuclear, and biomass. Fossil fuels and nuclear require waste management services,
while the biomass includes production using waste products as an energy source.
WATER (PUBLIC AND PRIVATE)
Substantial amounts of water are used in the electricity generation process for all of the
energy sources except wind power. As shown in Table C10 (U.S. DoE, 2006),
80
the intensity of
water use in electricity generation varies greatly, both between and within energy sources. This

79
Again, this weight is not based on empirical data, but aims to reflect the lesser use of these sectors inputs
in the electricity generation sectors when compared with petrochemicals and industrial gases.
80
U.S. Department of Energy (2006) Energy Demands on Water Resources. Report to Congress on the
Interdependency of Energy and Water. Accessed February 29, 2012 http://www.sandia.gov/energy-
water/docs/121-RptToCongress-EWwEIAcomments-FINAL.pdf

262

makes identification of shares of water inputs to electricity generation by energy type difficult.
An estimate is calculated from the U.S. Department of Energy data (2006), which shows that, for
the same plant-types, nuclear power has about double the mid-point water intensity of fossil,
biomass, and waste power, while solar is about 1.5 times this intensity. Geothermal has about
four times the water intensity, while natural gas has about half the water intensity. An exact
comparison is not provided for hydro, but it is assumed that the intensity is equal to natural gas.
Similar adjustments are not necessary for the outputs of the electricity generation,
transmission, and distribution sector (the rows of the I-O table). It is reasonable to assume that the
electricity produced will be consumed by purchasing sectors according to the energy mix
determined by the state-energy type matrix.
Table C1: Electricity generation energy mix by sector (without adjusting for inter-state flows).

ELCL ELGS ELOL ELOF ELNU ELBM ELGT ELHY ELSL ELWD
ABEEF 0.4431 0.2242 0.0087 0.0027 0.1730 0.0167 0.0109 0.1040 0.0005 0.0160
ADARY 0.4414 0.2253 0.0088 0.0027 0.1722 0.0175 0.0109 0.1045 0.0005 0.0161
AOLVS 0.4414 0.2253 0.0088 0.0027 0.1722 0.0175 0.0109 0.1045 0.0005 0.0161
APOUL 0.4414 0.2253 0.0088 0.0027 0.1722 0.0175 0.0109 0.1045 0.0005 0.0161
AFISH 0.3055 0.3041 0.0122 0.0026 0.1666 0.0216 0.0167 0.1557 0.0008 0.0143
AOTH 0.3055 0.3041 0.0122 0.0026 0.1666 0.0216 0.0167 0.1557 0.0008 0.0143
CRUD 0.3803 0.4434 0.0117 0.0026 0.0958 0.0096 0.0053 0.0322 0.0003 0.0187
COAL 0.5728 0.2064 0.0079 0.0017 0.1247 0.0116 0.0053 0.0619 0.0002 0.0075
OMIN 0.4420 0.3940 0.0062 0.0025 0.1012 0.0106 0.0024 0.0241 0.0001 0.0169
CNSR 0.3830 0.2814 0.0190 0.0028 0.1955 0.0242 0.0090 0.0746 0.0004 0.0099
MFML 0.4790 0.1954 0.0087 0.0024 0.2146 0.0213 0.0065 0.0626 0.0003 0.0092
MOML 0.4790 0.1954 0.0087 0.0024 0.2146 0.0213 0.0065 0.0626 0.0003 0.0092
MANM 0.4790 0.1954 0.0087 0.0024 0.2146 0.0213 0.0065 0.0626 0.0003 0.0092
MPTY 0.4790 0.1954 0.0087 0.0024 0.2146 0.0213 0.0065 0.0626 0.0003 0.0092
MFSH 0.4790 0.1954 0.0087 0.0024 0.2146 0.0213 0.0065 0.0626 0.0003 0.0092
MOFD 0.4790 0.1954 0.0087 0.0024 0.2146 0.0213 0.0065 0.0626 0.0003 0.0092
MCHM 0.4345 0.2538 0.0095 0.0044 0.2224 0.0204 0.0058 0.0408 0.0003 0.0081
MPET 0.3310 0.3895 0.0071 0.0040 0.1626 0.0184 0.0145 0.0570 0.0007 0.0153
MOND 0.4511 0.2104 0.0097 0.0023 0.2332 0.0223 0.0073 0.0557 0.0004 0.0077
MPRM 0.5989 0.1449 0.0043 0.0045 0.1868 0.0140 0.0021 0.0392 0.0001 0.0051
MORD 0.5098 0.1741 0.0061 0.0034 0.1839 0.0167 0.0045 0.0935 0.0002 0.0076
MSEM 0.3978 0.2573 0.0089 0.0023 0.2008 0.0208 0.0099 0.0906 0.0005 0.0112
MODR 0.4765 0.2080 0.0082 0.0030 0.1995 0.0183 0.0059 0.0712 0.0003 0.0091

263

TAIR 0.3899 0.2766 0.0278 0.0021 0.1992 0.0200 0.0080 0.0653 0.0004 0.0107
TRUK 0.4760 0.2141 0.0088 0.0027 0.2025 0.0189 0.0062 0.0612 0.0003 0.0094
TWAT 0.3344 0.3059 0.0474 0.0045 0.1929 0.0258 0.0050 0.0772 0.0002 0.0066
TRAL 0.5553 0.1675 0.0074 0.0021 0.1814 0.0129 0.0028 0.0588 0.0001 0.0116
TOTH 0.3995 0.2973 0.0194 0.0025 0.1808 0.0182 0.0077 0.0637 0.0004 0.0104
TLTP 0.3331 0.2925 0.0237 0.0020 0.2313 0.0241 0.0083 0.0761 0.0004 0.0084
COMC 0.2560 0.3553 0.0237 0.0025 0.2052 0.0244 0.0222 0.0954 0.0011 0.0142
INFO 0.3307 0.2826 0.0262 0.0022 0.2066 0.0226 0.0109 0.1062 0.0005 0.0115
PELE 0.4441 0.2385 0.0084 0.0031 0.2018 0.0194 0.0075 0.0669 0.0004 0.0098
GASU 0.4441 0.2385 0.0084 0.0031 0.2018 0.0194 0.0075 0.0669 0.0004 0.0098
PWAT 0.4441 0.2385 0.0084 0.0031 0.2018 0.0194 0.0075 0.0669 0.0004 0.0098
SANT 0.4441 0.2385 0.0084 0.0031 0.2018 0.0194 0.0075 0.0669 0.0004 0.0098
WTRD 0.3934 0.2656 0.0148 0.0026 0.2128 0.0223 0.0083 0.0698 0.0004 0.0099
RTRD 0.3905 0.2680 0.0170 0.0027 0.2041 0.0233 0.0092 0.0749 0.0004 0.0098
REST 0.3401 0.2890 0.0244 0.0027 0.2114 0.0269 0.0114 0.0835 0.0005 0.0101
BANK 0.3986 0.2554 0.0209 0.0049 0.2113 0.0207 0.0074 0.0719 0.0003 0.0087
SECB 0.2826 0.3083 0.0336 0.0021 0.2529 0.0237 0.0079 0.0800 0.0004 0.0086
INSR 0.4145 0.2372 0.0186 0.0044 0.2278 0.0222 0.0058 0.0594 0.0003 0.0098
OODW 0.3821 0.2850 0.0178 0.0028 0.2069 0.0230 0.0087 0.0631 0.0004 0.0102
HOTR 0.3611 0.2986 0.0358 0.0026 0.1883 0.0224 0.0108 0.0710 0.0005 0.0091
PSRV 0.3862 0.2597 0.0338 0.0027 0.2031 0.0240 0.0088 0.0716 0.0004 0.0097
VSRV 0.3467 0.2873 0.0284 0.0025 0.2127 0.0269 0.0105 0.0744 0.0005 0.0101
WAST 0.3747 0.2632 0.0160 0.0026 0.2113 0.0232 0.0081 0.0905 0.0004 0.0098
OBSV 0.3612 0.2729 0.0318 0.0025 0.2161 0.0258 0.0092 0.0707 0.0004 0.0094
ENTR 0.3522 0.2972 0.0239 0.0031 0.1996 0.0231 0.0119 0.0783 0.0006 0.0101
EDUC 0.3597 0.2663 0.0382 0.0023 0.2240 0.0269 0.0072 0.0672 0.0003 0.0079
MEDC 0.4045 0.2531 0.0198 0.0027 0.2101 0.0240 0.0072 0.0694 0.0003 0.0090
OSOC 0.3756 0.2550 0.0281 0.0025 0.2145 0.0248 0.0073 0.0823 0.0003 0.0095
GELE 0.3989 0.2705 0.0208 0.0028 0.2044 0.0243 0.0074 0.0612 0.0004 0.0094
TLTG 0.3331 0.2925 0.0237 0.0020 0.2313 0.0241 0.0083 0.0761 0.0004 0.0084
GVUT 0.3644 0.2329 0.1058 0.0019 0.1712 0.0367 0.0057 0.0730 0.0003 0.0080
FGML 0.4002 0.2446 0.0483 0.0023 0.1851 0.0270 0.0077 0.0747 0.0003 0.0098
OGOV 0.3856 0.2581 0.0364 0.0025 0.1967 0.0261 0.0082 0.0763 0.0004 0.0097
SGGV 0.3888 0.2661 0.0178 0.0027 0.2045 0.0234 0.0089 0.0773 0.0004 0.0101
NCMP 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
Total 0.4958 0.2099 0.0102 0.0028 0.1955 0.0163 0.0035 0.0577 0.0002 0.0083
Source: Author‘s calculations based on US EPA eGRID data (electricity generation energy mix by state) and US BEA
Regional Economic Account data (sectoral output by state).

264

Table C2: Electricity generation energy mix by sector (adjusted for inter-state flows).

ELCL ELGS ELOL ELOF ELNU ELBM ELGT ELHY ELSL ELWD
ABEEF 0.4673 0.2111 0.0082 0.0024 0.1713 0.0154 0.0089 0.0995 0.0004 0.0155
ADARY 0.4673 0.2111 0.0082 0.0024 0.1713 0.0154 0.0089 0.0995 0.0004 0.0155
AOLVS 0.4673 0.2111 0.0082 0.0024 0.1713 0.0154 0.0089 0.0995 0.0004 0.0155
APOUL 0.4673 0.2111 0.0082 0.0024 0.1713 0.0154 0.0089 0.0995 0.0004 0.0155
AFISH 0.3297 0.2861 0.0115 0.0022 0.1597 0.0224 0.0134 0.1607 0.0006 0.0136
AOTH 0.3297 0.2861 0.0115 0.0022 0.1597 0.0224 0.0134 0.1607 0.0006 0.0136
CRUD 0.3926 0.4319 0.0116 0.0025 0.0945 0.0089 0.0042 0.0351 0.0002 0.0185
COAL 0.5847 0.1973 0.0075 0.0017 0.1267 0.0091 0.0047 0.0607 0.0002 0.0075
OMIN 0.4518 0.3840 0.0061 0.0024 0.1015 0.0098 0.0020 0.0255 0.0001 0.0169
CNSR 0.4156 0.2585 0.0163 0.0028 0.1980 0.0176 0.0074 0.0740 0.0003 0.0095
MFML 0.5001 0.1827 0.0080 0.0023 0.2136 0.0170 0.0053 0.0620 0.0003 0.0089
MOML 0.5001 0.1827 0.0080 0.0023 0.2136 0.0170 0.0053 0.0620 0.0003 0.0089
MANM 0.5001 0.1827 0.0080 0.0023 0.2136 0.0170 0.0053 0.0620 0.0003 0.0089
MPTY 0.5001 0.1827 0.0080 0.0023 0.2136 0.0170 0.0053 0.0620 0.0003 0.0089
MFSH 0.5001 0.1827 0.0080 0.0023 0.2136 0.0170 0.0053 0.0620 0.0003 0.0089
MOFD 0.5001 0.1827 0.0080 0.0023 0.2136 0.0170 0.0053 0.0620 0.0003 0.0089
MCHM 0.4616 0.2352 0.0086 0.0042 0.2206 0.0158 0.0046 0.0412 0.0002 0.0079
MPET 0.3634 0.3617 0.0062 0.0036 0.1574 0.0164 0.0115 0.0645 0.0006 0.0147
MOND 0.4675 0.1985 0.0090 0.0022 0.2297 0.0207 0.0058 0.0589 0.0003 0.0074
MPRM 0.6038 0.1401 0.0040 0.0045 0.1885 0.0128 0.0017 0.0394 0.0001 0.0051
MORD 0.5154 0.1684 0.0058 0.0033 0.1830 0.0166 0.0036 0.0963 0.0002 0.0075
MSEM 0.4207 0.2399 0.0081 0.0022 0.2002 0.0174 0.0079 0.0925 0.0004 0.0108
MODR 0.4901 0.1978 0.0077 0.0029 0.1983 0.0168 0.0048 0.0726 0.0002 0.0088
TAIR 0.4135 0.2595 0.0266 0.0022 0.1977 0.0162 0.0065 0.0674 0.0003 0.0102
TRUK 0.4955 0.2013 0.0081 0.0026 0.2015 0.0159 0.0050 0.0606 0.0002 0.0092
TWAT 0.3627 0.2841 0.0445 0.0043 0.1921 0.0226 0.0041 0.0790 0.0002 0.0065
TRAL 0.5681 0.1615 0.0038 0.0020 0.1834 0.0109 0.0023 0.0563 0.0001 0.0115
TOTH 0.4225 0.2805 0.0183 0.0025 0.1796 0.0146 0.0062 0.0654 0.0003 0.0101
TLTP 0.3668 0.2697 0.0190 0.0021 0.2323 0.0183 0.0069 0.0765 0.0003 0.0081
COMC 0.3035 0.3224 0.0166 0.0024 0.2002 0.0196 0.0177 0.1037 0.0009 0.0131
INFO 0.3666 0.2612 0.0144 0.0021 0.2066 0.0186 0.0087 0.1104 0.0004 0.0110
PELE 0.4637 0.2238 0.0077 0.0030 0.2004 0.0168 0.0060 0.0688 0.0003 0.0095
GASU 0.4637 0.2238 0.0077 0.0030 0.2004 0.0168 0.0060 0.0688 0.0003 0.0095
PWAT 0.4637 0.2238 0.0077 0.0030 0.2004 0.0168 0.0060 0.0688 0.0003 0.0095
SANT 0.4637 0.2238 0.0077 0.0030 0.2004 0.0168 0.0060 0.0688 0.0003 0.0095
WTRD 0.4222 0.2455 0.0129 0.0026 0.2122 0.0176 0.0067 0.0706 0.0003 0.0094
RTRD 0.4201 0.2474 0.0148 0.0027 0.2042 0.0183 0.0074 0.0753 0.0003 0.0094
REST 0.3817 0.2613 0.0184 0.0027 0.2135 0.0192 0.0091 0.0841 0.0004 0.0095
BANK 0.4249 0.2384 0.0147 0.0039 0.2150 0.0172 0.0061 0.0713 0.0003 0.0083
SECB 0.3112 0.2911 0.0239 0.0020 0.2540 0.0202 0.0063 0.0826 0.0003 0.0082
INSR 0.4378 0.2207 0.0153 0.0038 0.2301 0.0180 0.0047 0.0600 0.0002 0.0095
OODW 0.4154 0.2621 0.0150 0.0027 0.2062 0.0174 0.0070 0.0641 0.0003 0.0097
HOTR 0.3944 0.2764 0.0260 0.0025 0.1916 0.0178 0.0090 0.0731 0.0004 0.0088
PSRV 0.4279 0.2391 0.0149 0.0027 0.2089 0.0180 0.0071 0.0718 0.0003 0.0092

265

VSRV 0.3933 0.2607 0.0145 0.0026 0.2174 0.0187 0.0085 0.0745 0.0004 0.0095
WAST 0.4030 0.2426 0.0143 0.0026 0.2118 0.0180 0.0065 0.0915 0.0003 0.0094
OBSV 0.4057 0.2492 0.0151 0.0026 0.2211 0.0186 0.0074 0.0712 0.0004 0.0089
ENTR 0.3857 0.2732 0.0186 0.0030 0.2003 0.0185 0.0096 0.0810 0.0005 0.0095
EDUC 0.4006 0.2445 0.0189 0.0024 0.2329 0.0200 0.0057 0.0671 0.0003 0.0075
MEDC 0.4337 0.2338 0.0148 0.0027 0.2126 0.0186 0.0058 0.0692 0.0003 0.0086
OSOC 0.4067 0.2371 0.0190 0.0025 0.2174 0.0199 0.0059 0.0822 0.0003 0.0091
GELE 0.4315 0.2493 0.0152 0.0028 0.2079 0.0177 0.0060 0.0603 0.0003 0.0091
TLTG 0.3668 0.2697 0.0190 0.0021 0.2323 0.0183 0.0069 0.0765 0.0003 0.0081
GVUT 0.4757 0.2028 0.0173 0.0029 0.2080 0.0180 0.0047 0.0627 0.0002 0.0076
FGML 0.4399 0.2257 0.0337 0.0024 0.1898 0.0193 0.0064 0.0733 0.0003 0.0094
OGOV 0.4312 0.2360 0.0172 0.0026 0.2040 0.0182 0.0067 0.0746 0.0003 0.0092
SGGV 0.4191 0.2455 0.0149 0.0026 0.2050 0.0181 0.0072 0.0777 0.0003 0.0096
NCMP 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
Total 0.4952 0.2077 0.0098 0.0027 0.1960 0.0164 0.0037 0.0601 0.0002 0.0083
Source: Author‘s calculations based on US EPA eGRID data (electricity generation energy mix by state) and US BEA
Regional Economic Account data (sectoral output by state).
Table C3: Disaggregated Electricity Generation, Transmission, and Distribution columns of the
input-output table (as consuming sectors, $m).
Sector ELCL ELGS ELOL ELOF ELNU ELBM ELGT ELHY ELSL ELWD
ABEEF 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
ADARY 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
AOLVS 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
APOUL 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
AFISH 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
AOTH 0.00 0.00 0.00 0.00 0.00 0.13 0.00 0.00 0.00 0.00
COAL 16999 0.00 0.00 0.00 0.00 0.13 0.00 0.00 0.00 0.00
CRUD 0.00 0.00 19799 0.00 0.00 0.00 0.00 0.00 0.00 0.00
OMIN 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
CNSR 1978 1453.21 98.06 14.69 1009.53 125.21 46.55 385.15 2.18 51.31
MFML 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
MOML 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
MANM 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
MPTY 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
MFSH 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
MOFD 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
MCHM 79.55 46.47 1.74 0.81 42.76 1.18 0.33 2.36 0.02 0.47
MPET 0.00 0.00 4505.5 0.00 0.00 0.00 0.00 0.00 0.00 0.00
MOND 177.05 82.56 3.81 0.89 91.50 8.74 2.86 21.87 0.14 3.03
MPRM 90.08 21.80 0.65 0.68 28.09 2.10 0.32 5.90 0.01 0.77
MORD 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
MSEM 13.34 8.63 0.30 0.08 6.73 0.70 0.33 3.04 0.02 0.38
MODR 828.15 361.52 14.26 5.21 346.74 31.79 10.30 123.66 0.49 15.73
TAIR 20.73 14.71 1.48 0.11 10.59 1.07 0.42 3.47 0.02 0.57
TRUK 146.77 66.00 2.72 0.82 62.44 5.84 1.90 18.87 0.09 2.90

266

TWAT 76.32 69.80 10.82 1.02 44.02 5.90 1.14 17.62 0.05 1.52
TRAL 1354.15 408.43 18.10 5.11 442.45 31.56 6.78 143.31 0.32 28.41
TOTH 2101.12 1563.29 102.03 13.27 951.04 95.75 40.75 335.03 1.95 54.81
TLTP 2.64 2.32 0.19 0.02 1.83 0.19 0.07 0.60 0.00 0.07
COMC 18.02 25.00 1.67 0.18 14.44 1.71 1.56 6.72 0.08 1.00
INFO 99.77 85.25 7.91 0.65 62.33 6.81 3.30 32.05 0.16 3.48
PELE 1.70 0.91 0.03 0.01 0.77 0.07 0.03 0.26 0.00 0.04
GASU 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
PWAT 58.37 15.67 1.10 0.41 53.05 2.55 3.96 4.40 0.07 0.00
SANT 58.32 31.31 1.10 0.41 26.50 2.55 0.99 8.79 0.05 1.29
WTRD 226.55 152.96 8.54 1.51 122.55 12.87 4.78 40.18 0.23 5.69
RTRD 13.60 9.33 0.59 0.10 7.11 0.81 0.32 2.61 0.02 0.34
REST 280.34 238.25 20.10 2.24 174.29 22.21 9.36 68.80 0.45 8.33
BANK 343.25 219.91 18.02 4.20 181.96 17.80 6.40 61.90 0.30 7.45
SECB 75.02 81.86 8.92 0.55 67.14 6.29 2.10 21.24 0.10 2.28
INSR 95.17 54.45 4.27 1.01 52.30 5.10 1.33 13.64 0.06 2.26
OODW 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
HOTR 175.82 145.39 17.41 1.25 91.66 10.92 5.24 34.56 0.23 4.41
PSRV 13.27 8.92 1.16 0.09 6.98 0.82 0.30 2.46 0.01 0.33
VSRV 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
WAST 19.00 13.35 0.81 0.13 10.72 1.18 0.41 4.59 0.02 0.50
OBSV 1850.41 1397.95 162.74 12.87 1106.83 132.34 46.98 362.14 2.25 48.12
ENTR 10.73 9.05 0.73 0.09 6.08 0.70 0.36 2.39 0.02 0.31
EDUC 89.03 65.92 9.46 0.57 55.44 6.66 1.77 16.64 0.08 1.95
MEDC 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
OSOC 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
GELE 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
TLTG 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
GVUT 248.26 79.35 72.07 1.31 233.29 25.03 15.64 24.88 0.27 0.00
FGML 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
OGOV 0.91 0.61 0.09 0.01 0.46 0.06 0.02 0.18 0.00 0.02
SGGV 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
NCMP 1.55 0.66 0.03 0.01 0.61 0.05 0.01 0.18 0.00 0.03
LAB 23633.2 10005.5 484.68 131.30 21808.83 1571.66 340.03 1551.59 15.90 800.27
KPI 8727.94 3695.12 178.99 48.49 4297.13 377.86 81.75 323.83 3.82 192.40
KOI 79254.4 33553.7 1625.37 440.31 39020.23 3431.18 742.34 2940.54 34.71 1747.11
ITX 16312.2 6906.04 334.53 90.62 8031.17 706.21 152.79 605.22 7.14 359.59
EMPLY 242.25 102.56 4.97 1.35 178.11 15.88 3.43 19.17 0.16 8.08
OUTP 169515 71767 3476.44 941.76 66855.96 5559.46 1202.80 19729.37 56.24 2830.80

267

Table C4: Disaggregated Electricity Generation, Transmission, and Distribution rows of the
input-output table (as producing sectors, $m). NB – these rows are transposed into columns for
presentation purposes.

ELCL ELGS ELOL ELOF ELNU ELBM ELGT ELHY ELSL ELWD
ABEEF 152.57 68.92 2.69 0.78 55.92 5.02 2.92 32.49 0.14 5.07
ADARY 160.73 72.61 2.83 0.83 58.91 5.29 3.07 34.23 0.15 5.34
AOLVS 102.71 46.40 1.81 0.53 37.65 3.38 1.96 21.87 0.09 3.41
APOUL 104.25 47.09 1.84 0.54 38.21 3.43 1.99 22.20 0.10 3.46
AFISH 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
AOTH 127.26 110.45 4.46 0.86 61.64 8.64 5.18 62.03 0.25 5.26
COAL 79.69 87.67 2.35 0.50 19.18 1.80 0.86 7.13 0.04 3.77
CRUD 584.57 197.21 7.52 1.66 126.63 9.07 4.70 60.70 0.20 7.51
OMIN 305.93 260.05 4.11 1.63 68.72 6.64 1.32 17.29 0.06 11.43
CNSR 718.11 446.54 28.18 4.79 342.05 30.46 12.71 127.90 0.59 16.40
MFML 72.51 26.48 1.16 0.34 30.98 2.46 0.76 8.99 0.04 1.28
MOML 105.68 38.60 1.69 0.49 45.15 3.59 1.11 13.10 0.05 1.87
MANM 208.35 76.10 3.32 0.98 89.01 7.07 2.20 25.82 0.11 3.69
MPTY 164.08 59.93 2.62 0.77 70.10 5.57 1.73 20.34 0.08 2.91
MFSH 28.84 10.53 0.46 0.14 12.32 0.98 0.30 3.57 0.01 0.51
MOFD 1031.44 376.74 16.46 4.83 440.64 35.01 10.88 127.83 0.52 18.26
MCHM 5495.89 2800.78 102.90 50.39 2626.09 188.26 54.63 490.48 2.64 93.95
MPET 420.92 419.04 7.22 4.13 182.34 19.04 13.32 74.77 0.64 16.98
MOND 2301.16 976.96 44.39 10.65 1130.38 102.10 28.66 289.68 1.38 36.62
MPRM 3163.98 734.16 20.86 23.83 987.88 66.82 9.13 206.64 0.41 26.60
MORD 9.25 3.02 0.10 0.06 3.28 0.30 0.06 1.73 0.00 0.13
MSEM 905.01 516.02 17.46 4.74 430.79 37.36 17.04 198.98 0.83 23.19
MODR 4322.44 1744.62 67.54 25.85 1748.93 148.09 42.30 640.64 2.01 77.75
TAIR 52.28 32.81 3.37 0.27 24.99 2.04 0.83 8.52 0.04 1.29
TRUK 64.27 26.11 1.05 0.34 26.13 2.06 0.65 7.86 0.03 1.19
TWAT 5.25 4.11 0.64 0.06 2.78 0.33 0.06 1.14 0.00 0.09
TRAL 8.34 2.37 0.06 0.03 2.69 0.16 0.03 0.83 0.00 0.17
TOTH 286.86 190.43 12.44 1.67 121.96 9.93 4.23 44.42 0.20 6.84
TLTP 9.37 6.89 0.49 0.05 5.94 0.47 0.18 1.95 0.01 0.21
COMC 369.95 393.01 20.27 2.88 244.06 23.83 21.54 126.35 1.04 16.00
INFO 364.02 259.33 14.30 2.11 205.11 18.50 8.67 109.60 0.42 10.92
ELCL 0.79 0.38 0.01 0.01 0.34 0.03 0.01 0.12 0.00 0.02
ELGS 0.42 0.20 0.01 0.00 0.18 0.02 0.01 0.06 0.00 0.01
ELOL 0.01 0.01 0.00 0.00 0.01 0.00 0.00 0.00 0.00 0.00
ELOF 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
ELNU 0.36 0.17 0.01 0.00 0.16 0.01 0.00 0.05 0.00 0.01
ELBM 0.03 0.02 0.00 0.00 0.01 0.00 0.00 0.01 0.00 0.00
ELGT 0.01 0.01 0.00 0.00 0.01 0.00 0.00 0.00 0.00 0.00
ELHY 0.12 0.06 0.00 0.00 0.05 0.00 0.00 0.02 0.00 0.00
ELSL 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
ELWD 0.02 0.01 0.00 0.00 0.01 0.00 0.00 0.00 0.00 0.00
GASU 4.38 2.11 0.07 0.03 1.89 0.16 0.06 0.65 0.00 0.09

268

PWAT 14.64 7.07 0.24 0.10 6.33 0.53 0.19 2.17 0.01 0.30
SANT 65.71 31.72 1.09 0.43 28.40 2.38 0.86 9.75 0.04 1.35
WTRD 1456.18 846.86 44.35 8.88 732.02 60.58 23.03 243.51 1.10 32.55
RTRD 5138.10 3026.03 181.16 32.45 2497.62 223.34 90.57 921.45 4.28 114.47
REST 6184.85 4234.24 297.45 43.67 3458.78 311.56 148.09 1361.77 7.01 154.17
BANK 448.06 251.37 15.46 4.14 226.66 18.14 6.39 75.16 0.30 8.79
SECB 392.52 367.23 30.19 2.55 320.44 25.51 7.99 104.20 0.38 10.32
INSR 10.15 5.12 0.36 0.09 5.33 0.42 0.11 1.39 0.01 0.22
OODW 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
HOTR 2336.36 1637.27 153.83 15.09 1135.31 105.69 53.60 432.90 2.32 52.17
PSRV 470.69 263.05 16.39 2.98 229.80 19.78 7.81 79.02 0.37 10.13
VSRV 24.89 16.50 0.92 0.16 13.76 1.18 0.53 4.72 0.03 0.60
WAST 0.00 0.00 0.00 0.00 0.00 50.71 0.00 0.00 0.00 0.00
OBSV 3883.72 2385.95 144.28 24.64 2116.97 177.84 70.63 681.23 3.36 85.08
ENTR 487.71 345.45 23.48 3.75 253.28 23.45 12.20 102.39 0.57 12.06
EDUC 164.63 100.48 7.78 1.00 95.72 8.22 2.36 27.60 0.11 3.07
MEDC 1545.80 833.16 52.70 9.53 757.60 66.46 20.71 246.62 0.98 30.63
OSOC 224.78 131.02 10.50 1.36 120.15 11.00 3.28 45.40 0.15 5.01
TLTG 68.00 50.01 3.52 0.38 43.06 3.40 1.28 14.18 0.06 1.51
GVUT 475.99 202.95 17.33 2.86 208.11 18.03 4.70 62.78 0.21 7.56
FGML 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
OGOV 756.87 414.31 30.17 4.57 358.19 31.93 11.69 130.97 0.55 16.23
SGGV 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
NCMP 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
HH1 3913.88 1657.00 80.27 21.74 3611.74 260.28 56.31 256.96 2.63 132.53
HH2 2938.26 1243.96 60.26 16.32 2711.43 195.40 42.28 192.90 1.98 99.50
HH3 6761.25 2862.49 138.66 37.56 6239.30 449.64 97.28 443.89 4.55 228.95
HH4 5390.07 2281.98 110.54 29.95 4973.97 358.45 77.55 353.87 3.63 182.52
HH5 11771.37 4983.61 241.41 65.40 10862.65 782.82 169.36 772.82 7.92 398.60
HH6 15932.92 6745.47 326.76 88.52 14702.94 1059.57 229.24 1046.04 10.72 539.52
HH7 4150.41 1757.14 85.12 23.06 3830.01 276.01 59.72 272.49 2.79 140.54
HH8 8362.36 3540.34 171.50 46.46 7716.81 556.11 120.32 549.01 5.63 283.17
HH9 3435.32 1454.40 70.45 19.09 3170.12 228.46 49.43 225.54 2.31 116.33
FG_D 319.42 135.23 6.55 1.77 157.26 13.83 2.99 11.85 0.14 7.04
FG_ND 581.22 246.07 11.92 3.23 286.16 25.16 5.44 21.56 0.25 12.81
SG 23846.74 10095.91 489.05 132.48 11740.74 1032.40 223.36 884.77 10.44 525.69
ENT 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
IV 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
STK 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
MOW 398.48 168.70 8.17 2.21 196.19 17.25 3.73 14.78 0.17 8.78
MUS 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
EOW 357.09 151.18 7.32 1.98 175.81 15.46 3.34 13.25 0.16 7.87
EUS 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
DIN 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
OUTP 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
SUPPLY 156884.2 65799.7 3097.2 855.2 62111.1 5192.4 1171.6 19040.2 54.9 2619.6
DD 156478.9 65629.7 3089.2 853.0 61950.7 5179.0 1168.6 18991.0 54.8 2612.8

269

DD_MAK 156478.9 65629.7 3089.2 853.0 61950.7 5179.0 1168.6 18991.0 54.8 2612.8
DD_ DIF 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Output_
Row
169866.5 71244.7 3353.5 925.9 67250.9 5622.1 1268.6 20615.8 59.4 2836.4
DS 169503.3 71092.3 3346.4 924.0 67107.1 5610.0 1265.9 20571.8 59.3 2830.3
DS_MAK 169503.3 71092.3 3346.4 924.0 67107.1 5610.0 1265.9 20571.8 59.3 2830.3
DS_DIF 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

Table C5: Disaggregated Electricity Generation, Transmission, and Distribution rows and
columns of the make table (transposed).
Sector ELCL ELGS ELOL ELOF ELNU ELBM ELGT ELHY ELSL ELWD GASU PWAT Total
ELCL 156661 0 0 0 0 0 0 0 0 0 0 0 156661
ELGS 0 66325 0 0 0 0 0 0 0 0 22740 0 89065
ELOL 0 0 3213 0 0 0 0 0 0 0 0 0 3213
ELOF 0 0 0 870 0 0 0 0 0 0 0 0 870
ELNU 0 0 0 0 61787 0 0 0 0 0 0 0 61787
ELBM 0 0 0 0 0 5138 0 0 0 0 0 0 5138
ELGT 0 0 0 0 0 0 1112 0 0 0 0 0 1112
ELHY 0 0 0 0 0 0 0 18233 0 0 0 2453 20686
ELSL 0 0 0 0 0 0 0 0 52 0 0 0 52
ELWD 0 0 0 0 0 0 0 0 0 2616 0 0 2616
Total 156661 66325 3213 870 61787 5138 1112 18233 52 2616 N/A N/A

Table C6: Chemical manufacturing subsectors value of shipments and their relevance for the
adjustment of inputs to electricity sector generation.
NAICS Sector description
Value of
sales,
shipments...
($‘000s)
Proportion
of total
chemical
production
Relevance for adjustment of chemical
inputs to electricity sector generation
by energy
32511
Petrochemical
manufacturing

77,867,192 10.87%
Predominantly used by fossil fuel
electricity generation
32512
Industrial gas
manufacturing

9,305,261 1.30%
Predominantly used by fossil fuel
electricity generation
32513
Synthetic dye and
pigment manufacturing

7,983,968 1.11% Not used
32518
Other basic inorganic
chemical manufacturing

29,631,955 4.14%
Predominantly nuclear electricity
generation
32519
Other basic organic
chemical manufacturing

101,340,819 14.15% Not significantly used
32521
Resin and synthetic
rubber manufacturing

92,697,081 12.94% No weighting across energy types
32522
Artificial and synthetic
fibers and filaments
manufacturing

8,168,386 1.14% No weighting across energy types
32531 Fertilizer manufacturing 17,783,586 2.48% Not used
32532
Pesticide and other
agricultural chemical
manufacturing

11,371,897 1.59% Not used

270

32541
Pharmaceutical and
medicine manufacturing

188,399,027 26.31% Not used
32551
Paint and coating
manufacturing

23,580,841 3.29% No weighting across energy types
32552 Adhesive manufacturing 10,489,481 1.46% No weighting across energy types
32561
Soap and cleaning
compound manufacturing

48,370,396 6.75% Not significantly used
32562
Toilet preparation
manufacturing

42,975,236 6.00% Not significantly used
32591
Printing ink
manufacturing

5,053,738 0.71% Not significantly used
32592
Explosives
manufacturing

1,745,029 0.24% Not used
32599
All other chemical
product and preparation
manufacturing

39,388,580 5.50% Not significantly used

Table C7: Proportions of chemical inputs to electricity generation by estimation category.
Category
Weighted value of
sales, shipments…
($‘000s)
Proportion
of total
Estimated values of chemical
inputs to electricity generation
($m)
Fossil fuel 87,172,453 50.03% 87.89
Nuclear 29,631,955 17.01% 29.88
No weighting across
energy types 33,733,947 19.36% 34.01
Not significant 23,712,877 13.61% 23.91
Total 174,251,232

Table C8: Proportions of chemical inputs to electricity generation energy type by estimation
category.
ELCL ELGS ELOL ELOF ELNU ELBM ELGT ELHY ELSL ELWD
State-energy
shares 0.4345 0.2538 0.0095 0.0044 0.2224 0.0204 0.0058 0.0408 0.0003 0.0081
Pre-
adjustment
allocation 76.34 44.59 1.67 0.78 39.08 3.58 1.01 7.17 0.05 1.43
Fossil fuel 54.38 31.77 1.19 0.56
Nuclear 29.88
No
weighting
across types 14.78 8.63 0.32 0.15 7.56 0.69 0.20 1.39 0.01 0.28
Not
significant 10.39 6.07 0.23 0.11 5.32 0.49 0.14 0.98 0.01 0.19
Post-
adjustment
allocation 79.55 46.47 1.74 0.81 42.76 1.18 0.33 2.36 0.02 0.47

271

Table C9: Adjustments made to fuel inputs.
Before Fuel Input adjustments
ELCL ELGS ELOL ELOF ELNU ELBM ELGT ELHY ELSL ELWD
AOTH 0.04 0.04 0.00 0.00 0.02 0.00 0.00 0.02 0.00 0.00
COAL 6465.24 7537.96 199.71 44.09 1627.88 163.92 90.27 546.86 4.39 318.65
CRUD 11341.44 4086.87 156.41 32.69 2468.28 229.01 104.99 1225.31 4.52 149.16
CNSR 1977.68 1453.21 98.06 14.69 1009.53 125.21 46.55 385.15 2.18 51.31
MCHM 76.34 44.59 1.67 0.78 39.08 3.58 1.01 7.17 0.05 1.43
MPET 1491.24 1754.91 31.86 18.06 732.53 82.71 65.24 256.90 3.15 68.92
WAST 19.00 13.35 0.81 0.13 10.72 1.18 0.41 4.59 0.02 0.50
GVUT 255.10 163.07 74.06 1.34 119.86 25.72 4.02 51.12 0.19 5.61
PWAT 61.99 33.29 1.17 0.44 28.17 2.71 1.05 9.34 0.05 1.37
After Fuel Input Adjustments

ELCL ELGS ELOL ELOF ELNU ELBM ELGT ELHY ELSL ELWD
AOTH

0.13

COAL 16998.97

CRUD

19798.67

MCHM 79.55 46.47 1.74 0.81 42.76 1.18 0.33 2.36 0.02 0.47
MPET

4505.53

WAST

50.71

GVUT 248.26 79.35 72.07 1.31 233.29 25.03 15.64 24.88 0.27 0.00
PWAT 58.37 15.67 1.10 0.41 53.05 2.55 3.96 4.40 0.07 0.00

Table C10: Water use by Thermoelectric Power Plants
(U.S. Dept of Energy, 2006)

272

APPENDIX D: SENSITIVITY TESTS FOR EQUITY ANALYSIS
Table D1: Comparison of broad-based and narrow-based policy results across emissions cap
levels. Level ($Bn) and percent changes to household total income distribution by bracket and
Gini coefficients.

Income
brackets
Base
case
Case 2: Cons 70 Case 3: No Cons
Emissions cap levels Emissions cap levels
3090 (3.7) 2990 (6.7) 2890 (9.7) 3090 (3.7) 2990 (6.7) 2890 (9.7)
Level %Δ Level %Δ Level %Δ Level %Δ Level %Δ Level %Δ
LT10k 142.8 133.4 -6.6 133.2 -6.7 129.4 -9.4 135.1 -5.4 131.1 -8.2 127.3 -10.8
10-15k 147.7 131.6 -10.9 131.1 -11.2 122.3 -17.2 135.5 -8.3 126.4 -14.4 117.8 -20.3
15-25k 412.6 355.1 -13.9 352.9 -14.4 313.2 -24.1 373.0 -9.6 331.7 -19.6 293.6 -28.8
25-35k 493.4 443.6 -10.1 441.9 -10.4 411.0 -16.7 457.2 -7.3 425.3 -13.8 395.6 -19.8
35-50k 730.1 682.4 -6.5 680.6 -6.8 648.1 -11.2 696.8 -4.6 663.0 -9.2 631.7 -13.5
50-75k 1105.7 1064.9 -3.7 1062.9 -3.9 1025.5 -7.3 1082.1 -2.1 1043.0 -5.7 1007.2 -8.9
75-100k 942.0 916.5 -2.7 915.1 -2.8 890.9 -5.4 928.6 -1.4 902.9 -4.1 879.9 -6.6
100-150k 839.5 832.5 -0.8 831.3 -1.0 810.3 -3.5 842.8 0.4 820.5 -2.3 800.6 -4.6
150k+ 3753.6 3776.2 0.6 3776.6 0.6 3789.8 1.0 3772.4 0.5 3781.6 0.7 3798.6 1.2
Total 8567.3 8336.3 -2.7 8325.7 -2.8 8140.5 -5.0 135.1 -5.4 131.1 -8.2 127.3 -10.8
Gini
coeff
Level Level Δ Level Δ Level Δ Level Δ Level Δ Level Δ
0.5684 0.5834 0.0150 0.5840 0.0156 0.5939 0.0255 0.5792 0.0108 0.5892 0.0208 0.5991 0.0307

Table D2: Comparison of broad-based and results across emissions cap levels (6%-16%). Level
($Bn) and percent changes to household total income by bracket, Gini coefficients.

Income
brackets
Base
case
Emissions cap levels (% reduction) Emissions cap levels (% reduction)
6% 8% 10% 12% 14% 16%
Level %Δ Level %Δ Level %Δ Level %Δ Level %Δ Level %Δ
LT10k 142.8 140.3 -1.7 137.8 -3.5 135.1 -5.4 132.5 -7.2 129.9 -9.0 127.5 -10.7
10-15k 147.7 146.9 -0.5 141.3 -4.3 135.2 -8.5 129.2 -12.5 123.3 -16.5 117.7 -20.3
15-25k 412.6 426.1 3.3 399.9 -3.1 371.2 -10.0 344.0 -16.6 317.6 -23.0 293.0 -29.0
25-35k 493.4 498.5 1.0 478.3 -3.1 456.1 -7.6 435.0 -11.8 414.6 -16.0 395.4 -19.9
35-50k 730.1 740.7 1.5 719.2 -1.5 695.8 -4.7 673.5 -7.7 652.0 -10.7 631.7 -13.5
50-75k 1105.7 1132.9 2.5 1107.6 0.2 1080.3 -2.3 1054.6 -4.6 1029.9 -6.9 1006.8 -8.9
75-100k 942.0 961.8 2.1 944.7 0.3 926.5 -1.6 909.6 -3.4 893.6 -5.1 878.9 -6.7
100-150k 839.5 872.0 3.9 857.1 2.1 841.3 0.2 826.6 -1.5 812.8 -3.2 800.0 -4.7
150k+ 3753.6 3779.4 0.7 3776.1 0.6 3777.3 0.6 3782.6 0.8 3791.7 1.0 3803.8 1.3
Total 8567.3 8698.6 1.5 8561.9 -0.1 8418.8 -1.7 8287.6 -3.3 8165.4 -4.7 8054.8 -6.0
Gini
coeff
Level Level Δ Level Δ Level Δ Level Δ Level Δ Level Δ
0.5684 0.5677 -0.0007 0.5733 0.0049 0.5798 0.0114 0.5863 0.0179 0.5929 0.0245 0.5994 0.0310

273

Table D3: Comparison of broad-based and narrow-based policy results across emissions cap
levels. Level ($Bn) and percent changes to household disposable income by bracket and Gini
coefficients.

Income
brackets
Base
case
Case 2: Cons 70 Case 3: No Cons
Emissions cap levels Emissions cap levels
3090 2990 2890 3090 2990 2890
Level %Δ Level %Δ Level %Δ Level %Δ Level %Δ Level %Δ
LT10k 545.1 512.5 -6.0 511.8 -6.1 497.8 -8.7 518.6 -4.9 504.1 -7.5 490.6 -10.0
10-15k 396.1 368.8 -6.9 367.9 -7.1 352.0 -11.1 375.9 -5.1 359.5 -9.3 344.1 -13.1
15-25k 697.7 634.4 -9.1 632.0 -9.4 588.4 -15.7 654.2 -6.2 608.7 -12.8 566.9 -18.7
25-35k 713.2 665.0 -6.8 663.4 -7.0 633.3 -11.2 678.5 -4.9 647.2 -9.3 618.4 -13.3
35-50k 1090.0 1035.3 -5.0 1033.2 -5.2 996.4 -8.6 1051.8 -3.5 1013.3 -7.0 978.0 -10.3
50-75k 1210.2 1160.0 -4.1 1158.0 -4.3 1122.3 -7.3 1176.4 -2.8 1139.2 -5.9 1105.0 -8.7
75-100k 978.6 943.3 -3.6 941.9 -3.8 916.9 -6.3 955.6 -2.4 929.2 -5.0 905.3 -7.5
100-150k 613.5 606.7 -1.1 605.9 -1.2 591.2 -3.6 614.0 0.1 598.3 -2.5 584.4 -4.8
150k+ 3057.0 3071.9 0.5 3072.1 0.5 3080.6 0.8 3069.8 0.4 3075.0 0.6 3086.6 1.0
Gini
coeff
Level Level Δ Level Δ Level Δ Level Δ Level Δ Level Δ
0.4121 0.4276 0.0155 0.4281 0.0160 0.4374 0.0253 0.4237 0.0116 0.4329 0.0208 0.4422 0.0302

Table D4: Comparison of broad-based policy results across emissions cap levels (6-12%). Level
($Bn) and percent changes to household disposable income by bracket and Gini coefficients.

Income
brackets
Base
case
Emissions cap levels (% reduction) Emissions cap levels (% reduction)
6% 8% 10% 12% 14% 16%
Level %Δ Level %Δ Level %Δ Level %Δ Level %Δ Level %Δ
LT10k 545.1 537.4 -1.4 528.4 -3.1 518.4 -4.9 508.9 -6.6 499.7 -8.3 491.0 -9.9
10-15k 396.1 396.8 0.2 386.5 -2.4 375.3 -5.3 364.5 -8.0 353.9 -10.7 344.0 -13.2
15-25k 697.7 713.1 2.2 683.9 -2.0 652.3 -6.5 622.2 -10.8 593.3 -15.0 566.3 -18.8
25-35k 713.2 719.5 0.9 699.3 -1.9 677.5 -5.0 656.8 -7.9 636.9 -10.7 618.3 -13.3
35-50k 1090.0 1102.3 1.1 1077.5 -1.2 1050.6 -3.6 1025.2 -5.9 1000.8 -8.2 978.0 -10.3
50-75k 1210.2 1224.3 1.2 1200.3 -0.8 1174.5 -3.0 1150.0 -5.0 1126.4 -6.9 1104.3 -8.8
75-100k 978.6 989.1 1.1 971.8 -0.7 953.4 -2.6 936.1 -4.3 919.5 -6.0 904.2 -7.6
100-150k 613.5 634.5 3.4 624.0 1.7 612.9 -0.1 602.6 -1.8 592.9 -3.4 584.0 -4.8
150k+ 3057.0 3078.1 0.7 3074.2 0.6 3073.6 0.5 3076.4 0.6 3082.3 0.8 3090.7 1.1
Gini
coeff
Level Level Δ Level Δ Level Δ Level Δ Level Δ Level Δ
0.4121 0.4132 0.0011 0.4182 0.0062 0.4242 0.0121 0.4302 0.0181 0.4364 0.0243 0.4425 0.0304

274

Table D5: Gini coefficients for household consumption of commodity groups by income bracket.
Scenarios across Case 2 (electricity efficiency improvements assumed to average 3.5% per year)
and Case 3 (no conservation).

Case 2: Cons 70 Case 3: No Cons

3090 (3.7) 2990 (6.7) 2890 (9.7) 3090 (3.7) 2990 (6.7) 2890 (9.7)

Base Level Δ Level Δ Level Δ Level Δ Level Δ Level Δ
FOOD 0.236 0.247 0.011 0.248 0.012 0.255 0.019 0.244 0.008 0.251 0.015 0.259 0.023
HOUS 0.355 0.373 0.017 0.373 0.018 0.385 0.030 0.368 0.012 0.379 0.024 0.391 0.036
GASO 0.243 0.262 0.019 0.263 0.020 0.277 0.034 0.257 0.014 0.270 0.027 0.284 0.041
LTRN 0.289 0.305 0.017 0.306 0.017 0.317 0.029 0.301 0.012 0.312 0.023 0.324 0.035
OTRA 0.435 0.452 0.016 0.452 0.017 0.464 0.029 0.447 0.012 0.458 0.023 0.471 0.035
HLTH 0.370 0.365 -0.005 0.365 -0.005 0.363 -0.007 0.367 -0.003 0.364 -0.006 0.362 -0.008
EQUP 0.456 0.474 0.018 0.475 0.019 0.488 0.031 0.469 0.013 0.481 0.025 0.495 0.038
WTER 0.222 0.226 0.004 0.226 0.004 0.228 0.006 0.225 0.003 0.227 0.005 0.230 0.007
ELEC 0.243 0.259 0.015 0.259 0.016 0.269 0.025 0.255 0.012 0.264 0.021 0.275 0.032
OFUE 0.337 0.348 0.010 0.348 0.011 0.355 0.018 0.345 0.008 0.352 0.014 0.359 0.022
OTHG 0.334 0.355 0.020 0.355 0.021 0.370 0.036 0.349 0.015 0.363 0.028 0.378 0.044
OTHS 0.470 0.507 0.037 0.508 0.038 0.539 0.069 0.495 0.025 0.524 0.054 0.558 0.088
Total 0.411 0.436 0.025 0.437 0.026 0.456 0.045 0.429 0.018 0.447 0.036 0.468 0.057

Table D6: Gini coefficients for household consumption of commodity groups by income bracket,
across emissions reductions (6-16%).

Emissions cap levels (% reduction) Emissions cap levels (% reduction)

6% 8% 10% 12% 14% 16%

Base Level Δ Level Δ Level Δ Level Δ Level Δ Level Δ
FOOD 0.236 0.238 0.001 0.241 0.005 0.245 0.009 0.250 0.013 0.255 0.019 0.260 0.024
HOUS 0.355 0.355 -0.001 0.361 0.006 0.368 0.013 0.376 0.020 0.384 0.028 0.392 0.036
GASO 0.243 0.242 -0.001 0.249 0.006 0.257 0.014 0.266 0.023 0.275 0.032 0.285 0.042
LTRN 0.289 0.289 0.000 0.295 0.006 0.301 0.013 0.308 0.020 0.316 0.027 0.324 0.036
OTRA 0.435 0.435 -0.001 0.440 0.005 0.447 0.012 0.455 0.020 0.463 0.028 0.471 0.036
HLTH 0.370 0.373 0.003 0.370 -0.001 0.367 -0.003 0.364 -0.006 0.363 -0.007 0.361 -0.009
EQUP 0.456 0.457 0.001 0.463 0.007 0.470 0.014 0.478 0.022 0.486 0.030 0.495 0.039
WTER 0.222 0.223 0.001 0.224 0.002 0.225 0.003 0.227 0.004 0.228 0.006 0.230 0.007
ELEC 0.243 0.247 0.004 0.251 0.008 0.256 0.013 0.262 0.019 0.269 0.025 0.276 0.033
OFUE 0.337 0.338 0.001 0.341 0.004 0.345 0.008 0.350 0.013 0.355 0.017 0.359 0.022
OTHG 0.334 0.335 0.001 0.342 0.008 0.350 0.016 0.359 0.024 0.368 0.034 0.379 0.045
OTHS 0.470 0.469 -0.001 0.481 0.011 0.497 0.027 0.515 0.045 0.536 0.066 0.559 0.089
Total 0.411 0.411 0.000 0.419 0.008 0.430 0.019 0.441 0.030 0.454 0.043 0.468 0.057

275

APPENDIX E: MULTI-SECTOR INCOME DISTRIBUTION MATRIX
The US Multi-Sector Income Distribution Matrix (MSIDM) is a central data set for the
USCGE model. The MSIDM presented below in Table 5 displays the income distribution of
payments (labor and capital income) from 169 sectors to households. The MSIDM is constructed
using data from the US Bureau of Labor Statistics and the US Internal Revenue Service. First,
labor and capital matrices are calculated separately before being summed to identify the total
MSIDM. It is important to note that labor and capital income does not account for all household
income. As shown in Table E9 below, ―constant income‖ – which refers to transfers and pensions
and annuities – accounts for around 15 percent of income. However, such income is not tied
directly to sector output, and hence is unlikely to change significantly in response to an ETP. As
such it is treated separately to labor and capital income in the modeling process.
LABOR MSIDM
The labor MSIDM takes data from US Bureau of Labor Statistics on the distribution of
wages for each sector and each occupation within that sector. The distribution is provided in
terms of the 10
th
, 25
th
, 50
th
, 75
th
, and 90
th
percentile of labor income. To estimate the labor income
distribution for each sector, first the 1
st
, 5
th
, 20
th
, 40
th
, 60
th
, 80
th
, 95
th
, and 99
th
percentile are
estimated through interpolation. This approach aims to overcome limitations with the original
data set such as the highest possible income being capped at $145600 per annum. This approach
also increases accuracy in the next step of the process, whereby the percentile income points are
allocated to the relevant brackets as a weighted share (e.g. the 1
st
percentile is assumed to account
for 3 percent, the 5
th
percentile is assumed to account for 4.5 percent, etc) of total employment
income within that sector-occupation.

276

The relevant brackets for this allocation process are labor income brackets as opposed to
total income brackets. As shown in Tables 8 and 9, the labor shares of total income vary across
brackets; in Table 10 these labor shares are used to estimate the labor income brackets for the
allocation procedure. The labor income brackets for the two lowest brackets are problematic
because they overlap – this is caused by individuals in the lowest bracket experiencing significant
capital income losses, as well as individuals in the second lowest bracket receiving significant
―constant‖ income such as retirement payments and government transfers (Tables E8 and E9).
For the allocation procedure, the labor income shares for the lowest income bracket are instead
calculated by setting negative income to zero; as labor income accounts for 62.37 percent of
positive income, the labor income lowest bracket upper limit is assumed to be $6,237.
CAPITAL INCOME
The capital income takes data from the US IRS Statistics of Income, and combines it with
IMPLAN data to estimate the distribution of income across sectors. This calculation process is
conducted in terms of various types of capital income, including Proprietary income, Dividend
and other investment capital income, household related capital income (such as rent), farm
income, and royalties income from sources such as Mining and Publishing.
RAS MATRIX BALANCING PROCEDURE FOR THE TOTAL INCOME SECTOR-
BRACKET MATRIX
Both the labor and capital MSIDM are balanced using the RAS biproportional matrix
weighting procedure. The RAS procedure forces the sum of row and column elements of a matrix
(in this case the pre-tax and pre-transfer income distribution matrix by sector and income bracket)
to conform to control totals included in the algorithm code. These control totals act as anchors, so
that the adjusted matrix is consistent with the macroeconomic data. As such, the control totals

277

must be external to the original matrix estimation process. In other words, the control totals
cannot be used to estimate the original matrix and then also to anchor the row and column sums.
The row control totals – the sums of income payments by income bracket – are the
combination of data from a number of sources. IRS Statistics of Income data is used as the row
control totals for both labor and capital. Negative income is reported for the lowest capital income
bracket; this value is set to zero in the RAS balancing procedure as the algorithm does not accept
negative values as inputs.
The column control totals – i.e. the sums of income payments by sector – are drawn from
the IMPLAN social accounting matrix, yet scaled down to reflect total household income, and
adjusted to match capital and labor income shares for the economy. However, the row control
totals – the sums of income payments by income bracket – are only presented indirectly within
the 2007 California social accounting matrix (IMPLAN, 2007); instead, the row control totals
must be calculated using other data from the matrix.
Tables 1 and 2 provide some analysis of the highest income bracket ($150k+) in the
MSIDM. In previous distributional analysis, the proportion of total income going to the highest
income bracket played an important role in the overall distributional impact of the policy. Where
income is clustered in the highest income bracket (Table 1), both high-skilled labor (Medicine,
Banking) and capital (Real estate, Households) can play a notable role. Sectors characterized by
low-skilled labor are prominent when income is clustered outside of the highest income bracket
(Table 2).

278

Table E1: Top ten largest proportions of payments (labor and capital income)
from sectors to households.
Sector
% of sector
total income
going to highest
bracket Labor Capital Total
Offices of health practitioners 0.9718 244,194 24,859 269,053
Home health care services 0.9395 27,093 3,790 30,883
Private households 0.9190 152 432,058 432,210
Outpatient, laboratory, and other
ambulatory care services 0.9169 47,579 21,668 69,248
Securities, commodity contracts, and
other financial investments and related
activities 0.8142 159,257 0 159,257
Internet and other information services 0.7668 28,331 17,074 45,405
Real estate 0.7656 37,326 610,489 647,816
Lessors of nonfinancial intangible assets 0.7550 994 28,099 29,093
Oil and gas extraction 0.7102 26,338 46,450 72,788
Tobacco manufacturing 0.6709 337 6,758 7,095

Table E2: Top ten smallest proportions of payments (labor and capital income)
from sectors to households.
Sector
% of sector
total income
going to highest
bracket Labor Capital Total
Death care services 0.0519 337 0 337
Support activities for agriculture and
forestry 0.0626 879 0 879
Footwear manufacturing 0.0758 30 22 52
Individual, family, community, and
vocational rehabilitation services 0.1010 2,773 1,808 4,582
Fiber, yarn, and thread mills 0.1043 153 7 160
Leather, hide tanning, finishing; Other
leather, allied product manufacturing 0.1046 38 48 86
Museums, historical sites, and similar
institutions 0.1284 550 0 550
Textile and fabric finishing and fabric
coating mills 0.1392 264 57 322
Fabric mills 0.1510 434 71 505
Food services and drinking places 0.1569 3,320 28,200 31,520

279

Table E3: MSIDM (labor and capital) in USCGE sector scheme, ($2007m)

LT10k 10-15k 15-25k 25-35k 35-50k 50-60k 60-75k 75-100k 100-150k 150k+ Total
ABEEF 0 2,148 177 110 199 113 188 293 311 1,795 5,334
ADARY 0 1,646 402 250 452 257 429 666 707 4,086 8,894
AOLVS 0 1,981 264 164 297 169 282 438 464 2,683 6,742
APOUL 0 2,416 114 71 128 73 121 188 199 1,152 4,461
AFISH 751 25 26 85 187 70 137 94 121 1,292 2,788
AOTH 3,804 22,714 7,201 4,557 3,681 1,652 2,825 3,512 3,653 25,230 78,829
COAL 1,371 14 76 330 1,027 823 1,184 1,217 629 3,233 9,904
CRUD 295 3,294 1,250 2,810 6,186 5,089 8,051 13,755 13,113 100,818 154,661
OMIN 2,524 117 156 568 1,674 1,335 1,916 2,100 1,279 9,319 20,988
CNSR 63 6,471 7,845 30,577 64,842 48,515 63,962 100,822 63,911 155,759 542,766
MFML 1 2 338 612 658 302 333 314 224 1,135 3,919
MOML 1 2 393 712 765 351 386 361 253 1,217 4,441
MANM 2 3 1,098 1,997 2,137 971 1,055 950 615 1,887 10,714
MPTY 1 2 755 1,373 1,469 667 724 651 419 1,257 7,318
MFSH 0 0 166 302 323 147 159 143 93 280 1,613
MOFD 6 47 4,001 7,190 7,790 3,622 4,047 3,977 3,085 20,551 54,317
MCHM 8,829 1,242 1,537 3,428 7,625 6,376 10,739 16,138 14,940 83,750 154,605
MPET 2,318 150 482 725 1,886 1,781 2,953 5,854 5,475 35,398 57,023
MOND 14,248 172 6,964 14,302 22,807 12,609 16,864 19,304 12,750 63,012 183,032
MPRM 6,248 221 459 2,293 5,312 3,312 4,676 4,834 3,307 19,242 49,903
MORD 593 9 144 669 1,383 890 1,486 2,496 1,703 5,183 14,554
MSEM 8,791 26 1,233 4,316 9,231 5,458 9,198 14,884 20,557 71,176 144,868
MODR 35,284 1,640 8,955 29,376 56,766 32,326 45,205 59,760 45,456 170,335 485,104
TAIR 974 420 472 1,331 2,836 2,430 4,046 4,344 2,995 20,405 40,254
TRUK 1,316 118 1,696 6,343 17,840 10,395 15,144 18,314 3,960 23,275 98,400
TWAT 38 79 85 184 457 400 558 771 997 3,845 7,414
TRAL 894 21 89 138 937 1,228 2,506 3,659 3,155 10,646 23,274
TOTH 2,664 19,172 3,421 8,368 13,141 7,250 11,327 11,330 5,812 20,573 103,059
TLTP 985 6 1,578 2,277 3,725 1,450 786 521 439 2,218 13,985
COMC 1,182 2,214 2,535 5,265 9,190 9,520 16,597 31,350 26,251 132,442 236,546
INFO 3,484 538 3,932 4,123 7,372 4,771 9,567 15,386 20,683 118,470 188,325
PELE 1,661 838 924 1,471 3,956 3,728 8,001 17,668 15,382 83,258 136,888
GASU 1,525 186 171 297 849 813 1,785 3,978 3,296 13,790 26,691
PWAT 565 17 14 25 72 70 154 345 279 964 2,504
SANT 565 17 14 25 72 70 154 345 279 964 2,504

280

WTRD 7 12,411 9,916 22,429 41,246 25,430 32,834 52,493 53,218 224,557 474,541
RTRD 1,277 9,936 119,820 97,383 73,703 24,543 26,722 26,773 24,965 140,154 545,277
REST 129 4,323 9,553 14,974 22,337 12,224 16,627 28,908 33,576 357,182 499,833
BANK 1,289 1,252 5,114 13,930 27,199 12,988 23,357 29,612 35,569 248,865 399,173
SECB 1,090 286 363 1,206 3,857 2,891 6,358 9,404 10,882 159,257 195,593
INSR 92 2,543 1,232 6,445 14,661 14,009 15,692 30,017 24,503 95,656 204,849
OODW 0 947 2,737 2,888 5,208 4,113 6,868 13,312 19,583 290,634 346,288
HOTR 794 288 105,925 46,042 26,489 7,483 8,377 7,426 6,480 46,639 255,944
PSRV 3,265 7,915 7,580 6,384 5,623 2,307 3,194 2,586 2,629 20,181 61,666
VSRV 0 49 89 194 464 389 721 1,098 1,505 6,152 10,660
WAST 1,150 24 538 1,421 3,621 1,764 3,024 2,902 1,953 9,094 25,490
OBSV 23,210 29,275 58,853 82,271 114,025 70,424 109,203 160,550 190,999 1,196,391 2,035,202
ENTR 1,761 12,440 12,873 9,772 8,651 3,409 5,521 6,735 5,974 28,246 95,382
EDUC 1,320 2,847 13,530 25,030 63,736 47,444 81,137 110,651 97,752 136,346 579,793
MEDC 712 457 4,113 11,625 22,639 19,294 28,187 53,981 51,433 420,438 612,879
OSOC 1,765 2,728 23,017 29,826 25,628 11,509 13,706 15,768 9,114 24,327 157,387
GELE 35 275 544 825 1,677 1,348 2,103 2,765 2,566 3,593 15,730
TLTG 39 -540 -510 -194 719 285 1,096 1,670 1,174 1,637 5,376
GVUT 0 0 0 0 0 0 0 0 0 0 0
FGML 2,318 3,747 7,199 10,709 21,370 17,287 26,726 35,059 32,680 45,763 202,858
OGOV 2,330 997 5,625 12,334 32,395 24,121 41,898 56,599 50,061 70,054 296,412
SGGV 1,103 2,854 10,158 20,109 49,934 37,766 64,192 86,270 77,017 107,789 457,192
NCMP 0 0 0 0 0 0 0 0 0 0 0
Total 144,670 163,023 457,236 552,265 822,451 510,058 765,087 1,095,341 1,010,423 4,843,593 10,364,146

Table E4: MSIDM (labor and capital) in USCGE sector scheme coefficients, (income bracket
proportion of total sector labor and capital income payments to households).

LT10k 10-15k 15-25k 25-35k 35-50k 50-60k 60-75k 75-100k 100-150k 150k+ Total
ABEEF 0.0000 0.4028 0.0332 0.0206 0.0373 0.0212 0.0353 0.0549 0.0582 0.3365 1.0000
ADARY 0.0000 0.1851 0.0452 0.0281 0.0508 0.0289 0.0482 0.0749 0.0794 0.4594 1.0000
AOLVS 0.0000 0.2938 0.0392 0.0244 0.0440 0.0250 0.0418 0.0649 0.0688 0.3980 1.0000
APOUL 0.0000 0.5416 0.0255 0.0158 0.0286 0.0163 0.0271 0.0422 0.0447 0.2582 1.0000
AFISH 0.2695 0.0090 0.0094 0.0304 0.0672 0.0250 0.0490 0.0338 0.0433 0.4633 1.0000
AOTH 0.0483 0.2881 0.0913 0.0578 0.0467 0.0210 0.0358 0.0446 0.0463 0.3201 1.0000
COAL 0.1385 0.0014 0.0077 0.0333 0.1037 0.0831 0.1196 0.1229 0.0635 0.3264 1.0000
CRUD 0.0019 0.0213 0.0081 0.0182 0.0400 0.0329 0.0521 0.0889 0.0848 0.6519 1.0000

281

OMIN 0.1203 0.0056 0.0075 0.0270 0.0797 0.0636 0.0913 0.1000 0.0610 0.4440 1.0000
CNSR 0.0001 0.0119 0.0145 0.0563 0.1195 0.0894 0.1178 0.1858 0.1177 0.2870 1.0000
MFML 0.0001 0.0006 0.0863 0.1560 0.1680 0.0772 0.0850 0.0800 0.0570 0.2897 1.0000
MOML 0.0001 0.0005 0.0885 0.1603 0.1723 0.0790 0.0869 0.0812 0.0571 0.2740 1.0000
MANM 0.0002 0.0002 0.1025 0.1864 0.1994 0.0906 0.0985 0.0887 0.0574 0.1761 1.0000
MPTY 0.0002 0.0002 0.1031 0.1876 0.2007 0.0912 0.0990 0.0890 0.0573 0.1717 1.0000
MFSH 0.0002 0.0002 0.1028 0.1871 0.2001 0.0909 0.0988 0.0889 0.0573 0.1737 1.0000
MOFD 0.0001 0.0009 0.0737 0.1324 0.1434 0.0667 0.0745 0.0732 0.0568 0.3784 1.0000
MCHM 0.0571 0.0080 0.0099 0.0222 0.0493 0.0412 0.0695 0.1044 0.0966 0.5417 1.0000
MPET 0.0407 0.0026 0.0085 0.0127 0.0331 0.0312 0.0518 0.1027 0.0960 0.6208 1.0000
MOND 0.0778 0.0009 0.0380 0.0781 0.1246 0.0689 0.0921 0.1055 0.0697 0.3443 1.0000
MPRM 0.1252 0.0044 0.0092 0.0460 0.1064 0.0664 0.0937 0.0969 0.0663 0.3856 1.0000
MORD 0.0407 0.0006 0.0099 0.0460 0.0950 0.0611 0.1021 0.1715 0.1170 0.3561 1.0000
MSEM 0.0607 0.0002 0.0085 0.0298 0.0637 0.0377 0.0635 0.1027 0.1419 0.4913 1.0000
MODR 0.0727 0.0034 0.0185 0.0606 0.1170 0.0666 0.0932 0.1232 0.0937 0.3511 1.0000
TAIR 0.0242 0.0104 0.0117 0.0331 0.0705 0.0604 0.1005 0.1079 0.0744 0.5069 1.0000
TRUK 0.0134 0.0012 0.0172 0.0645 0.1813 0.1056 0.1539 0.1861 0.0402 0.2365 1.0000
TWAT 0.0051 0.0107 0.0114 0.0248 0.0616 0.0539 0.0753 0.1040 0.1344 0.5187 1.0000
TRAL 0.0384 0.0009 0.0038 0.0059 0.0403 0.0528 0.1077 0.1572 0.1355 0.4574 1.0000
TOTH 0.0259 0.1860 0.0332 0.0812 0.1275 0.0703 0.1099 0.1099 0.0564 0.1996 1.0000
TLTP 0.0704 0.0004 0.1128 0.1628 0.2664 0.1037 0.0562 0.0373 0.0314 0.1586 1.0000
COMC 0.0050 0.0094 0.0107 0.0223 0.0389 0.0402 0.0702 0.1325 0.1110 0.5599 1.0000
INFO 0.0185 0.0029 0.0209 0.0219 0.0391 0.0253 0.0508 0.0817 0.1098 0.6291 1.0000
PELE 0.0121 0.0061 0.0067 0.0107 0.0289 0.0272 0.0585 0.1291 0.1124 0.6082 1.0000
GASU 0.0571 0.0070 0.0064 0.0111 0.0318 0.0305 0.0669 0.1491 0.1235 0.5166 1.0000
PWAT 0.2255 0.0066 0.0056 0.0099 0.0289 0.0278 0.0616 0.1378 0.1113 0.3850 1.0000
SANT 0.2255 0.0066 0.0056 0.0099 0.0289 0.0278 0.0616 0.1378 0.1113 0.3850 1.0000
WTRD 0.0000 0.0262 0.0209 0.0473 0.0869 0.0536 0.0692 0.1106 0.1121 0.4732 1.0000
RTRD 0.0023 0.0182 0.2197 0.1786 0.1352 0.0450 0.0490 0.0491 0.0458 0.2570 1.0000
REST 0.0003 0.0086 0.0191 0.0300 0.0447 0.0245 0.0333 0.0578 0.0672 0.7146 1.0000
BANK 0.0032 0.0031 0.0128 0.0349 0.0681 0.0325 0.0585 0.0742 0.0891 0.6235 1.0000
SECB 0.0056 0.0015 0.0019 0.0062 0.0197 0.0148 0.0325 0.0481 0.0556 0.8142 1.0000
INSR 0.0005 0.0124 0.0060 0.0315 0.0716 0.0684 0.0766 0.1465 0.1196 0.4670 1.0000
OODW 0.0000 0.0027 0.0079 0.0083 0.0150 0.0119 0.0198 0.0384 0.0566 0.8393 1.0000
HOTR 0.0031 0.0011 0.4139 0.1799 0.1035 0.0292 0.0327 0.0290 0.0253 0.1822 1.0000
PSRV 0.0529 0.1284 0.1229 0.1035 0.0912 0.0374 0.0518 0.0419 0.0426 0.3273 1.0000
VSRV 0.0000 0.0046 0.0083 0.0182 0.0436 0.0365 0.0676 0.1030 0.1412 0.5771 1.0000

282

WAST 0.0451 0.0009 0.0211 0.0557 0.1421 0.0692 0.1186 0.1138 0.0766 0.3568 1.0000
OBSV 0.0114 0.0144 0.0289 0.0404 0.0560 0.0346 0.0537 0.0789 0.0938 0.5878 1.0000
ENTR 0.0185 0.1304 0.1350 0.1024 0.0907 0.0357 0.0579 0.0706 0.0626 0.2961 1.0000
EDUC 0.0023 0.0049 0.0233 0.0432 0.1099 0.0818 0.1399 0.1908 0.1686 0.2352 1.0000
MEDC 0.0012 0.0007 0.0067 0.0190 0.0369 0.0315 0.0460 0.0881 0.0839 0.6860 1.0000
OSOC 0.0112 0.0173 0.1462 0.1895 0.1628 0.0731 0.0871 0.1002 0.0579 0.1546 1.0000
GELE 0.0022 0.0175 0.0346 0.0524 0.1066 0.0857 0.1337 0.1758 0.1631 0.2284 1.0000
TLTG 0.0072 -0.1004 -0.0948 -0.0361 0.1337 0.0531 0.2038 0.3107 0.2183 0.3044 1.0000
GVUT 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
FGML 0.0114 0.0185 0.0355 0.0528 0.1053 0.0852 0.1317 0.1728 0.1611 0.2256 1.0000
OGOV 0.0079 0.0034 0.0190 0.0416 0.1093 0.0814 0.1413 0.1909 0.1689 0.2363 1.0000
SGGV 0.0024 0.0062 0.0222 0.0440 0.1092 0.0826 0.1404 0.1887 0.1685 0.2358 1.0000
NCMP 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
Total 0.0140 0.0157 0.0441 0.0533 0.0794 0.0492 0.0738 0.1057 0.0975 0.4673 1.0000

Table E5: MSIDM (labor) in USCGE sector scheme, ($2007m).

LT10k 10-15k 15-25k 25-35k 35-50k 50-60k 60-75k 75-100k 100-150k 150k+ Total
ABEEF 0 2,075 0 0 0 0 0 0 0 0 2,077
ADARY 0 1,478 0 0 0 0 0 0 0 0 1,480
AOLVS 0 1,871 0 0 0 0 0 0 0 0 1,872
APOUL 0 2,369 0 0 0 0 0 0 0 0 2,371
AFISH 751 20 13 71 163 51 105 34 32 27 1,268
AOTH 3,804 22,075 5,661 3,559 1,881 592 1,055 672 455 1,015 40,769
COAL 1,371 0 46 300 973 783 1,118 1,090 444 792 6,917
CRUD 295 2,936 437 2,009 4,711 4,021 6,268 10,305 8,097 34,116 73,195
OMIN 2,524 85 75 483 1,515 1,216 1,718 1,711 710 1,310 11,346
CNSR 63 5,773 6,473 29,353 62,756 47,088 61,579 96,390 57,650 83,916 451,041
MFML 1 0 332 605 646 293 317 282 178 461 3,114
MOML 1 0 387 705 753 341 370 329 207 538 3,630
MANM 2 0 1,092 1,992 2,128 965 1,044 930 585 1,519 10,257
MPTY 1 0 751 1,370 1,463 663 718 640 403 1,045 7,055
MFSH 0 0 165 301 322 146 158 141 88 230 1,550
MOFD 6 0 3,862 7,041 7,522 3,410 3,692 3,288 2,070 5,371 36,262
MCHM 8,829 1,029 964 2,839 6,571 5,560 9,375 13,508 11,089 28,047 87,811
MPET 2,318 0 124 373 1,267 1,318 2,180 4,377 3,332 6,088 21,376
MOND 14,248 0 6,538 13,877 22,053 12,037 15,910 17,475 10,087 25,783 138,008

283

MPRM 6,248 163 303 2,133 5,024 3,088 4,302 4,113 2,249 3,892 31,515
MORD 593 0 119 643 1,335 853 1,424 2,375 1,526 2,563 11,430
MSEM 8,791 0 1,156 4,234 9,084 5,341 9,003 14,506 20,001 62,884 135,000
MODR 35,284 1,286 8,004 28,398 55,016 30,969 42,938 55,388 39,053 77,690 374,026
TAIR 974 404 427 1,285 2,753 2,365 3,938 4,135 2,689 15,929 34,899
TRUK 1,316 0 1,438 6,099 17,415 10,087 14,630 17,342 2,563 5,203 76,093
TWAT 38 72 65 163 419 370 509 676 857 1,805 4,974
TRAL 894 0 23 67 808 1,124 2,333 3,321 2,655 3,059 14,285
TOTH 2,664 19,080 3,229 8,190 12,834 7,032 10,964 10,647 4,836 8,471 87,948
TLTP 985 0 1,561 2,260 3,695 1,426 746 444 326 563 12,005
COMC 1,182 1,774 1,458 4,195 7,298 8,092 14,213 26,785 19,614 40,294 124,906
INFO 3,484 378 3,475 3,638 6,484 4,076 8,406 13,126 17,354 69,313 129,736
PELE 1,661 569 201 727 2,624 2,695 6,276 14,339 10,507 12,690 52,289
GASU 1,525 137 48 175 632 649 1,511 3,451 2,529 3,054 13,710
PWAT 565 12 4 16 56 58 135 308 226 273 1,653
SANT 565 12 4 16 56 58 135 308 226 273 1,653
WTRD 7 12,071 9,043 21,546 39,674 24,226 30,824 48,627 47,576 144,399 377,992
RTRD 1,277 9,504 118,808 96,397 71,971 23,256 24,574 22,677 19,035 59,921 447,420
REST 129 3,280 6,542 11,795 16,606 7,697 9,069 14,258 12,024 37,327 118,728
BANK 1,289 703 3,557 12,297 24,260 10,677 19,498 22,141 24,591 86,909 205,921
SECB 1,090 231 285 1,152 3,775 2,854 6,295 9,307 10,771 159,257 195,017
INSR 92 2,408 888 6,099 14,045 13,538 14,906 28,507 22,300 64,462 167,245
OODW 0 0 0 0 0 0 0 0 0 0 0
HOTR 794 132 105,506 45,610 25,715 6,883 7,374 5,493 3,648 5,637 206,792
PSRV 3,265 7,834 7,378 6,182 5,264 2,034 2,738 1,711 1,355 2,200 39,962
VSRV 0 45 79 184 447 376 699 1,056 1,445 5,315 9,648
WAST 1,150 0 476 1,357 3,508 1,677 2,879 2,622 1,545 3,256 18,469
OBSV 23,210 27,631 55,087 78,772 107,885 65,404 100,819 142,848 161,150 556,200 1,319,005
ENTR 1,761 12,348 12,660 9,565 8,286 3,140 5,071 5,877 4,731 11,494 74,932
EDUC 1,320 519 10,354 21,732 59,910 43,623 76,841 105,519 91,897 124,053 535,769
MEDC 712 7 3,177 10,764 21,154 18,247 26,437 50,696 46,747 362,826 540,766
OSOC 1,765 2,650 22,828 29,640 25,299 11,261 13,292 14,976 7,964 8,423 138,098
GELE 35 7 180 446 1,242 911 1,615 2,193 1,923 2,690 11,241
TLTG 39 8 233 577 1,606 1,178 2,089 2,836 2,487 3,479 14,533
GVUT 0 0 0 0 0 0 0 0 0 0 0
FGML 2,318 81 2,236 5,550 15,438 11,320 20,083 27,263 23,901 33,442 141,630
OGOV 2,330 163 4,497 11,161 31,047 22,764 40,388 54,827 48,066 67,253 282,495

284

SGGV 1,103 240 6,620 16,431 45,705 33,512 59,456 80,713 70,759 99,006 413,546
NCMP 0 0 0 0 0 0 0 0 0 0 0
Total 144,670 143,460 418,871 514,376 763,092 461,345 692,018 966,583 826,553 2,335,762 7,266,731

Table E6: MSIDM (labor) in USCGE sector scheme coefficients, (income bracket
proportion of total sector labor income payments to households).

LT10k 10-15k 15-25k 25-35k 35-50k 50-60k 60-75k 75-100k 100-150k 150k+ Total
ABEEF 0.0000 0.9990 0.0002 0.0001 0.0002 0.0001 0.0001 0.0002 0.0001 0.0001 1.0000
ADARY 0.0000 0.9990 0.0002 0.0001 0.0002 0.0001 0.0001 0.0002 0.0001 0.0001 1.0000
AOLVS 0.0000 0.9990 0.0002 0.0001 0.0002 0.0001 0.0001 0.0002 0.0001 0.0001 1.0000
APOUL 0.0000 0.9990 0.0002 0.0001 0.0002 0.0001 0.0001 0.0002 0.0001 0.0001 1.0000
AFISH 0.5927 0.0158 0.0102 0.0561 0.1286 0.0402 0.0829 0.0266 0.0255 0.0214 1.0000
AOTH 0.0933 0.5415 0.1389 0.0873 0.0461 0.0145 0.0259 0.0165 0.0112 0.0249 1.0000
COAL 0.1983 0.0000 0.0066 0.0434 0.1406 0.1132 0.1617 0.1575 0.0643 0.1145 1.0000
CRUD 0.0040 0.0401 0.0060 0.0274 0.0644 0.0549 0.0856 0.1408 0.1106 0.4661 1.0000
OMIN 0.2225 0.0075 0.0066 0.0426 0.1335 0.1072 0.1514 0.1508 0.0625 0.1155 1.0000
CNSR 0.0001 0.0128 0.0144 0.0651 0.1391 0.1044 0.1365 0.2137 0.1278 0.1860 1.0000
MFML 0.0002 0.0000 0.1065 0.1942 0.2074 0.0940 0.1018 0.0907 0.0571 0.1481 1.0000
MOML 0.0002 0.0000 0.1065 0.1942 0.2074 0.0940 0.1018 0.0907 0.0571 0.1481 1.0000
MANM 0.0002 0.0000 0.1065 0.1942 0.2074 0.0940 0.1018 0.0907 0.0571 0.1481 1.0000
MPTY 0.0002 0.0000 0.1065 0.1942 0.2074 0.0940 0.1018 0.0907 0.0571 0.1481 1.0000
MFSH 0.0002 0.0000 0.1065 0.1942 0.2074 0.0940 0.1018 0.0907 0.0571 0.1481 1.0000
MOFD 0.0002 0.0000 0.1065 0.1942 0.2074 0.0940 0.1018 0.0907 0.0571 0.1481 1.0000
MCHM 0.1005 0.0117 0.0110 0.0323 0.0748 0.0633 0.1068 0.1538 0.1263 0.3194 1.0000
MPET 0.1084 0.0000 0.0058 0.0175 0.0593 0.0617 0.1020 0.2047 0.1559 0.2848 1.0000
MOND 0.1032 0.0000 0.0474 0.1006 0.1598 0.0872 0.1153 0.1266 0.0731 0.1868 1.0000
MPRM 0.1983 0.0052 0.0096 0.0677 0.1594 0.0980 0.1365 0.1305 0.0714 0.1235 1.0000
MORD 0.0518 0.0000 0.0104 0.0562 0.1168 0.0746 0.1246 0.2078 0.1335 0.2242 1.0000
MSEM 0.0651 0.0000 0.0086 0.0314 0.0673 0.0396 0.0667 0.1075 0.1482 0.4658 1.0000
MODR 0.0943 0.0034 0.0214 0.0759 0.1471 0.0828 0.1148 0.1481 0.1044 0.2077 1.0000
TAIR 0.0279 0.0116 0.0122 0.0368 0.0789 0.0678 0.1128 0.1185 0.0770 0.4564 1.0000
TRUK 0.0173 0.0000 0.0189 0.0802 0.2289 0.1326 0.1923 0.2279 0.0337 0.0684 1.0000
TWAT 0.0077 0.0145 0.0130 0.0327 0.0843 0.0745 0.1024 0.1359 0.1723 0.3629 1.0000
TRAL 0.0626 0.0000 0.0016 0.0047 0.0565 0.0787 0.1633 0.2325 0.1859 0.2141 1.0000
TOTH 0.0303 0.2169 0.0367 0.0931 0.1459 0.0800 0.1247 0.1211 0.0550 0.0963 1.0000
TLTP 0.0820 0.0000 0.1300 0.1883 0.3077 0.1188 0.0621 0.0370 0.0272 0.0469 1.0000
COMC 0.0095 0.0142 0.0117 0.0336 0.0584 0.0648 0.1138 0.2144 0.1570 0.3226 1.0000

285

INFO 0.0269 0.0029 0.0268 0.0280 0.0500 0.0314 0.0648 0.1012 0.1338 0.5343 1.0000
PELE 0.0318 0.0109 0.0038 0.0139 0.0502 0.0515 0.1200 0.2742 0.2009 0.2427 1.0000
GASU 0.1112 0.0100 0.0035 0.0128 0.0461 0.0473 0.1102 0.2517 0.1845 0.2228 1.0000
PWAT 0.3416 0.0074 0.0026 0.0095 0.0341 0.0350 0.0816 0.1865 0.1366 0.1650 1.0000
SANT 0.3416 0.0074 0.0026 0.0095 0.0341 0.0350 0.0816 0.1865 0.1366 0.1650 1.0000
WTRD 0.0000 0.0319 0.0239 0.0570 0.1050 0.0641 0.0815 0.1286 0.1259 0.3820 1.0000
RTRD 0.0029 0.0212 0.2655 0.2155 0.1609 0.0520 0.0549 0.0507 0.0425 0.1339 1.0000
REST 0.0011 0.0276 0.0551 0.0993 0.1399 0.0648 0.0764 0.1201 0.1013 0.3144 1.0000
BANK 0.0063 0.0034 0.0173 0.0597 0.1178 0.0519 0.0947 0.1075 0.1194 0.4220 1.0000
SECB 0.0056 0.0012 0.0015 0.0059 0.0194 0.0146 0.0323 0.0477 0.0552 0.8166 1.0000
INSR 0.0006 0.0144 0.0053 0.0365 0.0840 0.0809 0.0891 0.1705 0.1333 0.3854 1.0000
OODW 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
HOTR 0.0038 0.0006 0.5102 0.2206 0.1244 0.0333 0.0357 0.0266 0.0176 0.0273 1.0000
PSRV 0.0817 0.1960 0.1846 0.1547 0.1317 0.0509 0.0685 0.0428 0.0339 0.0551 1.0000
VSRV 0.0000 0.0046 0.0082 0.0191 0.0464 0.0390 0.0725 0.1095 0.1498 0.5509 1.0000
WAST 0.0622 0.0000 0.0258 0.0735 0.1900 0.0908 0.1559 0.1420 0.0836 0.1763 1.0000
OBSV 0.0176 0.0209 0.0418 0.0597 0.0818 0.0496 0.0764 0.1083 0.1222 0.4217 1.0000
ENTR 0.0235 0.1648 0.1690 0.1276 0.1106 0.0419 0.0677 0.0784 0.0631 0.1534 1.0000
EDUC 0.0025 0.0010 0.0193 0.0406 0.1118 0.0814 0.1434 0.1969 0.1715 0.2315 1.0000
MEDC 0.0013 0.0000 0.0059 0.0199 0.0391 0.0337 0.0489 0.0937 0.0864 0.6709 1.0000
OSOC 0.0128 0.0192 0.1653 0.2146 0.1832 0.0815 0.0963 0.1084 0.0577 0.0610 1.0000
GELE 0.0031 0.0006 0.0160 0.0397 0.1105 0.0810 0.1437 0.1951 0.1710 0.2393 1.0000
TLTG 0.0027 0.0006 0.0160 0.0397 0.1105 0.0810 0.1438 0.1952 0.1711 0.2394 1.0000
GVUT 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
FGML 0.0164 0.0006 0.0158 0.0392 0.1090 0.0799 0.1418 0.1925 0.1688 0.2361 1.0000
OGOV 0.0082 0.0006 0.0159 0.0395 0.1099 0.0806 0.1430 0.1941 0.1701 0.2381 1.0000
SGGV 0.0027 0.0006 0.0160 0.0397 0.1105 0.0810 0.1438 0.1952 0.1711 0.2394 1.0000
NCMP 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
Total 0.0199 0.0197 0.0576 0.0708 0.1050 0.0635 0.0952 0.1330 0.1137 0.3214 1.0000

286

Table E7: MSIDM (capital) in USCGE sector scheme, ($2007m).

LT10k 10-15k 15-25k 25-35k 35-50k 50-60k 60-75k 75-100k 100-150k 150k+ Total
ABEEF 0 74 177 110 198 113 188 293 310 1,795 3,257
ADARY 0 168 402 250 452 257 429 666 706 4,086 7,415
AOLVS 0 110 264 164 297 169 281 437 464 2,683 4,870
APOUL 0 47 113 70 127 72 121 188 199 1,152 2,090
AFISH 0 5 13 14 24 19 31 60 88 1,264 1,520
AOTH 0 639 1,539 998 1,800 1,060 1,770 2,840 3,198 24,215 38,060
COAL 0 14 30 30 55 39 66 127 185 2,441 2,987
CRUD 0 359 813 801 1,476 1,068 1,783 3,450 5,016 66,702 81,466
OMIN 0 33 82 85 159 119 199 388 570 8,008 9,642
CNSR 0 698 1,372 1,225 2,086 1,427 2,382 4,432 6,261 71,843 91,725
MFML 0 2 7 7 12 10 16 31 46 674 805
MOML 0 2 7 7 12 10 16 31 46 679 811
MANM 0 3 6 5 9 6 11 20 29 367 457
MPTY 0 2 3 3 5 4 6 12 17 212 263
MFSH 0 0 1 1 1 1 1 3 4 51 63
MOFD 0 47 139 148 268 213 355 689 1,015 15,180 18,055
MCHM 0 214 573 589 1,053 817 1,364 2,630 3,852 55,703 66,794
MPET 0 150 358 352 620 463 774 1,477 2,143 29,310 35,646
MOND 0 172 426 425 754 571 954 1,829 2,663 37,229 45,024
MPRM 0 58 156 161 288 224 374 722 1,057 15,350 18,389
MORD 0 9 25 26 47 37 62 120 177 2,620 3,124
MSEM 0 26 77 81 147 117 195 378 556 8,292 9,868
MODR 0 354 951 978 1,750 1,357 2,267 4,372 6,403 92,645 111,078
TAIR 0 16 45 46 83 65 108 209 307 4,476 5,354
TRUK 0 118 258 244 425 308 514 972 1,397 18,072 22,307
TWAT 0 7 20 21 38 29 49 95 139 2,041 2,440
TRAL 0 21 66 71 129 104 174 338 499 7,587 8,989
TOTH 0 92 193 178 307 218 363 683 976 12,102 15,111
TLTP 0 6 16 17 31 24 40 77 113 1,655 1,980
COMC 0 440 1,077 1,070 1,892 1,428 2,384 4,565 6,637 92,147 111,640
INFO 0 160 457 484 887 695 1,160 2,259 3,328 49,158 58,588
PELE 0 269 723 744 1,332 1,033 1,725 3,329 4,875 70,568 84,599
GASU 0 49 123 123 217 165 275 527 767 10,735 12,981
PWAT 0 4 10 9 16 12 19 37 53 691 851
SANT 0 4 10 9 16 12 19 37 53 691 851

287

WTRD 0 340 873 883 1,572 1,204 2,010 3,865 5,642 80,158 96,549
RTRD 0 433 1,012 986 1,732 1,287 2,149 4,096 5,930 80,232 97,857
REST 0 1,042 3,012 3,178 5,731 4,526 7,558 14,650 21,552 319,855 381,105
BANK 0 549 1,557 1,633 2,939 2,311 3,859 7,471 10,978 161,956 193,252
SECB 0 55 78 54 82 37 62 97 111 0 576
INSR 0 135 343 346 616 471 786 1,510 2,203 31,194 37,603
OODW 0 947 2,737 2,888 5,208 4,113 6,868 13,312 19,583 290,634 346,288
HOTR 0 156 420 432 774 600 1,002 1,934 2,832 41,002 49,152
PSRV 0 81 202 203 359 273 456 875 1,274 17,981 21,704
VSRV 0 4 10 10 17 13 22 41 60 836 1,012
WAST 0 24 62 63 113 87 145 280 408 5,838 7,021
OBSV 0 1,644 3,766 3,499 6,140 5,020 8,384 17,702 29,849 640,192 716,196
ENTR 0 92 213 207 365 270 450 859 1,243 16,752 20,451
EDUC 0 2,328 3,176 3,297 3,826 3,821 4,296 5,132 5,855 12,293 44,024
MEDC 0 450 936 862 1,484 1,047 1,749 3,285 4,686 57,612 72,113
OSOC 0 78 188 186 329 248 414 792 1,150 15,904 19,289
GELE 0 269 364 378 435 437 487 572 644 903 4,489
TLTG 0 -548 -742 -772 -887 -893 -994 -1,166 -1,313 -1,843 -9,157
GVUT 0 0 0 0 0 0 0 0 0 0 0
FGML 0 3,666 4,963 5,159 5,932 5,968 6,643 7,796 8,779 12,321 61,227
OGOV 0 833 1,128 1,173 1,348 1,356 1,510 1,772 1,995 2,801 13,917
SGGV 0 2,613 3,538 3,678 4,229 4,254 4,736 5,558 6,258 8,783 43,646
NCMP 0 0 0 0 0 0 0 0 0 0 0
Total 0 19,563 38,364 37,889 59,358 48,713 73,069 128,758 183,870 2,507,831 3,097,415

Table E8: MSIDM (capital) in USCGE sector scheme coefficients, (income bracket
proportion of total sector capital income payments to households).

LT10k 10-15k 15-25k 25-35k 35-50k 50-60k 60-75k 75-100k
100-
150k 150k+ Total
ABEEF 0.0000 0.0226 0.0542 0.0337 0.0609 0.0346 0.0578 0.0898 0.0953 0.5510 1.0000
ADARY 0.0000 0.0226 0.0542 0.0337 0.0609 0.0346 0.0578 0.0898 0.0953 0.5510 1.0000
AOLVS 0.0000 0.0226 0.0542 0.0337 0.0609 0.0346 0.0578 0.0898 0.0953 0.5510 1.0000
APOUL 0.0000 0.0226 0.0542 0.0337 0.0609 0.0346 0.0578 0.0898 0.0953 0.5510 1.0000
AFISH 0.0000 0.0034 0.0089 0.0090 0.0161 0.0124 0.0207 0.0398 0.0581 0.8317 1.0000
AOTH 0.0000 0.0168 0.0404 0.0262 0.0473 0.0279 0.0465 0.0746 0.0840 0.6363 1.0000
COAL 0.0000 0.0045 0.0102 0.0100 0.0183 0.0132 0.0221 0.0426 0.0619 0.8171 1.0000
CRUD 0.0000 0.0044 0.0100 0.0098 0.0181 0.0131 0.0219 0.0423 0.0616 0.8188 1.0000

288

OMIN 0.0000 0.0034 0.0085 0.0088 0.0165 0.0123 0.0206 0.0403 0.0591 0.8306 1.0000
CNSR 0.0000 0.0076 0.0150 0.0134 0.0227 0.0156 0.0260 0.0483 0.0683 0.7832 1.0000
MFML 0.0000 0.0029 0.0081 0.0085 0.0153 0.0120 0.0200 0.0388 0.0570 0.8373 1.0000
MOML 0.0000 0.0029 0.0081 0.0085 0.0153 0.0120 0.0200 0.0388 0.0570 0.8373 1.0000
MANM 0.0000 0.0058 0.0123 0.0115 0.0198 0.0142 0.0237 0.0446 0.0639 0.8043 1.0000
MPTY 0.0000 0.0058 0.0123 0.0115 0.0198 0.0142 0.0237 0.0446 0.0639 0.8043 1.0000
MFSH 0.0000 0.0059 0.0124 0.0116 0.0200 0.0142 0.0238 0.0448 0.0641 0.8033 1.0000
MOFD 0.0000 0.0026 0.0077 0.0082 0.0148 0.0118 0.0197 0.0382 0.0562 0.8408 1.0000
MCHM 0.0000 0.0032 0.0086 0.0088 0.0158 0.0122 0.0204 0.0394 0.0577 0.8339 1.0000
MPET 0.0000 0.0042 0.0100 0.0099 0.0174 0.0130 0.0217 0.0414 0.0601 0.8222 1.0000
MOND 0.0000 0.0038 0.0095 0.0094 0.0167 0.0127 0.0212 0.0406 0.0591 0.8269 1.0000
MPRM 0.0000 0.0031 0.0085 0.0087 0.0157 0.0122 0.0203 0.0392 0.0575 0.8347 1.0000
MORD 0.0000 0.0028 0.0080 0.0084 0.0151 0.0119 0.0199 0.0385 0.0567 0.8388 1.0000
MSEM 0.0000 0.0026 0.0078 0.0083 0.0149 0.0118 0.0197 0.0383 0.0563 0.8403 1.0000
MODR 0.0000 0.0032 0.0086 0.0088 0.0158 0.0122 0.0204 0.0394 0.0576 0.8341 1.0000
TAIR 0.0000 0.0030 0.0083 0.0086 0.0155 0.0121 0.0202 0.0390 0.0573 0.8359 1.0000
TRUK 0.0000 0.0053 0.0116 0.0109 0.0190 0.0138 0.0230 0.0436 0.0626 0.8101 1.0000
TWAT 0.0000 0.0030 0.0083 0.0086 0.0154 0.0121 0.0202 0.0390 0.0572 0.8363 1.0000
TRAL 0.0000 0.0023 0.0073 0.0079 0.0144 0.0116 0.0193 0.0376 0.0556 0.8441 1.0000
TOTH 0.0000 0.0061 0.0127 0.0118 0.0203 0.0144 0.0240 0.0452 0.0646 0.8008 1.0000
TLTP 0.0000 0.0030 0.0083 0.0086 0.0155 0.0121 0.0202 0.0390 0.0572 0.8360 1.0000
COMC 0.0000 0.0039 0.0097 0.0096 0.0169 0.0128 0.0214 0.0409 0.0595 0.8254 1.0000
INFO 0.0000 0.0027 0.0078 0.0083 0.0151 0.0119 0.0198 0.0386 0.0568 0.8390 1.0000
PELE 0.0000 0.0032 0.0085 0.0088 0.0157 0.0122 0.0204 0.0393 0.0576 0.8342 1.0000
GASU 0.0000 0.0038 0.0094 0.0094 0.0167 0.0127 0.0212 0.0406 0.0591 0.8270 1.0000
PWAT 0.0000 0.0051 0.0113 0.0107 0.0187 0.0136 0.0228 0.0432 0.0622 0.8123 1.0000
SANT 0.0000 0.0051 0.0113 0.0107 0.0187 0.0136 0.0228 0.0432 0.0622 0.8123 1.0000
WTRD 0.0000 0.0035 0.0090 0.0091 0.0163 0.0125 0.0208 0.0400 0.0584 0.8302 1.0000
RTRD 0.0000 0.0044 0.0103 0.0101 0.0177 0.0131 0.0220 0.0419 0.0606 0.8199 1.0000
REST 0.0000 0.0027 0.0079 0.0083 0.0150 0.0119 0.0198 0.0384 0.0566 0.8393 1.0000
BANK 0.0000 0.0028 0.0081 0.0084 0.0152 0.0120 0.0200 0.0387 0.0568 0.8381 1.0000
SECB 0.0000 0.0950 0.1346 0.0938 0.1424 0.0647 0.1080 0.1690 0.1923 0.0003 1.0000
INSR 0.0000 0.0036 0.0091 0.0092 0.0164 0.0125 0.0209 0.0402 0.0586 0.8296 1.0000
OODW 0.0000 0.0027 0.0079 0.0083 0.0150 0.0119 0.0198 0.0384 0.0566 0.8393 1.0000
HOTR 0.0000 0.0032 0.0085 0.0088 0.0157 0.0122 0.0204 0.0393 0.0576 0.8342 1.0000
PSRV 0.0000 0.0037 0.0093 0.0093 0.0166 0.0126 0.0210 0.0403 0.0587 0.8285 1.0000
VSRV 0.0000 0.0039 0.0095 0.0095 0.0168 0.0127 0.0213 0.0407 0.0593 0.8263 1.0000

289

WAST 0.0000 0.0034 0.0089 0.0090 0.0161 0.0124 0.0207 0.0398 0.0582 0.8315 1.0000
OBSV 0.0000 0.0023 0.0053 0.0049 0.0086 0.0070 0.0117 0.0247 0.0417 0.8939 1.0000
ENTR 0.0000 0.0045 0.0104 0.0101 0.0179 0.0132 0.0220 0.0420 0.0608 0.8192 1.0000
EDUC 0.0000 0.0529 0.0721 0.0749 0.0869 0.0868 0.0976 0.1166 0.1330 0.2792 1.0000
MEDC 0.0000 0.0062 0.0130 0.0120 0.0206 0.0145 0.0243 0.0456 0.0650 0.7989 1.0000
OSOC 0.0000 0.0040 0.0098 0.0097 0.0171 0.0128 0.0215 0.0410 0.0596 0.8245 1.0000
GELE 0.0000 0.0599 0.0811 0.0843 0.0969 0.0975 0.1085 0.1273 0.1434 0.2012 1.0000
TLTG 0.0000 0.0599 0.0811 0.0843 0.0969 0.0975 0.1085 0.1273 0.1434 0.2012 1.0000
GVUT 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
FGML 0.0000 0.0599 0.0811 0.0843 0.0969 0.0975 0.1085 0.1273 0.1434 0.2012 1.0000
OGOV 0.0000 0.0599 0.0811 0.0843 0.0969 0.0975 0.1085 0.1273 0.1434 0.2012 1.0000
SGGV 0.0000 0.0599 0.0811 0.0843 0.0969 0.0975 0.1085 0.1273 0.1434 0.2012 1.0000
NCMP 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
Total 0.0000 0.0063 0.0124 0.0122 0.0192 0.0157 0.0236 0.0416 0.0594 0.8097 1.0000

Table E9: U.S. Total Income by Type of Income, 2007.

Income
Brackets
Labor
Income
Capital
Income
Variable
Income
Subtotal Transfers
Pension
and
Annuities
Other
Income
Constant
Income
Subtotal
Total
Income
LT$10k 120.0 -30.1 90.0 48.8 23.6 -68.6 3.8 93.7
10-15k 119.0 14.0 133.0 35.5 22.3 1.2 59.0 192.0
15-25k 347.5 27.4 374.9 59.9 56.7 3.3 119.8 494.8
25-35k 426.7 27.1 453.8 43.7 59.6 2.9 106.2 560.1
35-50k 633.1 42.4 675.5 51.8 89.2 3.4 144.4 819.9
50-60k 382.7 34.8 417.6 28.4 67.2 2.6 98.2 515.7
60-75k 574.1 52.2 626.4 42.6 100.8 3.8 147.2 773.6
75-100k 801.9 92.1 894.0 48.1 150.9 5.8 204.7 1,098.7
100-150k 685.7 131.5 817.2 29.3 134.1 6.1 169.5 986.7
150k+ 1,937.8 1,792.9 3,730.7 59.5 295.2 41.3 395.9 4,126.6
Total 6,028.8 2,184.3 8,213.1 447.5 999.5 1.7 1,448.7 9,661.8

290

Table E10: U.S. Labor and Capital Income Shares by
Income Bracket, 2007.

Income
Brackets
Variable Income Total Income
Labor
Income
Capital
Income
Labor
Income
Capital
Income
Constant
Income
LT$10k 1.334 -0.334 1.281 -0.321 0.040
10-15k 0.895 0.105 0.620 0.073 0.307
15-25k 0.927 0.073 0.702 0.055 0.242
25-35k 0.940 0.060 0.762 0.048 0.190
35-50k 0.937 0.063 0.772 0.052 0.176
50-60k 0.917 0.083 0.742 0.068 0.190
60-75k 0.917 0.083 0.742 0.068 0.190
75-100k 0.897 0.103 0.730 0.084 0.186
100-150k 0.839 0.161 0.695 0.133 0.172
150k+ 0.519 0.481 0.470 0.434 0.096
Total 0.734 0.266 0.624 0.226 0.150

Table E11: U.S. Labor Income Brackets, Based on
Income Shares, 2007.
Total
Income
Brackets
Total Income
Bracket Upper
Limit ('000s)
Labor Income
Share of Total
Income
Labor Income
Bracket Upper
Limit ($)
LT$10k 10 1.281 12,806
81

10-15k 15 0.620 9,297
15-25k 25 0.702 17,559
25-35k 35 0.762 26,668
35-50k 50 0.772 38,606
50-60k 60 0.742 44,529
60-75k 75 0.742 55,662
75-100k 100 0.730 72,990
100-150k 150 0.695 104,249
150k+ 0.470
Total 0.624

81
As described in the text, when labor income shares are calculated using positive income only, this labor
income bracket upper limit is instead $6,237

291

APPENDIX F: REGIONALIZATION MODEL
The purpose of Appendix F and Table F1 is to present all of the equations in Dixon and
Rimmer (2004)‘s regionalization approach. At this time, most of the necessary data has been
obtained, especially with respect to output shares, trade data, and some demand and inter-state
flow data. Further work is required to complete the inter-state flow data sets, as well as to check
through additional data requirements. Once all the data is collected and appropriately formatted,
the computational requirements are relatively straightforward. Although ideally this model would
be solved through an equilibrium computational model, it is feasible to run a bootstrap-type
version of the model by guessing initial values for key variables and then iterating the various sets
of equations in sequence. This produces results which are closely converging to the equilibrium
state.
One notable characteristic of the Dixon & Rimmer (2004) model is that many variables
are percentage change as opposed to levels. This follows the computational approach of the
various models employed at the Monash Center of Policy Studies, and has some useful properties
in this case. The percentage change approach allows for variables in the regionalizing model to
follow a straightforward format:
var_R(r) = var + relevant_R(r) – sum
g
SHVAR(g) * relevant_R(g)
Where:
var_R is a regional variable and var is the corresponding national variable – e.g.
percentage changes in intermediate demands at the state and national levels;
relevant_R is a regional variable relevant to determining the gap between var_R and var,
e.g. the percentage change in output at the state level;

292

SHVAR(g) is the coefficient giving the share of region g in the national level of var, e.g.
state shares in national output.
Results from the USCGE model are presented in levels form, though these can easily be
changed to percentage change form in most cases. The only restriction on implementing the full
version of Dixon & Rimmer‘s regional model would be if particular national variables are not
provided in the USCGE model. In that case, proxies could be identified.
Variables in Bold are national model results in % change
Table F1: Regionalization Model Equations, Variable Descriptions and Data Sources.
T1: Output by industry j in region r
(All, j,IND)(All, r,REG)
x0ind_IR(j,r) =

x0ind(j) + sum
i
{H0CI(i,j) * x0_CSR(i,‖dom‖,r)} – sum
g
(JOBSH(j,g) *
sum
i
{H0CI(i,j)*x0_CSR(i,‖dom‖,g)}

Code Description Source
x0ind_IR(j,r) % change. Output by industry j in region r Equation
x0ind(j) % change. Output by j National
Model
H0CI(i,j) Share of comm i in output of j Make
matrix
x0_CSR(i,‖dom‖,r) % change. Supply of dom i from r – ―percent changes in
the outputs of commodities in r‖ [in equation above this is
weighted by commodity shares in national output of j]
T14 (dom)
r
JOBSH(j,g) Share of g output in j Shares in
MSSM
x0_CSR(i,‖dom‖,g) % change. Supply of dom i from g – ―percent changes in
the outputs of commodities in g‖
T14 (dom)
g
T2: Disposable income in region r
(All,r,REG)
dispy_R(r) =

hdy + GAMMA * sum
j
{LABSH(j,r)*[labind(j) + x0ind_IR(j,r) – x0ind(j)]
- GAMMA * sum
g
{SHLAB(g)} sum
j
{ LABSH(j,g) * [labind(j) + x0ind_IR(j,g) – x0ind(j)]

―Relevant‖ variable is a modified measure of employment in region r.
Assume % change in employment in industry j in region r is determined by the % change in
national employment in industry j and by the deviation between the % changes in j‘s output in
region r and j‘s national output.

293

GAMMA set at 1, but may be <1 if, e.g., effects of employment changes in region r on
household consumption are softened by social security payments, other non-labor incomes and
changes in savings behavior.
Code Description Source
dispy_R(r) Disposable income in region r Equation
hdy Scalar household disposable income National
model
GAMMA Scalar. Parameter, controls sensitivity of regional
disposable income to regional employment
Discretionary
LABSH(j,r) Share of j in r‘s labor income

= JOBSH(j,r)*LABIND(j) /
sum
k
(JOBSH(k,r)*LABIND(k)

Where LABIND(j) = Value of labor input

MSSM

labind(j) % change. Employment by industry j. National
Model
x0ind_IR(j,r) % change. Output by industry and region T1
x0ind(j) % change. Output by j National
Model
SHLAB(g) Share of region r in national labor income

= sum
i
{JOBSH(j,r) * LABIND(j) / sum
i
{LABIND(k)}

MSSM
T3: Demand for commodity i from source s by households in region

(All,i,COM)(All,s,SOURCE)(All,r,REG)

x3_CSR(i,s,r) = x3cs(i,s) + dispy_R(r) – sum
g
{REGSH3(i.g)} * dispy_R(g)

x3_CSR(i,s,r) % change. Consumption of i,s by households in r T3
Dom/Imp!
x3cs(i,s) % change. Consumption by households of i,s National
Model
Dom/Imp
dispy_R(r) Disposable income in region r T2
REGSH3(i.g) Share of region r in consumption of i
―Assume that household in region r account for x per
cent of national household consumption of commodity i
(domestic and imported) if region r accounts for x per
cent of national disposable income.‖

T2
(Disposable
income)
dispy_R(g) Disposable income in region g T2
T4: Demand for commodity i to be delivered from region r to export (demand for i by region
of exit)

(All,i,COM)(All,s,SOURCE)(All,r,REG)

x4_CR(i,r) = x4(i) + reg4_CR(i,r) – sum
g
{REGSH4(i.g) * reg4_CR(i,g)

294

T4 specifies percentage movements in exports of i from ports in state r. ―Relevant‖ variable
[reg4_CR(i,r)] is the percentage change in r‘s share of the exports of i. Normally exogenous.
x4_CR(i,r) % change. Exports of i leaving US from region r T4
x4(i) % change. Exports of commodity i National
Model
reg4_CR(i,r) % change. Share of US exports of i leaving from r
(usually exogenous). Relevant variable
USA Trade
Online
REGSH4(i.g) Share of region r as a port for exports of i.
―Goods exports by ports of exit. If ports in region r
account for x per cent of US goods exports, then we
assumed that x per cent of the exports of each good
leave the US from region r‖
USA Trade
Online
reg4_CR(i,g) % change. Share of US exports of i leaving from r
(usually exogenous). Relevant variable
USA Trade
Online
T5: Demand for commodity i from source s to be delivered from region r to governments

(All,i,COM)(All,s,SOURCE)(All,r,REG)

x5_CSR(i,s,r) = x5cs(i,s)
x5_CSR(i,s,r) % change. Consumption of i,s in r by government T5 Dom/Imp
x5cs(i,s) % change. Consumption by governments of i,s National
Model
Dom/Imp
T6: Demand for commodity i from source s to be delivered from region r to inventories

(All,i,COM)(All,s,SOURCE)(All,r,REG)

d_x6_CSR(i,s,r) = SUPSH(i,s,r)*d_x6cs(i,s)

d_x6_CSR(i,s,r) Change. Inventory accumulation of i,s in r T6 Dom/Imp
SUPSH(i,s,r) Share of region r in the supply of i,s
d_x6cs(i,s) Change. Inventory accumulation of i,s National
Model
Dom/Imp
T7: Gross Regional Product

(All,r,REG)

gspreal_R(r)

= gdpreal + sum
j
{VADSH(j,r) * x0ind_IR(j,r)} – sum
g
{SHVAD(g) sum
j
{VADSH(j,g) *
x0ind_IR(j,g)}}

gspreal_R(r) % change. Gross State product in r T7
gdpreal Scalar. National. Real GDP National
Model

295

VADSH(j,r) Share of industry j in value-added in region r

[JOBSH(j,r)*VADD(j)]/[sum
k
{JOBSH(k,r)*VADD(k)]
JOBSH – share of region r in employment and output in
industry j.
VADD – Value added in industry j/k

Shares in
MSSM

x0ind_IR(j,r) % change. Output by industry j and region r T1
SHVAD(g) Share of region r in national value-added

sum
j
{JOBSH(j,r)}*VADD(j)/sum
k
{VADD(k)}

Shares in
MSSM
VADSH(j,g) Share of industry j in value-added in region g. (As
above)
As above
x0ind_IR(j,g) % change. Output by industry j and region g T1
T8: Employment by region

(All,r,REG)

emp_R(r) =

emp_hours + sum
j
{LABSH(j,r) * [labind(j) + x0ind_IR(j,r) – x0ind(j)] – sum
g
{SHLAB(g)
sum
j
{LABSH(j,g) * [labind(j) + x0ind_IR(j,g) – x0ind(j)]

emp_R(r) % change. Reg‘l Employment in r T8
emp_hours Scalar. Nat’l Aggregate employment National
Model
LABSH(j,r) Share of industry j in r‘s labor income See T2
labind(j) % change. Nat’l Employment by industry National
Model
x0ind_IR(j,r) % change. Reg‘l Output by industry j and region r T1
x0ind(j) Nat’l Output by industry National
Model
SHLAB(g) Share of region r in national labor income See T2
LABSH(j,g) Share of industry j in g‘s labor income See T2
x0ind_IR(j,g) % change. Reg‘l Output by industry j and region g T1
T9: Demand for non-margin commodity i from source s by agents in r

(All,i,NMARG)(All,s,SOURCE)(All,r,REG)

dem_CSR(i,s,r) =

(1/TOTDEMREG(i,s,r)) *

(sum
j
{JOBSH(j,r)} *BAS1(i,s,j) * [x1csi(i,s,j) + x0ind_IR(j,r) – sum
g
{JOBSH(j,g) *
x0ind_IR(j,g)]

296

+ sum
j
{JOBSH(j,r) * BAS2(i,s,j) * [x2csi(i,s,j) + x0ind_IR(j,r) – sum
g
{JOBSH(j,g) *
x0ind_IR(j,g)]

+ REGSH3(i,r) * BAS3(i,s) * x3_CSR(i,s,r)

+ SOURCEDOM(s) * REGSH4(i,r) * BAS4(i) * x4_CR(i,r)

+ REGSH5(i,r) * BAS5(i,s) *x5_CSR(i,s,r)

+ 100* d_x6_CSR(i,s,r)}

NB – Similar to T13 below – see description from DR there.
dem_CSR(i,s,r) % change. Demand for i,s in region r. NMARG T9 Dom/Imp
TOTDEMREG(i,s,r) Demand for i,s in region r

sum
j
{JOBSH(jr)}*[BAS1(i,s,j) +
BAS2(i,s,j)+REGSH3(i,r)*BAS3(i,s) +
SOURCEDOM(s)*REGSH4(i,r) +
REGSH5(i,r)*BAS5(i,s) + SUPSH(i,s,r)*BAS6(i,s)

SUPSH(i,s,r) – Share of region r in the supply of i,s
BAS6(i,s)
Dom/Imp

See below for
most

For dom good,
output share.
For import,
share of
nation‘s
imports that
arrive through
ports in region
r.
JOBSH(j,r) Share of region r in employment and output in
industry j
MSSM
BAS1(i,s,j) Intermediate use of i,s by industry j, nat‘l. National
Model Data
Dom/Imp
x1csi(i,s,j) % change. Nat’l. Intermediate demand for i,s by
industry j
National
Model
Dom/Imp
x0ind_IR(j,r) % change. Reg‘l. Output by industry and region r T1 r
JOBSH(j,g) Share of region r in employment and output in
industry g
MSSM
x0ind_IR(j,g)] % change. Reg‘l. Output by industry and region g T1 g
BAS2(i,s,j) Investment use of i,s by industry j, nat‘l National
Model Data
Dom/Imp
x2csi(i,s,j) % change. Investment demand for i,s by industry j National
Model

297

Dom/Imp
REGSH3(i,r) Share of region r in consumption of i

DR assume households in region r account for x
percent of national household consumption of
commodity i (dom & imp) if region r accounts for x
per cent of national disposable income.

T2 (disposable
income)
BAS3(i,s) Household use of i,s, nat‘l National
Model Data
x3_CSR(i,s,r) % change. Reg‘l. Consumption of i,s by households in
r
T3 (dom/imp)
SOURCEDOM(s) One for s = domestic, zero for s = imported
REGSH4(i,r) Share of region r as a port for exports of i See T4
BAS4(i) Exports of i, national National
Model Data
x4_CR(i,r) % change. Reg‘l. Exports of i leaving US from region
r
T4
REGSH5(i,r) Share of region r in govt consumption of i
―Used regional output shares…assumed that demands
for delivery from region r are satisfied by production
in region r‖
MSSM
BAS5(i,s) Government consumption of i,s, national National
Model Data
x5_CSR(i,s,r) % change. Consumption of i,s in r by governments T5 Dom/Imp
d_x6_CSR(i,s,r) Change T6 Dom/Imp
T10: Use of margin commodity m to facilitate flows of i,s from r

(All,m,MARG)(All,i,NMARG)(All,s,SOURCE)(All,r,REG)

xmargf_MNSR(m,i,s,r) = xmarg_MCS(m,i,s) + x0_CSR(i,s,r) – sum
g
{WTF(m,i,s,g) *
x0_CSR(i,s,g)
xmargf_MNSR(m,i,s
,r)
% change. Reg‘l. Use of margin commodity m to
facilitate the flow of i,s from r
T10
xmarg_MCS(m,i,s) % change. Reg’l. Margin demand for m by agents
in r
National
Model
x0_CSR(i,s,r) % change. Reg‘l. Supply of i,s from region r T14 r
WTF(m,i,s,g) Share of the margin use of m on flows of i,s
associated with i,s flows from r

= DMF(m,i,s,r)/ sum
d
{DMF(m,i,s,d)

DMF – Margin use of m on flows of i,s from r

x0_CSR(i,s,g) % change. Reg‘l. Supply of i,s from region g T14 g
T11: Use of margin commodity m to facilitate flows of i,s to r

(All,m,MARG)(All,i,NMARG)(All,s,SOURCE)(All,r,REG)

298

xmargt_MNSR(m,i,s,r) = xmarg_MCS(m,i,s) + dem_CSR(i,s,r) – sum{WTT(m,i,s,g) *
dem_CSR(i,s,g)
xmargt_MNSR(m,i,s,r) % change. Reg‘l Use of margin commodity m to
facilitate the flow of i,s to r
T11
xmarg_MCS(m,i,s) % change. Reg’l.
dem_CSR(i,s,r) % change. Reg‘l. Demand for i,s in region r T9 MARG
T13 NMARG
T12: Margin demands for margin commodity m by agents in r

(All,m,MARG)(All,r,REG)

xmarg_MR(m,r) =

(1/DM(m,r)) * sum
iNMARG
sum
s
{BETA(m)*DMF(m,i,s,r) *xmargf_MNSR(m,i,s,r) + (1-
BETA(m))*DMT(m,i,s,r) *xmargt_MNSR(m,i,s,r)}

―We assume that agent in each region r are responsible for organizing the reaction BETA(m) of
the margin services of type m required on the flows of i,s from r. Consequently we assume that
agents in each region r are responsible for organizing the fraction [1-BETA(m)] of the margin
services of type m required on the flows of i,s to r. In the illustrative application in section 3,
BETA(m) is 0.5 for all m except retail trade. For retail trade we assume that all th margin
service is organized by the receiving region, i.e. BETA(retail) = 0. With the BETA(m)s
assumed constant, we obtain T12, in which the percentage change in the demand for margin
service m by agents in region r is a weighted average of the percentage changes in the demands
for m to be used in facilitating flows from and to region r. The weights BETA, DMF, DM are
the shares of the total margin demand for m in r that are associated with flows of i,s from r and
flows of i,s to r.‖

xmarg_MR(m,r) % change. Margin demand for m by agents in r T12
DM(m,r) Margin demand for m by agents in r.
BETA(m) Fraction of the margin us of m on flows from a region
that is organized by agents in the region.

DMF(m,i,s,r) Margin use of m on flows of i,s from r
xmargf_MNSR(m,i,s,r) % change. Reg‘l. Use of margin commodity m to
facilitate the flow of i,s from r
T10
DMT(m,i,s,r) Margin use of m on flows of i,s to r See above
xmargt_MNSR(m,i,s,r) % change. Reg‘l. Use of margin commodity m to
facilitate the flow of i,s to r
T11
T13: Demand for margin commodity i from source s by agents in r

(All,i,MARG)(All,s,SOURCE)(All,r,REG)

dem_CSR(i,s,r) =

(1/TOTDEMREG(i,s,r)) *

{(sum
j
{JOBSH(j,r)} *BAS1(i,s,j) * [x1csi(i,s,j) + x0ind_IR(j,r) – sum
g
{JOBSH(j,g) *

299

x0ind_IR(j,g)]

+ sum
j
{JOBSH(j,r) * BAS2(i,s,j) * [x2csi(i,s,j) + x0ind_IR(j,r) – sum
g
{JOBSH(j,g) *
x0ind_IR(j,g)]

+ REGSH3(i,r) * BAS3(i,s) * x3_CSR(i,s,r)

+ SOURCEDOM(s) * REGSH4(i,r) * BAS4(i) * x4_CR(i,r)

+ REGSH5(i,r) * BAS5(i,s) *x5_CSR(i,s,r)

+ 100* d_x6_CSR(i,s,r) + SOURCEDOM(s)*DM(i,r)*xmarg_MR(i,r)}

i.e. very similar to T9 – see definition and data above. ―Has an extra term covering demands in
region r for commodity i from the domestic source (s=dom) to be used as a margin service.
Margin services connect producers (or ports of entry in the case of imports) to users. Margin
commodities include transport, retail trade, and wholesale trade. All demands for margin
services are satisfied by domestic production and margin commodities can be used not only as
margin services but also directly (e.g. air transport used to move employees between work
sites).

― The weights applied to different demands on the RHSs of T9 and T13 are the shares of each
demand in total demand. For example, the weight given to the percentage change in region r‘s
household demand for i,s [x3_CSR(i,s,r)] in determining the percentage change in r‘s total
demand for i,s is the share of household consumption in r‘s total demand for i,s. This share in
calculated as HHShare(i,s,r) = REGSH3(i,r)*BAS3(i,s)/TOTDEMREG(i,s,r)…Thus the
numerator is household consumption of i,s in r. The denominator is the total demand for i,s in r.

―The only exceptions to the use of shares as weights on the RHSs of T9 and T13 are for
inventories. Because the inventory variable [d_x6_CSR(i,s,r)] refers to a change (not a
percentage change), the appropriate coefficients in T9 and T13 are 100/TOTDEMREG(i,s,r).‖

For intermediate demand and investment ―we assume that the percentage change in
intermediate demand for i,s by industry j in region r varies from that by industry j for the nation
to the extent that the percentage change in j‘s output in region r [x0ind_IR(j,r)] varies from the
percentage change in j‘s output for the nation. The percentage change in industry j‘s output for
the nation is calculated as a weighted average of the percentage changes in j‘s outputs in the
regions. Similarly, we assume that the differences between the percentage changes in j‘s
investment demands for i,s at the regional level reflect differences between percentage changes
in j‘s output at the regional and national levels.‖
dem_CSR(i,s,r) % change. Reg‘l. Demand for i,s in region r. MARG T13
SOURCEDOM(s) One for s = domestic, zero for s = imported
DM(i,r) Margin demand for i by agents in r
xmarg_MR(i,r) % change. Margin demand for i by agents in r. T12
T14: Supply of commodity i from source s out of region r

(All,i,COM)(All,s,SOURCE)(All,r,REG)

300

x0_CSR(i,s,r) =

(1/SCSR(i,s,r)) * sum
g
{SHIN(i,s,r,g)*TOTDEMREG(i,s,g) *

[dem_CSR(i,s,g)+shin_CSRR(i,s,r,g) – sum
k
{SHIN(i,s,k,g) * shin_CSRR(i,s,k,g)}

―T14 determines the percentage change in the supply of i,s from region r. For domestic
commodities supply is output in region r. For imported commodities, supply is the volume of
imports of i coming into the US via ports in region r.‖
Assume percentage change in demand for i,s from r by region g is given by last bracketed
function, ―that is, g‘s demand for i,s from r moves with g‘s demand for i,s [dem_CSR(i,s,r)]
and with percentage changes [shin_CSRR(i,s,r,r)] in the share of g‘s demands for i,s satisfied
from r. The share changes, shin_CSRR(i,s,r,g) are normally set exogenously. The final term on
the RHS ensures that exogenous movements in the shin_CSRR(i,s,r,g)s cannot lead to a
violation of the adding-up condition…With the dem(i,s,r,g)s given, T14 determines the
percentage change in the supply of i,s from r as a weighted average of the dem(i,s,r,g)s across
all regions g. The weights, SHIN(i,s,r,g)*TOTDEMREG(i,s,g)/SCSR(i,s,r) are regional shares
in the demand for i,s supplied from r.
x0_CSR(i,s,r) % change.Reg‘l. Supply of i,s from region r T14
SCSR(i,s,r) Supply of i,s from region r See above
SHIN(i,s,r,g) Share of region r in satisfying region g‘s demand for i,s See above
TOTDEMREG(i,s,g) Demand for i,s in region g See T9
dem_CSR(i,s,g) % change. Reg‘l Demand for i,s in region g T9, T13
shin_CSRR(i,s,r,g) % change. Reg‘l. Share of region r in satisfying region g‘s
demand for i,s (usually exogenous)

SHIN(i,s,k,g) Share of region r in satisfying region g‘s demand for i,s See above
shin_CSRR(i,s,k,g) % change.

301

APPENDIX G: SENATORS AND REPRESENTATIVES WHO “BREAK RANK” ON
CLIMATE BILLS
Table G1: Ideology Influence on Senate Voting, Bills with Explicit Climate Change Component.
Independent Variables
Year Bill
Liberal Index
Liberal Index (LI) and Non-
Climate Environment Bill
Voting (NoCC)
R
2
Coeff. R
2
Coeff.
2011 Global Warming Pollution 0.82 0.017*** 0.86 LI: 0.007***
NoCC:0.033***
2009 National Security and Climate
Change
0.78 0.016*** 0.91 LI: 0.0001
NoCC:0.0112***
2009 Polar Bear Protections and Global
Warming
0.78 0.016*** 0.91 LI: 0.000
NoCC:0.012***
2008 Global Warming 0.65 0.014*** 0.71 LI: 0.006*
NoCC:0.007***
2007 CAFÉ & Energy Efficiency I 0.32 0.010*** 0.36 LI: 0.001
NoCC:0.008**
2007 CAFÉ & Energy Efficiency II 0.31 0.009*** 0.43 LI: 0.001
NoCC:0.009***
2007 CAFÉ & Clean Energy 0.55 0.131*** 0.63 LI: 0.001
NoCC:0.012***
2007 Water Resources – Global Warming 0.55 0.013*** 0.65 LI: 0.004
NoCC:0.010***
***Significant at the 0.001 level; ** Significant at the 0.01 level; *Significant at the 0.1 level.
Table G2: Ideology Influence on House Voting, Bills with Explicit Climate Change Component.
Independent Variables
Year Proposal
Liberal Index
Liberal Index (LI) and
Non-Climate Environment
Bill Voting (NoCC)
R
2
Coeff. R
2
Coeff.
2011 Global Warming Pollution 0.77 0.016*** 0.91 LI: -0.002*
NoCC:0.012***
2011 Climate Change Adaptation 0.75 0.016*** 0.89 LI: -0.0001
NoCC:0.011***
2009 Climate & Clean Energy 0.71 0.015*** 0.72 LI: 0.008***
NoCC:0.005***
2007 Global Warming & National Security 0.77 0.016*** 0.83 LI: 0.005***
NoCC:0.008***
2007 Reduce Global Warming 0.67 0.014*** 0.79 LI: 0.001
NoCC:0.011***
***Significant at the 0.001 level; ** Significant at the 0.01 level; *Significant at the 0.1 level.

302

Table G3: Ideology Influence on Senate Voting, Bills with Content Related to Climate Change.
Independent Variables
Year Proposal
Liberal Index
Liberal Index (LI) and Non-
Climate Environment Bill
Voting (NoCC)
R
2
Coeff. R
2
Coeff.
2011 Ending Ethanol Subsidies 0.00 0.000 0.26 LI: -0.022***
NoCC:0.077***
2010 Biodiesel Tax Credit 0.12 0.006*** 0.11 LI: 0.002
NoCC:0.003
2010 Funding Renewable Energy 0.63 0.015*** 0.68 LI: 0.007**
NoCC:0.005***
2009 Clean Energy Recovery 0.77 0.016*** 0.91 LI: 0.0001
NoCC:0.012***
2008 Clean Energy Tax Credits I 0.71 0.015*** 0.85 LI: 0.003
NoCC:0.012***
2008 Clean Energy Tax Credits II 0.74 0.015*** 0.85 LI: 0.003
NoCC:0.011***
2008 Energy Prices 0.69 0.015*** 0.82 LI: 0.001
NoCC:0.013***
2008 Low-Income Energy Assistance 0.70 0.015*** 0.79 LI: 0.003
NoCC:0.011***
2007 Undermining Renewable Electricity 0.57 0.013*** 0.75 LI: -0.001
NoCC:0.013***
2007 Oil Subsidies Repeal 0.54 0.013*** 0.60 LI: 0.002
NoCC:0.010***
2007 Biofuels 0.49 0.012*** 0.58 LI: 0.001
NoCC:0.010***
***Significant at the 0.001 level; ** Significant at the 0.01 level; *Significant at the 0.1 level.
Table G4: Ideology Influence on House Voting, Bills with Content Related to Climate Change.
Independent Variables
Year Proposal
Liberal Index
Liberal Index (LI) and
Non-Climate Environment
Bill Voting (NoCC)
R
2
Coeff. R
2
Coeff.
2011 Flood Insurance 0.01 0.001** 0.04 LI: -0.003*
NoCC:0.003**
2011 Light Bulb Energy Efficiency
Standards
0.75 0.016*** 0.86 LI: 0.002*
NoCC:0.032***
2011 Energy Efficiency and Renewable
Energy
0.17 0.006*** 0.16 LI: 0.004*
NoCC:0.001
2010 Energy Efficiency 0.74 0.015*** 0.82 LI: 0.003**
NoCC:0.009***
2009 Clean Energy Recovery 0.78 0.016*** 0.86 LI: 0.002*
NoCC:0.009***
2008 Clean Energy Tax Credits I 0.76 0.015*** 0.78 LI: 0.005***
NoCC:0.093***

303

2008 Undermining Clean Energy Tax
Credits
0.79 0.016*** 0.79 LI: 0.010***
NoCC:
0.005***
2008 Clean Energy Tax Credits II 0.70 0.015*** 0.70 LI: 0.005***
NoCC:
0.007***
2008 Gutting Renewable Energy 0.75 0.016*** 0.77 LI: 0.007***
NoCC: 0.07***
2007 Renewable Electricity Standard 0.56 0.013*** 0.58 LI: 0.003**
NoCC:
0.008***
2007 Clean Energy 0.73 0.015*** 0.81 LI: 0.002*
NoCC:
0.010***
2007 CAFÉ & Clean Energy 0.77 0.016*** 0.81 LI: 0.006***
NoCC:
0.008***
***Significant at the 0.001 level; ** Significant at the 0.01 level; *Significant at the 0.1 level.

Abstract (if available)

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The spatial economic impact of live music in Orange County, CA

PDF

Household carbon footprints: how to encourage adoption of emissions‐reducing behaviors and technologies

PDF

Climate change, air pollution and health: how policy can induce adaptation

PDF

The role of leadership and collaboration as a catalyst for regional economic development: a grounded theory study

PDF

Philanthropic foundations and social impact bonds: understanding impact investment approaches among philanthropic foundations

PDF

Civic associations, local governance and conflict prevention in Indonesia

PDF

A better method for measuring housing affordability and the role that affordability played in the mobility outcomes of Latino-immigrants following the Great Recession

PDF

Measuring the drivers of economic, energy, and environmental changes: an index decomposition analysis

PDF

Structure, agency, and the Kuznets Curve: observations and implications for sustainability planning in U.S. cities

PDF

A framework for evaluating urban policy and its impact on social determinants of health (SDoH)

PDF

A time of crisis: the Australian experience and what can California learn?

PDF

The impact of mobility and government rental subsidies on the welfare of households and affordability of markets

PDF

Intradepartmental collaboration in the public organizations: implications to practice in an era of resource scarcity and economic uncertainty

PDF

Why go green? Cities' adoption of local renewable energy policies and urban sustainability certifications

PDF

Global health diplomacy: a new era of health in U.S. foreign policy

PDF

Organizational structuring for homeland security: a U.S. state-by-state comparative analysis

PDF

Urban universities' campus expansion projects in the 21st century: a case study of the University of Southern Calfornia's "Village at USC" project and its potential economic and social impacts on...

PDF

A view from below: the development and role of organizational social capital in neighborhood regeneration in Los Angeles

PDF

Three essays on the political economy of federal spending

PDF

The role of public policy in the decisions of parents and caregivers: an examination of work, fertility, and informal caregiving

Asset Metadata

Creator Prager, Fynnwin (author)

Core Title The economic and political impacts of U.S. federal carbon emissions trading policy across households, sectors and states

School School of Policy, Planning and Development

Degree Doctor of Philosophy

Degree Program Policy, Planning, and Development

Publication Date 04/29/2013

Defense Date 02/22/2013

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag climate policy,computable general equilibrium modeling,economic impacts,income distribution,OAI-PMH Harvest,political economy,regional equity

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Painter, Gary Dean (committee chair), Rose, Adam Z. (committee member), Sellers, Jefferey M. (committee member)

Creator Email fprager@usc.edu,fynnprager@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-246338

Unique identifier UC11288061

Identifier etd-PragerFynn-1613.pdf (filename),usctheses-c3-246338 (legacy record id)

Legacy Identifier etd-PragerFynn-1613.pdf

Dmrecord 246338

Document Type Dissertation

Rights Prager, Fynnwin

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA