*• _ *
l ®!
PRO"^
Regulatory Impact Analysis for the Final
Revised Cross-State Air Pollution Rule
(CSAPR) Update for the 2008 Ozone NAAQS

-------
ii

-------
EPA-452/R-21-002
March 2021
Regulatory Impact Analysis for the Final Revised Cross-State Air Pollution Rule (CSAPR)
Update for the 2008 Ozone NAAQS
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Health and Environmental Impacts Division
Research Triangle Park, NC
in

-------
CONTACT INFORMATION
This document has been prepared by staff from the Office of Air Quality Planning and
Standards, U.S. Environmental Protection Agency. Questions related to this document should be
addressed to Robin Langdon, U.S. Environmental Protection Agency, Office of Air Quality
Planning and Standards, C439-02, Research Triangle Park, North Carolina 27711 (email:
langdon.robin@epa.gov).
ACKNOWLEDGEMENTS
In addition to EPA staff from the Office of Air Quality Planning and Standards, personnel from
the Office of Atmospheric Programs and the Office of Policy's National Center for
Environmental Economics contributed data and analysis to this document.
iv

-------
TABLE OF CONTENTS
LIST OF TABLES	ix
LIST OF FIGURES	xiii
EXECUTIVE SUMMARY	ES-1
Overview	ES-1
ES. 1 Identifying Needed Emissions Reductions and Description of the Remedy	ES-2
ES.2 Baseline and Analysis Years	ES-5
ES.3 Emissions and Air Quality Modeling	ES-6
ES.4 Control Strategies and Emissions Reductions	ES-8
ES.5 Cost Impacts	ES-11
ES.6 Benefits	ES-13
ES.6.1 Health Benefits Estimates	ES-13
ES.6.2 Climate Benefits Estimates	ES-18
ES.6.3 Total Benefits	ES-21
ES.6.4 Unquantified Health and Welfare Benefits Categories	ES-23
ES.7 Results of Benefit-Cost Analysis	ES-24
CHAPTER 1: INTRODUCTION AND BACKGROUND	1-1
Overview	1-1
1.1	Background	1-3
1.1.1	Role of Executive Orders in the Regulatory Impact Analysis	1-4
1.1.2	Alternatives Analyzed	1-4
1.1.3	The Need for Air Quality or Emissions Regulation	1-5
1.2	Overview and Design of the RIA	1-6
1.2.1	Methodology for Identifying Needed Reductions	1-6
1.2.2	States Covered by the Rule	1-8
1.2.3	Regulated Entities	1-8
1.2.4	Baseline and Analysis Years	1-9
1.2.5	Emissions Controls, Emissions, and Cost Analysis Approach	1-10
1.2.6	Benefits Analysis Approach	1-11
1.3	Organization of the Regulatory Impact Analysis	1-11
CHAPTER 2: ELECTRIC POWER SECTOR PROFILE	2-1
Overview	2-1
v

-------
2.1	Background	2-1
2.2	Power Sector Overview	2-1
2.2.1	Generation	2-2
2.2.2	Transmission	2-9
2.2.3	Distribution	2-10
2.3	Sales, Expenses, and Prices	2-10
2.3.1	Electricity Prices	2-11
2.3.2	Prices of Fossil Fuels Used for Generating Electricity	2-15
2.3.3	Changes in Electricity Intensity of the U.S. Economy from 2014 to 2018	2-16
2.4	Deregulation and Restructuring	2-17
CHAPTER 3: EMISSIONS AND AIR QUALITY IMPACTS	3-1
Overview	3-1
3.1 ACE Air Quality Modeling Platform	3-1
3.2. Applying Modeling Outputs to Create Spatial Fields	3-3
3.3	Application of ACE Approach for the Revised CSAPR Update	3-7
3.4	Spatial Distribution of Air Quality Impacts	3-9
3.5	Uncertainties and Limitations of ACE Approach	3-13
3.6	References	3-15
APPENDIX 3A: METHODOLOGY FOR DEVELOPING AIR QUALITY SURFACES
	3A-1
3 A. 1 Air Quality Modeling Platform for the ACE Rule	3 A-l
3A. 1.1 Air Quality Model, Meteorology and Boundary Conditions	3A-l
3 A. 1.2 2011 and 2023 Emissions	3A-3
3 A. 1.3 2011 Model Evaluation for Ozone and PM2.5	3 A-7
3A.2 Source Apportionment Tags	3A-l 1
3A.3 Applying Source Apportionment Contributions to Create Air Quality Fields	3A-13
3A.3.2 Scaling Ratio Applied to Source Apportionment Tags	3A-13
3A.4 Creating Fused Fields Based on Observations and Model Surfaces	3A-16
3 A. 5 References	3 A-19
CHAPTER 4: COST, EMISSIONS, AND ENERGY IMPACTS	4-1
Overview	4-1
4.1 Regulatory Control Alternatives	4-1
4.1.2 Regulatory Control Alternatives Analyzed	4-2
vi

-------
4.2	Power Sector Modeling Framework	4-5
4.3	EPA's Power Sector Modeling of the Baseline Run and Three Regulatory Control
Alternatives	4-7
4.3.1 EPA's IPM Baseline Run v.6	4-8
4.3.2. Methodology for Evaluating the Regulatory Control Alternatives	4-8
4.3.3 Methodology for Estimating Compliance Costs	4-14
4.4	Estimated Impacts of the Regulatory Control Alternatives	4-17
4.4.1	Emission Reduction Assessment	4-17
4.4.2	Impact of Emissions Reductions on Maintenance and Nonattainment Monitors 4-26
4.4.3	Compliance Cost Assessment	4-26
4.4.4	Impacts on Fuel Use, Prices and Generation Mix	4-29
4.5	Social Costs	4-37
4.6	Limitations	4-38
4.7	References	4-40
CHAPTER 5: BENEFITS	5-1
Overview	5-1
5.1	Estimated Human Health Benefits	5-2
5.2.1	Health Impact Assessment for Ozone and PM2.5	5-5
5.2.1.1	Selecting Air Pollution Health Endpoints to Quantify	5-8
5.2.1.2	Calculating Counts of Air Pollution Effects Using the Health Impact Function
	5-11
5.2.1.3	Quantifying Ozone-Attributable Premature Mortality	5-12
5.2.1.4	Quantifying PM2.5-Attributable Premature Mortality	5-13
5.2.2	Economic Valuation Methodology for Health Benefits	5-15
5.2.3	Characterizing Uncertainty in the Estimated Benefits	5-17
5.2.4	Estimated Number and Economic Value of Health Benefits	5-21
5.2	Estimated Climate Benefits from Reducing CO2	5-27
5.3	Total Benefits	5-36
5.4	Unquantified Benefits	5-38
5.4.1	NO: Health Benefits	5-40
5.4.2	Ozone Welfare Benefits	5-41
5.4.3	NO2 Welfare Benefits	5-41
5.4.4	Visibility Impairment Benefits	5-42
5.5	References	5-43
vii

-------
CHAPTER 6: STATUTORY AND EXECUTIVE ORDER REVIEWS	6-1
Overview	6-1
6.1	Executive Order 12866: Regulatory Planning and Review	6-1
6.2	Paperwork Reduction Act	6-1
6.3	Regulatory Flexibility Act	6-1
6.3.1	Identification of Small Entities	6-3
6.3.2	Overview of Analysis and Results	6-7
6.3.2.1	Methodology for Estimating Impacts of the Revised CSAPR Update on Small
Entities	6-7
6.3.2.2	Results	6-9
6.3.3	Summary of Small Entity Impacts	6-11
6.4	Unfunded Mandates Reform Act	6-11
6.5	Executive Order 13132: Federalism	6-12
6.6	Executive Order 13175: Consultation and Coordination with Indian Tribal
Governments	6-12
6.7	Executive Order 13045: Protection of Children from Environmental Health & Safety
Risks 	6-14
6.8	Executive Order 13211: Actions that Significantly Affect Energy Supply, Distribution,
or Use 	6-14
6.9	National Technology Transfer and Advancement Act	6-14
6.10	Executive Order 12898: Federal Actions to Address Environmental Justice in Minority
Populations and Low-Income Populations	6-14
CHAPTER 7: COMPARISON OF BENEFITS AND COSTS	7-1
Overview	7-1
7.1 Results	7-2
viii

-------
LIST OF TABLES
Table ES-1. Illustrative NOx Ozone Season Emission Budgets (Tons) Evaluated	ES-4
Table ES-2. NOx Mitigation Strategies Represented in Modeling of the Regulatory Control
Alternatives	ES-9
Table ES-3. Estimated 2021, 2025, 2030, 2035, and 2040a EGU Emissions Reductions in the 12
States of NOx, SO2, and CO2 and More and Less Stringent Alternatives (Tons)	ES-10
Table ES-4. National Compliance Cost Estimates (millions of 2016$) for the Regulatory Control
Alternatives	ES-12
Table ES-5. Estimated Avoided Ozone-Related Premature Respiratory Mortality and Illnesses
for the Final and More and Less Stringent Alternatives for 2021 (95% Confidence Interval)
	ES-15
Table ES-6. Estimated Avoided PM2.5 and Ozone-Related Mortality and Illnesses for the Final
and More and Less Stringent Alternatives for 2024 (95% Confidence Interval)	ES-16
Table ES-7. Estimated Discounted Economic Value of Ozone-Attributable Premature Mortality
and Illnesses for the Final Policy Scenarios in 2021 (95% Confidence Interval; millions of
2016$)	ES-17
Table ES-8. Estimated Discounted Economic Value of Avoided Ozone and PM2.5-Attributable
Premature Mortality and Illnesses for the Final Policy Scenario in 2024 (95% Confidence
Interval; millions of 2016$)	ES-18
Table ES-9. Interim Global Social Cost of Carbon Values (2016$/Metric Tonne CO2)	ES-20
Table ES-10. Estimated Total Annual Global Climate Benefits (2021-40) from Changes in CO2
Emissions (Millions of 2016$)	ES-20
Table ES-11. Combined Health Benefits and Climate Benefits for the Final Rule and More and
Less Stringent Alternatives for 2021 (millions of 2016$)	ES-21
Table ES-12. Combined Health Benefits and Climate Benefits for the Final Rule and More and
Less Stringent Alternatives for 2025 (millions of 2016$)	ES-22
Table ES-13. Combined Health Benefits and Climate Benefits for the Final Rule and More and
Less Stringent Alternatives for 2030 (millions of 2016$)	ES-23
Table ES-14. Benefits, Costs, and Net Benefits of the Final Rule and More and Less Stringent
Alternatives for 2021 for the U.S. (millions of 2016$)	ES-25
Table ES-15. Benefits, Costs, and Net Benefits of the Final Rule and More and Less Stringent
Alternatives for 2025 for the U.S. (millions of 2016$)	ES-25
IX

-------
Table ES-16. Benefits, Costs, and Net Benefits of the Final Rule and More and Less Stringent
Alternatives for 2030 for the U.S. (millions of 2016$)	ES-26
Table ES-17. Summary of Annual Values, Present Values and Equivalent Annualized Values for
the 2021-2040 Timeframe for Estimated Compliance Costs, Benefits, and Net Benefits for the
Final Rule (millions of 2016$, discounted to 2021)	ES-28
Table 2-1. Total Net Summer Electricity Generating Capacity by Energy Source, 2014 and 2018
	2-3
Table 2-2. Net Generation in 2014 and 2018 (Trillion kWh = TWh)	2-6
Table 2-3. Coal and Natural Gas Generating Units, by Size, Age, Capacity, and Average Heat
Rate in 2018	2-7
Table 2-4. Total U.S. Electric Power Industry Retail Sales, 2014 and 2018 (billion kWh)	2-11
Table 3 A. 1. Model Performance Statistics by Region for PM2.5	3 A-10
Table 3A.2. Model Performance Statistics by Region for Ozone on Days Above 60 ppb (May-
Sep)	3 A-11
Table 3A.3. Source Apportionment Tags	3A-12
Table 4-1. Illustrative NOx Ozone Season Emission Budgets (Tons) Evaluated	4-4
Table 4-2. NOx Mitigation Strategies Represented in Modeling of the Regulatory Control
Alternatives 	4-12
Table 4-3. Summary of Methodology for Calculating Compliance Costs Estimated Outside of
IPVI for Revised CSAPR Update Final Rule, 2021 (2016$)	4-16
Table 4-4. EGU Ozone Season NOx Emissions and Emissions Changes (thousand tons) for the
Baseline Run and the Regulatory Control Alternatives	4-18
Table 4-5. EGU Annual Emissions and Emissions Changes for NOx, SO2, PM2.5, and CO2 for
the Regulatory Control Alternatives	4-21
Table 4-6. National Compliance Cost Estimates (millions of 2016$) for the Regulatory Control
Alternatives 	4-27
Table 4-7. 2021 Projected U.S. Power Sector Coal Use for the Baseline Run and the Regulatory
Control Alternatives	4-30
Table 4-8. 2021 Projected U.S. Power Sector Natural Gas Use for the Baseline Run and the
Regulatory Control Alternatives	4-30
Table 4-9. 2021 Projected Minemouth and Power Sector Delivered Coal Price for the Baseline
Run and the Regulatory Control Alternatives	4-31
x

-------
Table 4-10. 2021 Projected Henry Hub and Power Sector Delivered Natural Gas Price for the
Baseline Run and the Regulatory Control Alternatives	4-31
Table 4-11. 2021 Projected U.S. Generation by Fuel Type for the Baseline Run and the
Regulatory Control Alternatives	4-31
Table 4-12. 2021 and 2025 Projected U.S. Capacity by Fuel Type for the Baseline Run and the
Regulatory Control Alternatives	4-32
Table 4-13. Average Retail Electricity Price by Region for the Baseline Run and the Regulatory
Control Alternatives, 2021	4-33
Table 4-14. Average Retail Electricity Price by Region for the Baseline Run and the Regulatory
Control Alternatives, 2025	4-35
Table 5-1. Health Effects of Ambient Ozone and PM2.5	5-10
Table 5-2. Estimated Avoided Ozone-Related Premature Respiratory Mortality and Illnesses for
the Final and More and Less Stringent Alternatives for 2021 (95% Confidence Interval)	5-22
Table 5-3. Estimated Avoided PM2.5 and Ozone-Related Mortality and Illnesses for the Final and
More and Less Stringent Alternatives for 2024 (95% Confidence Interval)	5-23
Table 5-4. Estimated Discounted Economic Value of Ozone-Attributable Premature Mortality
and Illnesses for the Final Policy Scenarios in 2021 (95% Confidence Interval; millions of
2016$)	5-24
Table 5-5. Estimated Discounted Economic Value of Avoided Ozone and PM2.5-Attributable
Premature Mortality and Illnesses for the Final Policy Scenario in 2024 (95% Confidence
Interval; millions of 2016$)	5-25
Table 5-6. Stream of Human Health Benefits from 2021 through 2040: Monetized Benefits
Quantified as Sum of Long-Term Ozone Mortality and Long-Term PM2.5 Mortality (Discounted
at 3%; 95% Confidence Interval; millions of 2016$)	5-26
Table 5-7. Stream of Human Health Benefits from 2021 through 2040: Monetized Benefits
Quantified as Sum of Short-Term Ozone Mortality and Long-Term PM2.5 Mortality (Discounted
at 7%; 95% Confidence Interval; millions of 2016$)	5-27
Table 5-8. Interim Global Social Cost of Carbon Values, 2020-2050 (2016$/Metric Tonne CO2)
	5-31
Table 5-9. Estimated Global Climate Benefits from Changes in CO2 Emissions 2021 - 2040
(Millions of 2016$)	5-35
Table 5-10. Combined Health Benefits and Climate Benefits for the Final Rule and More and
Less Stringent Alternatives for 2021 (millions of 2016$)	5-36
XI

-------
Table 5-11. Combined Health Benefits and Climate Benefits for the Final Rule and More and
Less Stringent Alternatives for 2025 (millions of 2016$)	5-37
Table 5-12. Combined Health Benefits and Climate Benefits for the Final Rule and More and
Less Stringent Alternatives for 2030 (millions of 2016$)	5-38
Table 5-13. Unquantified Health and Welfare Benefits Categories	5-39
Table 6-1. SBA Size Standards by NAICS Code	6-6
Table 6-2. Projected Impact of the Revised CSAPR Update on Small Entities in 2021	6-10
Table 6-3. Incremental Annual Costs under the Revised CSAPR Update Summarized by
Ownership Group and Cost Category in 2021 (2016$ millions)	6-11
Table 7-1. Benefits, Costs, and Net Benefits of the Final Rule and More and Less Stringent
Alternatives for 2021 for the U.S. (millions of 2016$)	7-3
Table 7-2. Benefits, Costs, and Net Benefits of the Final Rule and More and Less Stringent
Alternatives for 2025 for the U.S. (millions of 2016$)	7-3
Table 7-3. Benefits, Costs, and Net Benefits of the Final Rule and More and Less Stringent
Alternatives for 2030 for the U.S. (millions of 2016$)	7-4
Table 7-4. Summary of Annual Values, Present Values and Equivalent Annualized Values for
the 2021-2040 Timeframe for Estimated Compliance Costs, Benefits, and Net Benefits for the
Final Rule (millions of 2016$, discounted to 2021)	7-6
Xll

-------
LIST OF FIGURES
Figure 2-1. National New Build and Retired Capacity (MW) by Fuel Type, 2014-2018	2-4
Figure 2-2. Regional Differences in Generating Capacity (MW), 2018	2-5
Figure 2-3. Cumulative Distribution in 2018 of Coal and Natural Gas Electricity Capacity and
Generation, by Age	2-8
Figure 2-4. Fossil Fuel-Fired Electricity Generating Facilities, by Size	2-9
Figure 2-5. Real National Average Electricity Prices (including taxes) for Three Major End-Use
Categories	2-12
Figure 2-6. Relative Increases in Nominal National Average Electricity Prices for Major End-
Use Categories (including taxes), With Inflation Indices	2-13
Figure 2-7. Real National Average Electricity Prices for Three Major End-Use Categories
(including taxes), 1960-2019 (2018$)	2-14
Figure 2-8. Relative Change in Real National Average Electricity Prices (2018$) for Three Major
End-Use Categories (including taxes)	2-14
Figure 2-9. Relative Real Prices of Fossil Fuels for Electricity Generation; Change in National
Average Real Price per MMBtu Delivered to EGU	2-15
Figure 2-10. Relative Growth of Electricity Generation, Population and Real GDP Since 2014 16
Figure 2-11. Relative Change of Real GDP, Population and Electricity Generation Intensity
Since 2014	2-17
Figure 2-12. Status of State Electricity Industry Restructuring Activities	2-18
Figures 2-13. and 2-14. Capacity and Generation Mix by Ownership Type, 2014 & 2018	2-20
Figure 3-1. Air Quality Modeling Domain	3-2
Figure 3-2. Map of change in May-September MDA8 ozone (ppb): 2021 baseline - less
stringent regulatory alternative (scale: + 0.10 ppb)	3-10
Figure 3-3. Map of change in May-September MDA8 ozone (ppb): 2021 baseline - rule (scale:
+ 0.50 ppb)	3-10
Figure 3-4. Map of change in May-September MDA8 ozone (ppb): 2021 baseline - more
stringent regulatory alternative (scale: + 0.50 ppb)	3-11
Figure 3-5. Map of change in May-September MDA8 ozone (ppb): 2024 baseline - less
stringent regulatory alternative (scale: + 0.10 ppb)	3-11
Xlll

-------
Figure 3-6. Map of change in May-September MDA8 ozone (ppb): 2024 baseline - rule (scale:
• 0.50 ppb)	3-12
Figure 3-7. Map of change in May-September MDA8 ozone (ppb): 2024 baseline - more
stringent regulatory alternative (scale: + 0.50 ppb)	3-12
Figure 3-8. Map of change in annual mean PM2.5 (]ag/m3): 2024 baseline - rule (scale: + 0.005
;;g/ms)	3-13
Figure 3A-1. Air Quality Modeling Domain	3A-2
Figure 3 A-2. NOAA Climate Regions	3 A-9
Figure 4-1. Electricity Market Module Regions	4-37
Figure 5-1. Stylized Relationship between the PM2.5 Concentrations Considered in Epidemiology
Studies and our Confidence in the Estimated PM-related Premature Deaths	5-18
Figure 5-2. Estimated Percentage of PM2.5-Related Deaths and Number of Individuals Exposed
by Annual Mean PM2.5 Level in 2024	5-20
Figure 5-3. Frequency Distribution of SC-CO2 Estimates for 2030	5-33
xiv

-------
EXECUTIVE SUMMARY
Overview
This action is taken in response to the United States Court of Appeals for the District of
Columbia Circuit's (D.C. Circuit) September 13, 2019 remand of the Cross-State Air Pollution
Rule (CSAPR) Update.1 The CSAPR Update finalized Federal Implementation Plans (FIPs) for
22 states to address their interstate pollution-transport obligations under the Clean Air Act
(CAA) for the 2008 ozone National Ambient Air Quality Standards (NAAQS).2 The D.C. Circuit
found that the CSAPR Update, which was published on October 26, 2016 as a partial remedy to
address upwind states' obligations prior to the 2018 Moderate area attainment date under the
2008 ozone NAAQS, was unlawful to the extent it allowed those states to continue their
significant contributions to downwind ozone problems beyond the statutory dates by which
downwind states must demonstrate their attainment of the air quality standards. This rule will
resolve 21 states' outstanding interstate ozone transport obligations with respect to the 2008
ozone NAAQS.3
This action finds that for 9 of the 21 states with remanded FIPs (Alabama, Arkansas,
Iowa, Kansas, Mississippi, Missouri, Oklahoma, Texas, and Wisconsin), their projected 2021
ozone season nitrogen oxides (NOx) emissions do not significantly contribute to a continuing
downwind nonattainment and/or maintenance problem; therefore the CSAPR Update fully
addresses their interstate ozone transport obligations with respect to the 2008 ozone NAAQS.
This action also finds that for the 12 remaining states (Illinois, Indiana, Kentucky, Louisiana,
Maryland, Michigan, New Jersey, New York, Ohio, Pennsylvania, Virginia, and West Virginia),
their projected 2021 ozone season NOx emissions significantly contribute to downwind states'
nonattainment and/or maintenance problems for the 2008 ozone NAAQS.
1	Wisconsin v. EPA, 938 F.3d 303 (D.C. Cir. 2019). This action will also have the effect of addressing the
outstanding obligations that resulted from the D.C. Circuit's vacatur of the CSAPR Close-Out in New York v. EPA,
781 Fed. App'x 4 (D.C. Cir. 2019).
2	The 2008 ozone NAAQS is an 8-hour standard that was set at 75 parts per billion (ppb). See 73 FR 16436 (March
27, 2008).
3	In the CSAPR Update, EPA found that the finalized Tennessee emissions budget fully addressed Tennessee's good
neighbor obligation with respect to the 2008 ozone NAAQS. As such, Tennessee is not considered in this rule, and
the number of states included is reduced from 22 to 21 states.
ES-1

-------
EPA is creating an additional geographic group and ozone season trading program
comprised of these 12 upwind states with remaining linkages to downwind air quality problems
in 2021. This new group, Group 3, will be covered by a new CSAPRNOx Ozone Season (May 1
- September 30) Group 3 trading program and will no longer be subject to Group 2 budgets.
Aside from the removal of the 12 covered states from the current Group 2 program, this rule
leaves unchanged the budget stringency and geography of the existing CSAPR NOx Ozone
Season Group 1 and Group 2 trading programs. The electric generating units (EGUs) covered by
the FIPs and subject to the budget are fossil-fired EGUs with >25 megawatt (MW) capacity.
ES.l Identifying Needed Emissions Reductions and Description of the Remedy
To reduce interstate emission transport under the authority provided in CAA section
110(a)(2)(D)(i)(I), this rule further limits ozone season NOx emissions from EGUs in 12 states
using the same framework used by EPA in developing the CSAPR (the interstate transport
framework). The interstate transport framework provides a 4-step process to address the
requirements of the good neighbor provision for ground-level ozone and fine particulate matter
(PM2.5) NAAQS: (1) identifying downwind receptors that are expected to have problems
attaining or maintaining the NAAQS; (2) determining which upwind states contribute to these
identified problems in amounts sufficient to "link" them to the downwind air quality problems
(i.e., here, an amount of contribution equal to or greater than 1 percent of the NAAQS); (3) for
states linked to downwind air quality problems, identifying upwind emissions that significantly
contribute to downwind nonattainment or interfere with downwind maintenance of the NAAQS;
and (4) for states that are found to have emissions that significantly contribute to nonattainment
or interfere with maintenance of the NAAQS downwind, implementing the necessary emissions
reductions through enforceable measures. In this action, EPA applies this 4-step interstate
transport framework to respond to the D.C. Circuit's remand and revise the CSAPR Update with
respect to the 2008 ozone NAAQS.
The remedy that emerges from the 4-step interstate transport framework here are state
emissions budgets for EGUs implemented as an interstate cap-and-trade program.4 This
4 Section X.J. Executive Order 12898: Federal Actions to Address Environmental Justice in Minority Populations
and Low-Income Populations in the preamble describes EPA's ongoing response to the President's environmental
justice commitments.
ES-2

-------
regulatory impact analysis (RIA) evaluates how the EGUs covered by the rule are expected to
reduce their emissions in response to the requirements and flexibilities provided by the remedy
implemented by the Revised CSAPR Update and the estimated benefits, costs, and impacts of
doing so. The rule sets EGU ozone season NOx emissions budgets (allowable emission levels)
for 2021 and future years. EPA implements these reductions through FIPs in any state that does
not have an approved good neighbor SIP by the date this action is finalized. Furthermore, under
the FIPs, affected EGUs would participate in the CSAPR NOx ozone-season allowance trading
program. The allowance trading program essentially converts the EGU NOx emissions budget
for each of the 12 states subject to the FIP into a limited number of NOx ozone-season
allowances that, on a tonnage basis, equal the state's ozone season emissions budget. Starting in
2021, emissions from affected EGUs in the 12 states cannot exceed the sum of emissions budgets
but for the ability to use banked allowances from previous years for compliance. The seasonal
budgets decline in 2022, 2023 and 2024. No further reductions in budgets occur after 2024, and
budgets remain in place for future years. Furthermore, emissions from affected EGUs in a
particular state are subject to the CSAPR assurance provisions, which require additional
allowance surrender penalties (a total of 3 allowances per ton of emissions) on emissions that
exceed a state's CSAPR NOx ozone season assurance level, or 121 percent of the states'
emissions budget. EPA is creating a limited initial bank of allowances for use in the new Group 3
trading program by converting allowances banked in 2017-2020 under the existing Group 2
trading program at a formula-based conversion ratio. The target bank amount is based on the sum
of the states' "variability limits" - that is, the amounts by which emissions from a given state's
units can exceed the state's emission budget before incurring a penalty surrender ratio.
For the Revised CSAPR Update, the EGU ozone season NOx budgets for each state
reflect widely available EGU NOx reduction strategies including optimizing existing SCR and
SNCR, as well as installing state-of-the-art combustion controls, at an estimated representative
cost of $1,800 per ton. Furthermore, this RIA analyzes more and less stringent regulatory control
alternatives at estimated representative NOx control costs of $9,600 per ton and $500 per ton,
respectively. Table ES-1 shows the illustrative EGU NOx ozone season emission budgets that
are evaluated in this RIA.
All three scenarios are illustrative in nature, and the budgets included in the Revised
CSAPR Update scenario differ slightly from the budgets finalized in this rule. That is because
ES-3

-------
subsequent to completing the analysis of these three scenarios, EPA decided to set the budgets
using a different approach than assumed for the analysis. In particular, the modeling presented in
the RIA assumes that SNCR optimization is available in 2022, whereas the final budgets confirm
that SNCR optimization is available in 2021. The estimated incremental emission reductions
would be 1,163 tons if EPA had used the actual 2021 budgets, or a 1.1 percent tightening in the
modeled budget across the Group 3 states. The choice of 2021 or 2022 for the initial year of
SNCR optimization is an exogenous input into the analysis. EPA finds that the three illustrative
regulatory control alternatives presented in this RIA provide a reasonable approximation of the
impacts of the final rule, as well as an evaluation of the relative impacts of two regulatory
alternatives. This finding is supported by an analysis of the costs and impacts (but not the
benefits) of the final Revised CSAPR Update emission budgets that assumes 2021 for the initial
year of SNCR optimization and provided in the docket for this rulemaking.
Table ES-1. Illustrative NOx Ozone Season Emission Budgets (Tons) Evaluated
Final Rule
State
2021
2022
2023
2024
2025
Illinois
9,198
9,102
8,179
8,059
8,059
Indiana
13,085
12,582
12,553
9,564
9,564
Kentucky
15,307
14,051
14,051
14,051
14,051
Louisiana
15,389
14,818
14,818
14,818
14,818
Maryland
1,499
1,266
1,266
1,348
1,348
Michigan
12,732
12,290
9,975
9,786
9,786
New Jersey
1,253
1,253
1,253
1,253
1,253
New York
3,416
3,416
3,421
3,403
3,403
Ohio
9,690
9,773
9,773
9,773
9,773
Pennsylvania
8,379
8,373
8,373
8,373
8,373
Virginia
4,614
3,897
3,980
3,663
3,663
West Virginia
13,686
12,884
12,884
12,884
12,884
Total
108,248
103,703
100,525
96,974
96,974
Less-Stringent Alternative
State
2021
2022
2023
2024
2025
Illinois
9,348
9,348
8,393
8,272
8,272
Indiana
15,677
15,206
15,179
12,083
12,083
Kentucky
15,606
15,606
15,606
15,606
15,606
Louisiana
15,430
15,430
15,430
15,430
15,430
Maryland
1,501
1,267
1,267
1,350
1,350
Michigan
13,126
12,688
10,386
10,188
10,188
New Jersey
1,346
1,346
1,346
1,346
1,346
New York
3,463
3,463
3,468
3,450
3,450
Ohio
15,487
15,569
15,569
15,569
15,569
Pennsylvania
11,807
11,806
11,806
11,806
11,806
Virginia
4,661
4,270
4,357
4,021
4,021
West Virginia
15,017
15,017
15,017
15,017
15,017
Total
122,468
121,016
117,822
114,138
114,138
ES-4

-------
More-Stringent Alternative5
State
2021
2022
2023
2024
2025
Illinois
9,198
9,102
8,179
6,891
6,891
Indiana
13,085
12,582
12,553
8,430
8,430
Kentucky
15,307
14,051
14,051
9,775
9,775
Louisiana
15,389
14,818
14,818
12,622
12,622
Maryland
1,499
1,266
1,266
1,168
1,168
Michigan
12,732
12,290
9,975
7,344
7,344
New Jersey
1,253
1,253
1,253
1,257
1,257
New York
3,416
3,416
3,421
3,297
3,297
Ohio
9,690
9,773
9,773
9,222
9,222
Pennsylvania
8,379
8,373
8,373
7,851
7,851
Virginia
4,614
3,897
3,980
3,184
3,184
West Virginia
13,686
12,884
12,884
10,568
10,568
Total
108,248
103,703
100,525
81,609
81,609
ES.2 Baseline and Analysis Years
The rule sets forth the requirements to eliminate states' significant contribution to
downwind nonattainment or interference with maintenance of the 2008 ozone NAAQS. To
develop and evaluate control strategies for addressing these obligations, it is important to first
establish a baseline projection of air quality and electricity sector and related fuel market
conditions in the analysis period, taking into account currently on-the-books Federal regulations,
substantial Federal regulatory proposals, enforcement actions, state regulations, population,
expected electricity demand growth, and where possible, economic growth.6 Establishing this
baseline for the analysis then allows us to estimate the incremental costs and benefits of the
additional emissions reductions that will be achieved by the rule.
5	For the illustrative purposes in this RIA, EPA's analytical technique for assessing the more stringent alternative
presents emission reduction values incremental to the final rule's stringency prior to 2025. This does not reflect a
determination that new SCR controls could be installed on a fleetwide basis before the 2025 ozone season. See
sections VLB. 1, C. 1, and D. 1 of the preamble for further discussion.
6	The technical support document (TSD) for the 2016vl emissions modeling platform titled Preparation of
Emissions Inventories for 2016vl North American Emissions Modeling Platform is included in the docket for this
rule. The TSD includes additional discussion on mobile source rules included in the baseline. The future year onroad
emission factors account for changes in activity data and the impact of on-the-books rules that are implemented into
MOVES2014b. These rules include the Light Duty Vehicle GHG Rule for Model-Year 2017-2025 and the Tier 3
Motor Vehicle Emission and Fuel Standards Rule. Local inspection and maintenance (I/M) and other onroad mobile
programs are included, such as California LEVHI, the National Low Emissions Vehicle (LEV) and Ozone Transport
Commission (OTC) LEV regulations, local fuel programs, and Stage II refueling control programs. Regulations
finalized after the year 2014 are not included, such as the Safer Affordable Fuel Efficient (SAFE) Vehicles Final
Rule for Model Years 2021-2026 and the Final Rule for Phase 2 Greenhouse Gas Emissions Standards and Fuel
Efficiency Standards for Medium- and Heavy-Duty Engines and Vehicles.
ES-5

-------
The analysis in this RIA focuses on benefits, costs and certain impacts from 2021 through
2040. We focus on 2021 because it is by the 2021 ozone season, corresponding with the 2021
Serious area attainment date, that significant contribution from upwind states' must be
eliminated to the extent possible. It is also the first year in which some EGU NOx mitigation
technologies are available. In addition, impacts for 2022 to 2025 are important as these years
reflect the years in which additional NOx mitigation technologies are first available. Since
retrofits are assumed to have a 15-year book life and the budgets remain in place for future years
after 2025, costs and benefits from installation may persist beyond 2025. In order to capture
these streams, the RIA also provides costs through 2040.
Presenting estimated benefits, costs, and certain impacts in 2025 reflects the time needed to
make these retrofits on a regional scale and reflects full implementation of the policy. Additional
benefits and costs are expected to occur after 2025 as EGUs subject to this rule continue to
comply with the tighter allowance budget, which is below their baseline emissions, and these
costs and benefits are reflected in the estimates through 2040 provided in this RIA.
ES.3 Emissions and Air Quality Modeling
The air quality spatial fields for this rule were constructed using the method and air
quality modeling data developed to support the RIA for the Repeal of the Clean Power Plan, and
the Emission Guidelines for Greenhouse Gas Emissions from Existing Electric Utility
Generating Units (U.S. EPA 2019), also referred to the Affordable Clean Energy (ACE) rule.7
The foundational data from this approach includes the ozone contributions from EGU emissions
in each state based on the 2023 ACE EGU state-sector sector contribution modeling and the
2023 emissions for coal and non-coal fired EGUs that were input to that modeling.8
The air quality modeling used in the ACE analysis included annual model simulations for
a 2011 base year and a 2023 future year to provide hourly concentrations of ozone and primary
and secondarily formed PM2.5 component species (e.g., sulfate, nitrate, ammonium, elemental
7	Additional details on the modeling and methodology for developing spatial fields of air quality for EGU control
strategies are provided in Appendix 3 A.
8	The 2023 emissions used for the ACE modeling were derived from the 2011-based emissions platform whereas the
emissions used in the air quality modeling to project ozone design values and contributions for this final rule were
based on the more recent 2016 platform.
ES-6

-------
carbon (EC), organic aerosol (OA), and crustal material9) for both years nationwide. The
photochemical modeling results for 2011 and 2023, in conjunction with modeling to characterize
the air quality impacts from groups of emissions sources {i.e., source apportionment modeling)
and emissions data for the baseline and regulatory control alternatives, were used to construct the
air quality spatial fields that reflect the influence of emissions changes between the baseline and
the regulatory control alternatives.
The air quality model simulations {i.e., model runs) were performed using the
Comprehensive Air Quality Model with Extensions (CAMx) f. Our CAMx nationwide modeling
domain {i.e., the geographic area included in the modeling) covers all lower 48 states plus
adjacent portions of Canada and Mexico using a horizontal grid resolution of 12 x 12 km.
To estimate ozone-related benefits in 2021 and 2024 EPA applied the approach used in the
ACE final RIA using as input the ozone season EGU NOx emissions (tons) for the 2021 and
2024 baseline along with emissions for the rule and each of the two other regulatory control
alternatives. These emissions were applied using this approach and source apportionment data to
produce spatial fields of the May-September seasonal average MDA8 ozone and the April-
October seasonal average MDA1 ozone concentrations as described in Chapter 3.10 Similarly, to
estimate PM2.5 benefits in 2024 EPA applied the same approach using as input the non-ozone
season EGU NOx emissions (tons) for the 2024 baseline along with emissions for the rule and
each of the two other regulatory control alternatives.11 To estimate benefits in the years beyond
2024, the same 2021 and 2024 air quality surfaces were linearly extrapolated or interpolated to
2040.12 NOx reductions in the ozone season provide minimal PM2.5 benefits since PM2.5 nitrate
concentrations, which result from conversion of NOx emissions to nitrate, are minimal during
the warmer temperatures during the ozone season. Conversely, the conversion of nitrates to
9	Crustal material refers to metals that are commonly found in the earth's crust such as Aluminum, Calcium, Iron,
Magnesium, Manganese, Potassium, Silicon, Titanium and the associated oxygen atoms.
10	MDA8 is defined as maximum daily 8-hour average ozone concentration, and MDA1 is defined as the maximum
daily 1-hour ozone concentration.
11	See Chapter 5, section 5.2.1.2 Calculating Counts of Air Pollution Effects Using the Heath Impact Function for
additional discussion. We use BenMAP-CE to quantify individual risk and counts of estimated premature deaths and
illnesses attributable to photochemical modeled changes in summer season average ozone concentrations for the
year 2021, and summer season average ozone concentrations and annual mean PM2 5 for the year 2024 using a health
impact function.
12	Ozone air quality was modeled in 2021 and 2024, while the formation of PM2 5 was modeled only in 2024. This
assumes that ozone and PM2 5 formation reaches a steady state beyond 2024, which may or may not be the case
particularly the further out beyond 2024.
ES-7

-------
PM2.5 is much greater in cooler (non-ozone season) months, and thus it becomes worthwhile to
estimate PM2.5 benefits from NOx reductions in those months. As there no expected NOx
emissions changes outside of the ozone season in 2021, PM2.5 concentrations are not evaluated
for this model year. The changes in emissions of SO2 and directly emitted PIVh.swere partially
analyzed for this rule as described in Chapter 4. In particular, EPA did not analyze changes in
emissions due to generation shifting for either SO2 or PM2.5 because the majority of the
reductions associated with this rulemaking are tied to the operation of SCR and SNCR controls
(which do not affect SO2 emission rates) and state of the art combustion controls (which have a
minimal impact on SO2 emission rates). Additionally in order to meet the court-ordered timeline
for this rulemaking EPA prioritized fully capturing the impact of reductions from generation
shifting on NOx and CO2, but did not account for the relatively small amount of SO2 and primary
PM emissions reductions that would likely occur due to generation shifting. Hence total benefits
could be higher than those reported in this RIA.
ES.4 Control Strategies and Emissions Reductions
Before undertaking power sector analysis to evaluate compliance with the regulatory
control alternatives, EPA first considered available EGUNOx mitigation strategies that could be
implemented for the upcoming ozone season (i.e., the 2021 ozone season). EPA considered all
widely-used EGU NOx control strategies: optimizing NOx removal by existing, operational
selective catalytic reduction (SCRs) and turning on and optimizing existing idled SCRs;13
optimizing existing idled selective non-catalytic reduction (SNCRs); installation of (or upgrading
to) state-of-the-art NOx combustion controls; shifting generation to units with lower NOx
emission rates; and installing new SCRs and SNCRs. Similarly, as proposed, EPA determined
that the power sector could implement most of these NOx mitigation strategies, except
installation of state-of-the-art NOx combustion controls and new SCRs or SNCRs, for the 2021
ozone-season.
The EGU NOx mitigation strategies that are assumed to operate or are available to reduce
NOx in response to each of the regulatory control alternatives are shown in Table ES-2. EPA
13 Units may choose to idle SCRs and SNCRs in order to avoid fixed operation and maintenance (FOM) and variable
operation and maintenance (VOM) costs such as auxiliary fan power, catalyst costs, and additional administrative
costs (labor), depending on the prevailing CSAPR allowance price for those units otherwise not required to attain a
NOx emission rate that would require operating their SCRs and SNCRs more intensively.
ES-8

-------
analyzed ozone-season NOx emission reductions and the associated costs to the power sector of
implementing the EGU NOx ozone-season emissions budgets in each of the 12 states using the
Integrated Planning Model (IPM) and its underlying data and inputs as well as certain costs that
are estimated outside the model but use IPM inputs for their estimation. This analysis is also
used to identify changes in other air pollutants from the power sector that arise as a consequence
of complying with the budgets.
Table ES-2. NOx Mitigation Strategies Represented in Modeling of the Regulatory Control
Alternatives
Regulatory Control
Alternative
NOx Controls Implemented
Less Stringent Alternative	(1) Shift generation to minimize costs (costs estimated within IPM)	
(All controls above)
(2) Fully operating existing SCRs to achieve 0.08 lb/MMBtu NOx emission rate
(costs estimated outside IPM)
Final Rule	(3) Turn on idled SCRs (costs estimated outside IPM) and fully operate akin to
(2)
(4) Fully operate existing SNCRs (costs estimated outside IPM)
	(5) Install state-of-the-art combustion controls.	
More Stringent Alternative
(All controls above)
(6) In 2025, impose state emission limits commensurate with installation of new
SCRs on units without such controls. However, additional SCRs controls are
not a least-cost compliance strategy, (costs estimated within IPM)	
For the NOx controls identified in Table ES-2, under the rule and the more stringent
alternative, 47 units, not already doing so in 2019, are projected to fully operate existing SCRs, 4
units are projected to turn on idled SCRs, and 29 units are projected to fully operate existing
SNCRs. Under the less stringent alternative, no units are projected to either fully operate existing
SNCRs, SCRs or turn on idled SCRs. Under the rule and the more stringent alternative, 10 units
are projected to install state-of-the-art combustion controls, and under the less stringent
alternative no units are projected to install state-of-the-art combustion controls. The book-life of
the controls is assumed to be 15 years. Under the rule, the more stringent alternative, and the less
stringent alternative, no units are projected to install new SCRs.14 The book-life of the new SCRs
is assumed to be 15 years. For additional details, see the EGU NOx Mitigation Strategies Final
Rule TSD.
14 In the RIA for the proposed rule, units were exogenously forced to install SCR controls in IPM. In the modeling
for the final rule, the choice to install SCR controls was endogenous to the model, and no incremental SCR
installations occurred, with the model relying on greater levels of generation shifting instead.
ES-9

-------
Table ES-3 shows the emissions reductions expected from the rule in 2021, 2025, 2030,
2035, and 2040 as well as the more and less stringent alternatives analyzed.
Table ES-3. Estimated 2021, 2025, 2030, 2035, and 2040a EGU Emissions Reductions in the
12 States of NOx, SO2, and CO2 and More and Less Stringent Alternatives (Tons)b
2021
Final Rule
More Stringent
Alternative
Less Stringent
Alternative
NOx (annual)
16,000
16,000
2,000
NOx (ozone season)
16,000
16,000
2,000
SO2 (annual) * * *
CO2 (annual, thousand metric)
2025
NOx (annual)
21,000
37,000
2,000
NOx (ozone season)
19,000
34,000
2,000
SO2 (annual) * * *
CO2 (annual, thousand metric)
5,000
14,000
4,000
2030
NOx (annual)
16,000
27,000
2,000
NOx (ozone season)
13,000
25,000
2,000
SO2 (annual) * * *
CO2 (annual, thousand metric)
8,000
19,000
6,000
2035
NOx (annual)
15,000
26,000
2,000
NOx (ozone season)
13,000
25,000
2,000
SO2 (annual) * * *
CO2 (annual, thousand metric)
8,000
19,000
6,000
2040
NOx (annual)
14,000
25,000
2,000
NOx (ozone season)
13,000
24,000
2,000
SO2 (annual) * * *
CO2 (annual, thousand metric)
4,000
13,000
3,000
a The 2021-2040 emissions reductions estimates are based on IPM projections for CO2 and engineering analysis for
annual and ozone season NOx. SO2 and PM2 5 emissions were only partially analyzed. IPM was run for the
following years: 2021, 2023, 2025, 2030, 2035, 2040, 2045 and 2050. For more information, see Chapter 4 and the
Ozone Transport Policy Analysis Final Rule TSD.
b NOx emissions are reported in English (short) tons; CO2 is reported in metric tons.
*There are no annual SO2 and PM2 5 emissions reductions that come from turning on SCRs and SNCRs assuming
that nothing else changes, but EPA did not analyze the effects on SO2 and direct PM that may come from shifting
power generation, for example from coal-fired power plants to gas-fired or other types of power plants. EPA does
expect some changes in SO2 and PM2 5 emissions due to shifting of power generation.
The results of EPA's analysis show that, with respect to compliance with the EGU NOx emission
budgets in 2021, maximizing the use of existing operating SCRs provides the largest amount of
ozone season NOx emission reductions (47 percent, affecting 47 units), and turning on idled
SCRs produces an additional 37 percent (affecting 4 units) of the total ozone season NOx
reductions. Generation shifting primarily from coal to gas generation (16 percent) makes up the
ES-10

-------
remainder of the ozone season NOx reductions. EPA relied on Engineering Analysis to account
for changes in NOx (annual and ozone season), SO2, and direct PM. While this approach
captures the impact of generation shifting for NOx emissions, it does not fully capture the impact
of generation shifting for SO2 and PM in complying with the Revised CSAPR Update budgets.
Additionally in order to meet the court-ordered timeline for this rulemaking EPA prioritized fully
capturing the impact of reductions from generation shifting on NOx and CO2, but did not account
for the relatively small amount of SO2 and primary PM emissions reductions that would likely
occur due to generation shifting. Hence total benefits could be higher than those reported in this
RIA. EPA relied on IPM estimates to capture changes in CO2 emissions, which fully account for
the impact of generation shifting.
If EPA were to have included SNCR optimization in 2021, with respect to compliance
with the EGU NOx emission budgets in 2021, maximizing the use of existing operating SCRs
provides the largest amount of ozone season NOx emission reductions (44 percent, affecting 47
units), turning on idled SCRs produces an additional 35 percent (affecting 4 units) of the total
ozone season NOx reductions, and fully operating existing SNCRs produces an additional 6
percent of reductions (affecting 29 units). Generation shifting primarily from coal to gas
generation (15 percent) makes up the remainder of the ozone season NOx reductions.
ES.5 Cost Impacts
The estimates of the changes in the cost of supplying electricity for the regulatory control
alternatives are presented in Table ES-4. Since the rule does not result in any additional
recordkeeping, monitoring or reporting requirements, the costs associated with compliance with
monitoring, recordkeeping, and reporting requirements are not included within the estimates in
this table.
There are several notable aspects of the results presented in Table ES-4. The most notable
result is that the estimated annual compliance cost for the less stringent alternative is negative
(i.e., a cost reduction) in 2025, although this regulatory control alternative reduces NOx
emissions by 2,000 tons as shown in Table ES-3. While seemingly counterintuitive, estimating
negative compliance costs in a single year is possible given the assumption of perfect foresight in
IPM. IPM's objective function is to minimize the discounted net present value (NPV) of a stream
ES-11

-------
of annual total cost of generation over a multi-decadal time period. For example, with the
assumption of perfect foresight it is possible that on a national basis within the model the least-
cost compliance strategy may be to delay a new investment or economic retirement that was
projected to occur sooner in the baseline. Such a delay could result in a lowering of annual cost
in an early time period and increase it in later time periods. Since the less-stringent alternative is
designed to include only generation shifting, it does not necessitate full operation of existing
controls, or installation of new controls, leading to a negative annual cost estimate in 2025.
Table ES-4. National Compliance Cost Estimates (millions of 2016$) for the Regulatory
Control Alternatives

Final Rule
More-Stringent Alternative
Less-Stringent Alternative
2021-2025 (Annualized)
10.0
41.4
-2.9
2021-2040 (Annualized)
24.8
28.5
19.6
2021 (Annual)
5.1
5.2
1.6
2025 (Annual)
1.6
4.0
-14.9
2030 (Annual)
63.6
32.3
67.0
2035 (Annual)
18.2
41.2
14.3
2040 (Annual)
8.8
134.0
18.9
The 2021-2025 (Annualized) row reflects total estimated annual compliance costs levelized over the period 2021
through 2025, discounted using a 4.25 real discount rate.15 The 2021-2040 (Annualized) row reflects total estimated
annual compliance costs levelized over the period 2021 through 2040, and discounted using a 4.25 real discount
rate. This does not include compliance costs beyond 2040. The 2021 (Annual), 2025 (Annual), 2030 (Annual), 2035
(Annual), and 2040 (Annual) rows reflect annual estimates in each of those years.
Under the Revised CSAPR Update, fully operating existing SCR controls provides a large
share of the total emissions reductions. These options are selected in 2021, while upgrading to
state-of-the-art combustion controls is assumed to begin in 2022.16 Generation shifting costs are
positive in 2021, but negative in 2025. The result is that the costs in 2021 are higher than costs in
2025. Projected costs for the illustrative Revised CSAPR Update peak in 2030 at $63.6 million
(2016$) and annualized costs for the 2021-40 period are $24.8 million (2016$).
15 This table reports compliance costs consistent with expected electricity sector economic conditions. An NPV of
costs was calculated using a 4.25% real discount rate consistent with the rate used in IPM's objective function for
cost-minimization. The NPV of costs was then used to calculate the levelized annual values over a 5-year period
(2021-2025) and over a 20-year period (2021-2040) using the 4.25% rate as well. Tables ES-17 and 7-4 report the
NPV of the annual stream of costs from 2021-2040 using 3% and 7% consistent with OMB guidance.
ES-12

-------
ES.6 Benefits
ES. 6.1 Health Benefits Estimates
The Revised CSAPR Update is expected to reduce concentrations of ground-level ozone,
PM2.5, and CO2 in the atmosphere (see Chapter 3 for discussion). Reducing NOx emissions
generally reduces human exposure to ozone and the incidence of ozone-related health effects,
though the degree to which ozone is reduced will depend in part on local levels of VOCs as
discussed in Chapter 3. The rule would also reduce emissions of NOx throughout the year.
Because NOx is also a precursor to formation of ambient PM2.5, reducing these emissions would
reduce human exposure to ambient PM2.5 throughout the year and would reduce the incidence of
PM2.5-attributable health effects.17 Reducing emissions of NOx would also reduce ambient
exposure to NO2 and its associated health effects.
EPA historically has used evidence reported in the Integrated Science Assessment (ISA)
for the most recent NAAQS review to inform its approach for quantifying air pollution-
attributable health, welfare, and environmental impacts associated with that pollutant. The ISA
synthesizes the toxicological, clinical and epidemiological evidence to determine whether each
pollutant is causally related to an array of adverse human health outcomes associated with either
short-term (hours to less than one month) or long-term (one month to years) exposure; for each
outcome, the ISA reports this relationship to be causal, likely to be causal, suggestive of a causal
relationship, inadequate to infer a causal relationship or not likely to be a causal. We estimate the
incidence of air pollution-attributable premature deaths and illnesses using recently updated
methods, reflecting the new evidence reported in the 2019 PM2.5 and 2020 Ozone IS As and
accounting for recommendations from the Science Advisory Board. When updating each health
endpoint EPA considered: (1) the extent to which there exists a causal relationship between that
pollutant and the adverse effect; (2) whether suitable epidemiologic studies exist to support
quantifying health impacts; (3) and whether robust economic approaches are available for
estimating the value of the impact of reducing human exposure to the pollutant. Detailed
descriptions of these updates are available in the TSD for the Final Revised Cross-State Air
17 This RIA does not quantify PM2 5-related benefits associated with SO2 emission reductions. As discussed in
Chapter 4, EPA does not estimate significant SO2 emission reductions as a result of this action. Additionally, this
RIA does not estimate changes in emissions of directly emitted particles.
ES-13

-------
Pollution Rule for the 2008 Ozone NAAQS Update titled Estimating PM2.5- and Ozone-
Attributable Health Benefits.
Table ES-5 and Table ES-6 report the estimated number of reduced premature deaths and
illnesses in each year relative to the baseline along with the 95% confidence interval. The
number of estimated reduced deaths and illnesses from the final rule and more and less stringent
alternatives are calculated from the sum of individual reduced mortality and illness risk across
the population. Table ES-7 and Table ES-8 report the estimated economic value of avoided
premature deaths and illness in each year relative to the baseline along with the 95% confidence
interval. In each of these tables, for each discount rate and regulatory control alternative,
multiple benefits estimates are presented reflecting alternative ozone and PM2.5 mortality risk
estimates.
ES-14

-------
Table ES-5. Estimated Avoided Ozone-Related Premature Respiratory Mortality and
Illnesses for the Final and More and Less Stringent Alternatives for 2021 (95%
	Confidence Interval) a'b	


Final Rule
More Stringent
Alternative
Less Stringent
Alternative
Avoided premature respiratory mortality
Long-
term
Turner etal. (2016)°
190
190
18
exposure

(130 to 250)
(130 to 250)
(13 to 24)

Katsouyanni el al.
(2009)cd and Zanobetti el
9
9
1
Short-
al. (2008)'1 pooled
(4 to 14)
(4 to 14)
(0 to 1)
term
exposure
Katsouyanni el al.
(2009)cd
9
(-5 to 21)
9
(-5 to 21)
1
(-0 to 2)

Zanobetti et al. (2008)e
9
(4 to 13)
9
(4 to 13)
1
(0 to 1)
Morbidity effects
Long-
term
exposure
Asthma onset6
1,300
(1,100 to 1,500)
1,300
(1,100 to 1,500)
130
(110 to 150)
Allergic rhinitis
symptoms8
7,700
(4,000 to 11,000)
7,700
(4,000 to 11,000)
750
(400 to 1,100)

Hospital admissions—
21
21
2

respiratory"1
(-5 to 46)
(-5 to 46)
(-1 to 5)
Short-
term
ED visits—respiratoryf
440
(120 to 920)
440
(120 to 920)
43
(12 to 90)
Asthma symptoms
240,000
(-30,000 to 500,000)
240,000
(-30,000 to 500,000)
24,000
(-2,900 to 49,000)
exposure
Minor restricted-activity
120,000
120,000
12,000

daysd-f
(49,000 to 200,000)
(49,000 to 200,000)
(4,800 to 19,000)

School absence days
91,000
(-13,000 to 190,000)
91,000
(-13,000 to 190,000)
8,900
(-1,300 to 19,000)
" Values rounded to two significant figures.
b We estimated changes in annual mean PM25 and PM2 5 -related benefits in 2024, but not 2021. As discussed in
Chapter 4, in 2021, the only control measure expected to be adopted for compliance in the regulatory control
alternatives is optimization of existing SCRs, and this measure will operate only during the ozone season. As
discussed in Chapter 3, NOx reductions in the ozone season provide minimal PM2 5 benefits since PM2 5 nitrate
concentrations, which result from conversion of NOx emissions to nitrate, are minimal during the warmer
temperatures during the ozone season. Conversely, the conversion of nitrates to PM2 5 is much greater in cooler (non-
ozone season) months, and thus it becomes worthwhile to estimate PM2 5 benefits from NOx reductions in those
months. In 2024, the presence of additional control measures that operate year-round and other changes in market
conditions as a result of the rule lead to notable NOx reductions in the winter months.
0 Applied risk estimate derived from April-September exposures to estimates of O3 across the standard May-
September warm season.
d Converted O3 risk estimate metric from MDA1 to MDA8.
e Applied risk estimate derived from June-August exposures to estimates of O3 across the standard May-September
warm season.
f Applied risk estimate derived from full year exposures to estimates of O3 across the standard May-September warm
season.
g Converted O3 risk estimate metric from DA24 to MDA8.
ES-15

-------
Table ES-6. Estimated Avoided PM2.5 and Ozone-Related Mortality and Illnesses for the
Final and More and Less Stringent Alternatives for 2024 (95% Confidence
	Interval) a	


Final
More Stringent
Alternative
Less Stringent
Alternative
Avoided premature mortality



in
^ Long-term
Turner et al. (2016)b
4
(2 to 5)
4
(2 to 5)
	
^ exposure
Di et al. (2017)
4
(3 to 4)
4
(3 to 4)

Long-term
exposure
Turner etal. (2016)
230
(160 to 300)
410
(280 to 530)
19
(13 to 25)
ss
0
N
O Short-term ¦
exposure
Zanobetti et al (2008)° and
Katsouyanni et al (2009)b d
pooled
10
(4 to 16)
18
(7 to 29)
1
(0 to 1)
Katsouyanni et al (2009)b d
10
(-6 to 26)
18
(-10 to 46)
1
(-0 to 2)

Zanobetti et al (2008)°
10
(5 to 16)
18
(8 to 28)
1
(0 to 1)
PM2.5- related non-fatal heart attacks among adults


Short-term
Peters et al. (2001)
4
(1 to 6)
4
(1 to 6)

exposure
Pooled estimate
0
(0 to 1)
0
(0 to 1)

Morbidity effects
Asthma onsetbd
(PM2 5 & 03)
1,600
(1,400 to 1,800)
2,900
(2,500 to 3,300)
130
(110 to 150)

Allergic rhinitis
symptoms6
(PM2.5 & 03)
9,200
(4,900 to 13,000)
17,000
(8,700 to 24,000)
770
(400 to 1,100)
Long-term
exposure
Stroke (PM25)
0.2
(0 to 0.3)
0.2
(0 to 0.3)

Lung cancer (PM2 5)
0.4
(0.0 to 0.7)
0.4
(0.0 to 0.7)


Hospital Admissions -
Alzheimer's disease
(PM25)
2
(1 to 2)
2
(1 to 2)


Hospital Admissions-
Parkinson's disease
(PM2.5)
0.2
(0.1 to 0.3)
0.2
(0.1 to 0.3)


Hospital admissions-
cardiovascular (PM2 5)
0.5
(0.3 to 0.6)
0.5
(0.3 to 0.6)


ED visits- cardiovascular
(PM25)
1
(-0 to 2)
1
(0 to 2)

Short-term
Hospital admissions—
respiratory11 (PM2 5 & O3)
27
(-7 to 60)
49
(-12 to 110)
2
(-1 to 5)
exposure
ED visits —respiratory
(PM2 5 & 03)
530
(150 to 1,100)
950
(260 to 2,000)
44
(12 to 92)

Asthma symptomsf
(PM2 5 & 03)
290,000
(-37,000 to
610,000)
530,000
(-66,000 to
1,100,000)
25,000
(-3,000 to 51,000)
ES-16

-------
Minor restricted-activity
days (PM2 5 & O3)
150,000
(60,000 to 230,000)
260,000
(110,000 to
410,000)
12,000
(4,800 to 19,000)
Cardiac arrest (PM2 5)
0.05
(-0.02 to 0.11)
0.05
(-0.02 to 0.11)
--
Lost work days
380
380
--
(PM25)
(320 to 440)
(320 to 440)

School absence days
(03)
110,000
(-15,000 to
200,000
(-28,000 to
9,100
230,000)
410,000)
(-1,300 to 19,000)
a Values rounded to two significant figures.
b Applied risk estimate derived from April-September exposures to estimates of O3 across the standard May-
September warm season.
0 Converted O3 risk estimate metric from MDA1 to MDA8
d Applied risk estimate derived from June-August exposures to estimates of O3 across the standard May-September
warm season.
e Converted O3 risk estimate metric from DA24 to MDA8
f Applied risk estimate derived from full year exposures to estimates of O3 across the standard May-September warm
season.
Table ES-7. Estimated Discounted Economic Value of Ozone-Attributable Premature
Mortality and Illnesses for the Final Policy Scenarios in 2021 (95% Confidence
	Interval; millions of 2016$)a'b	


Final Rule
More Stringent Alternative
Less Stringent Alternative
3%
Discount
Rate
$230
($58 to
$480)°
$1,900 |
($210 to $5,000)d |
$260
($88 to
$520)°
and
$1,900
($210 to
$5,000)d
1 $22
| ($6 to
1 $47)°
and
$190
($20 to
$490)d
7%
Discount
Rate
$200
($38 to
$460)°
$1,700 |
($170 to $4,500)d |
$200
($38 to
$460)°
and
$1,700
($170 to
$4,500)d
|
|
1 $20
| ($4 to
1 $45)°
and
$170
($17 to
$440)d
a Values rounded to two significant figures. The two benefits estimates are separated by the word "and" to signify
that they are two separate estimates. The estimates do not represent lower- and upper-bound estimates and should
not be summed.
b We estimated changes in annual mean PM2 5 and PM25 -related benefits in 2024, but not 2021. As discussed in
Chapter 4, in 2021, the only control measure expected to be adopted for compliance in the regulatory control
alternatives is optimization of existing SCRs, and this measure will operate only during the ozone season. As
discussed in Chapter 3, NOx reductions in the ozone season provide minimal PM2 5 benefits since PM2.5 nitrate
concentrations, which result from conversion of NOx emissions to nitrate, are minimal during the warmer
temperatures during the ozone season. Conversely, the conversion of nitrates to PM2 5 is much greater in cooler (non-
ozone season) months, and thus it becomes worthwhile to estimate PM2 5 benefits from NOx reductions in those
months. In 2024, the presence of additional control measures that operate year-round and other changes in market
conditions as a result of the rule lead to notable NOx reductions in the winter months.
0 Sum of ozone mortality estimated using the pooled Katsouyanni et al. (2009) and Zanobetti and Schwartz (2008)
short-term risk estimate and the Di et al. (2017) long-term mortality risk estimate. As PM-related mortality
quantified using risk estimates from the Di et al. (2017) and Turner et al. (2016) are within 5% of one another, in the
interest of clarity and simplicity, we present the results estimated using the risk estimate from Di et al. (2017) alone.
d Sum of ozone mortality estimated using the long-term risk estimate and the Di et al. (2017) long-term mortality
risk estimate. As PM-related mortality quantified using risk estimates from the Di et al. (2017) and Turner et al.
(2016) are within 5% of one another, in the interest of clarity and simplicity, we present the results estimated using
the risk estimate from Di et al. (2017) alone.
ES-17

-------
Table ES-8. Estimated Discounted Economic Value of Avoided Ozone and PM2.5-
Attributable Premature Mortality and Illnesses for the Final Policy Scenario in
	2024 (95% Confidence Interval; millions of 2016$)a	
Final Rule
More Stringent Alternative
Less Stringent Alternative1"
3% Discount
Rate
$310
($72 to
$680)°
and
$2,400
($250 to
$6,200)d
$530
($130 to
$l,100)c
and
$4,200
($450 to
$ll,000)d
$22
($6 to $47)°
$190
($20 to
$490)d
7% Discount
Rate
$280
($48 to
$640)°
and
$2,100
($210 to
$5,600)d
$470
($84 to
$l,100)c
and
$3,800
($370 to
$9,900)d
$20 and
($4 to $45)°
$170
($17 to
$440)d
a Values rounded to two significant figures. The two benefits estimates are separated by the word "and" to signify
that they are two separate estimates. The estimates do not represent lower- and upper-bound estimates and should
not be summed.
b No PM-attributable benefits accrue for this scenario.
0 Sum of ozone mortality estimated using the pooled Katsouyanni et al. (2009) and Zanobetti and Schwartz (2008)
short-term risk estimate and the Di et al. (2017) long-term mortality risk estimate. As PM-related mortality
quantified using risk estimates from the Di et al. (2017) and Turner et al. (2016) are within 5% of one another, in the
interest of clarity and simplicity, we present the results estimated using the risk estimate from Di et al. (2017) alone.
d Sum of ozone mortality estimated using the long-term risk estimate and the Di et al. (2017) long-term mortality
risk estimate. As PM-related mortality quantified using risk estimates from the Di et al. (2017) and Turner et al.
(2016) are within 5% of one another, in the interest of clarity and simplicity, we present the results estimated using
the risk estimate from Di et al. (2017) alone.
ES. 6.2 Climate Benefits Estimates
We estimate the climate benefits for this final rulemaking using a measure of the social
cost of carbon (SC-CO2). The SC-CO2 is the monetary value of the net harm to society
associated with a marginal increase in CO2 emissions in a given year, or the benefit of avoiding
that increase. In principle, SC-CO2 includes the value of all climate change impacts, including
(but not limited to) changes in net agricultural productivity, human health effects, property
damage from increased flood risk and natural disasters, disruption of energy systems, risk of
conflict, environmental migration, and the value of ecosystem services. The SC-CO2, therefore,
reflects the societal value of reducing emissions of the gas in question by one metric ton. The
SC-CO2 is the theoretically appropriate value to use in conducting benefit-cost analyses of
policies that affect CO2 emissions.
We estimate the global social benefits of CO2 emission reductions expected from this
final rule using the SC-CO2 estimates presented in the Technical Support Document: Social Cost
of Carbon, Methane, and Nitrous Oxide Interim Estimates under Executive Order 13990 (IWG
2021). These SC-CO2 estimates are interim values developed under Executive Order (E.O.)
13990 for use in benefit-cost analyses until an improved estimate of the impacts of climate
ES-18

-------
change can be developed based on the best available science and economics. These SC-CO2
estimates are the same as those used in the 2016 Final CSAPR Update RIA.
The SC-CO2 estimates used in this analysis were developed over many years, using a
transparent process, peer-reviewed methodologies, the best science available at the time of that
process, and with input from the public. Specifically, an interagency working group (IWG) that
included the EPA and other executive branch agencies and offices used three integrated
assessment models (IAMs) to develop the SC-CO2 estimates and recommended four global
values for use in regulatory analyses. The SC-CO2 estimates were first released in February 2010
and updated in 2013 using new versions of each IAM. In 2015, as part of the response to public
comments received to a 2013 solicitation for comments on the SC-CO2 estimates, the IWG
announced a National Academies of Sciences, Engineering, and Medicine review of the SC-CO2
estimates to offer advice on how to approach future updates to ensure that the estimates continue
to reflect the best available science and methodologies. In January 2017, the National Academies
released their final report, Valuing Climate Damages: Updating Estimation of the Social Cost of
Carbon Dioxide, and recommended specific criteria for future updates to the SC-CO2 estimates,
a modeling framework to satisfy the specified criteria, and both near-term updates and longer-
term research needs pertaining to various components of the estimation process (National
Academies 2017). On January 20, 2021, President Biden issued Executive Order 13990, which
directed the IWG to ensure that the U.S. Government's (USG) estimates of the social cost of
carbon and other greenhouse gases reflect the best available science and the recommendations of
the National Academies (2017). The IWG was tasked with first reviewing the estimates currently
used by the USG and publishing interim estimates within 30 days of E.O. 13990 that reflect the
full impact of GHG emissions, including taking global damages into account.18 The interim SC-
CO2 estimates published in February 2021 are used here to estimate the climate benefits for this
final rulemaking.19
18	The E.O. instructs the IWG to undertake a fuller update of the SC-GHG estimates by January 2022.
19	The values used in the proposal RIA were interim values developed under E.O. 13783 (signed March 28, 2017)
for use in regulatory analyses. E.O. 13783 withdrew the TSDs used in the benefits analysis of the 2016 CSAPR
Update (U.S. EPA, 2016b) for describing the global social cost of greenhouse gas estimates as no longer
representative of government policy. Thus, EPA followed E.O. 13783 by using SC-CO2 estimates reflecting impacts
occurring within U.S. borders and 3% and 7% discount rates in our central analysis for the proposal RIA.
ES-19

-------
Table ES-9 summarizes the interim global SC-CO2 estimates for the years 2015 to 2050.
These estimates are reported in 2016 dollars but are otherwise identical to those presented in the
IWG's 2016 TSD (IWG 2016a). For purposes of capturing uncertainty around the SC-CO2
estimates in analyses, we emphasize the importance of considering all four of the SC-CO2
values. The SC-CO2 increases over time within the models - i.e., the societal harm from one
metric ton emitted in 2030 is higher than the harm caused by one metric ton emitted in 2025 -
because future emissions produce larger incremental damages as physical and economic systems
become more stressed in response to greater climatic change, and because GDP is growing over
time and many damage categories are modeled as proportional to GDP.
Table ES-9. Interim Global Social Cost of Carbon Values (2016$/Metric Tonne CO2)
Emissions
Year
2020
2025
2030
2035
2040
2045
2050
Discount Rate and Statistic
5%
Average
$13
$15
$18
$20
$23
$26
$29
3%
Average
$47
$52
$57
$63
$67
$73
$78
2.5%
Average
$71
$77
$83
$90
$95
$100
$110
3%
95th Percentile
$140
$160
$170
$190
$210
$220
$240
Note: These SC-CO2 values are identical to those used in the 2016 Final CSPAR Update RIA adjusted for inflation
to 2016 dollars using the annual GDP Implicit Price Deflator values in the U.S. Bureau of Economic Analysis'
(BEA) NIPA Table 1.1.9 (U.S. BEA 2021). The values are stated in $/metric tonne CO2 (1 metric tonne equals
1.102 short tons) and vary depending on the year of CO2 emissions. This table displays the values rounded to the
nearest dollar; the annual unrounded values used in the calculations in this RIA are available on OMB's website:
. Source: Technical Support Document: Social Cost of Carbon, Methane,
and Nitrous Oxide Interim Estimates under Executive Order 13990 (February 2021).
Table ES-10 shows the estimated monetary value of the estimated changes in CO2
emissions expected to occur over 2021 - 2040 for the Revised CSAPR Update, the more-
stringent alternative, and the less-stringent alternative. EPA estimated the dollar value of the
C02-related effects for each analysis year between 2021 and 2040 by applying the SC-CO2
estimates, shown in Table ES-9, to the estimated changes in CO2 emissions in the corresponding
year under the regulatory options.
Table ES-10. Estimated Total Annual Global Climate Benefits (2021-40) from Changes in
CO2 Emissions (Millions of 2016$)
Regulatory Alternative
Year
5% Discount
Rate
3% Discount
Rate
2.5%
Discount
Rate
3% Discount
Rate (95th
Percentile)
Final
2021
0
1
1
2
ES-20

-------

2022
46
143
206
434

2023
94
290
417
882

2024
102
311
444
946

2025
109
331
473
1,011

2030
128
373
525
1,146

2035
98
273
380
838

2040
127
340
467
1,043

2021
1
2
3
7

2022
76
237
341
720

2023
156
480
689
1,460
More-Stringent
2024
204
623
892
1,898
Alternative
2025
254
771
1,100
2,350

2030
323
939
1,322
2,885

2035
316
878
1,222
2,698

2040
383
1,025
1,410
3,146

2021
0
1
1
3

2022
39
122
176
371

2023
80
248
356
754
Less-Stringent
2024
81
248
355
755
Alternative
2025
82
248
353
755

2030
93
271
381
831

2035
73
203
282
623

2040
91
242
333
743
Note: We emphasize the importance and value of considering the benefits calculated using all four SC-CO2
estimates. As discussed in Chapter 5 and in the Technical Support Document: Social Cost of Carbon, Methane,
and Nitrous Oxide Interim Estimates under Executive Order 13990 (IWG 2021), a consideration of climate
benefits calculated using discount rates below 3 percent, including 2 percent and lower, are also warranted when
discounting intergenerational impacts.
ES. 6.3 Total Benefits
Table ES-11 through Table ES-13 present the total health and climate benefits for the
final rule and the more and less stringent alternatives.
Table ES-11. Combined Health Benefits and Climate Benefits for the Final Rule and More
	and Less Stringent Alternatives for 2021 (millions of 2016$)a

Health and Climate Benefits
Climate Benefits
Onlyb
SC-CO2 Discount
Rate and Statistic
(Discount Rate Applied to Health
Benefits)

3%
7%

Final Rule
5% (average)
$230 and $1,900
$200 and $1,700
$0
3% (average)
$230 and $1,900
$200 and $1,700
$1
2.5% (average)
$230 and $1,900
$200 and $1,700
$1
3% (95th percentile)
$230 and $1,900
$200 and $1,700
$2
More Stringent Alternative
5% (average)
$260 and $1,900
$200 and $1,700
$1
3% (average)
$260 and $1,900
$200 and $1,700
$2
2.5% (average)
$260 and $1,900
$200 and $1,700
$3
ES-21

-------

Health and Climate Benefits
Climate Benefits
Onlyb
SC-CO2 Discount
Rate and Statistic
(Discount Rate Applied to Health
Benefits)

3%
7%

3% (95th percentile)
$270 and $1,900
$210 and $1,700
$7
Less Stringent Alternative
5% (average)
$20 and $190
$20 and $170
$0
3% (average)
$20 and $190
$20 and $170
$1
2.5% (average)
$20 and $190
$20 and $170
$1
3% (95th percentile)
$20 and $190
$20 and $170
$3
a The two benefits estimates are separated by the word "and" to signify that they are two separate estimates. The
estimates do not represent lower- and upper-bound estimates and should not be summed.
b Climate benefits are based on changes (reductions) in CO2 emissions and are calculated using four different
estimates of the social cost of carbon (SC-CO2) (model average at 2.5 percent, 3 percent, and 5 percent discount
rates; 95th percentile at 3 percent discount rate). For purposes of this table, we show the benefits associated with the
model average at a 3 percent discount rate. However, we emphasize the importance and value of considering the
benefits calculated using all four SC-CO2 estimates; the additional benefit estimates range from $0.24 million to
$2.31 million in 2021 for the finalized option. As discussed in Chapter 5, a consideration of climate benefits
calculated using discount rates below 3 percent, including 2 percent and lower, are also warranted when discounting
intergenerational impacts.
Table ES-12. Combined Health Benefits and Climate Benefits for the Final Rule and More
	and Less Stringent Alternatives for 2025 (millions of 2016$)a

Health and Climate Benefits
Climate
Benefits Onlyb
SC-CO2 Discount
Rate and Statistic
(Discount Rate Applied to Health
Benefits)

3%
7%

Final Rule
5% (average)
$430 and $2,500
$400 and $2,300
$110
3% (average)
$650 and $2,700
$620 and $2,500
$330
2.5% (average)
$790 and $2,900
$760 and $2,700
$470
3% (95th percentile)
$1,300 and $3,400
$1,300 and $3,200
$1,000
More Stringent Alternative
5% (average)
$790 and $4,500
$740 and $4,000
$250
3% (average)
$1,300 and $5,000
$1,300 and $4,600
$770
2.5% (average)
$1,600 and $5,300
$1,600 and $4,900
$1,100
3% (95th percentile)
$2,900 and $6,600
$2,900 and $6,200
$2,400
Less Stringent Alternative
5% (average)
$100 and $280
$100 and $250
$80
3% (average)
$270 and $450
$270 and $420
$250
2.5% (average)
$370 and $550
$370 and $520
$350
3% (95th percentile)
$780 and $960
$780 and $930
$760
a The two benefits estimates are separated by the word "and" to signify that they are two separate estimates. The
estimates do not represent lower- and upper-bound estimates and should not be summed.
b Climate benefits are based on changes (reductions) in CO2 emissions and are calculated using four different
estimates of the social cost of carbon (SC-CO2) (model average at 2.5 percent, 3 percent, and 5 percent discount
ES-22

-------
rates; 95th percentile at 3 percent discount rate). For purposes of this table, we show the benefits associated with the
model average at a 3 percent discount rate. However, we emphasize the importance and value of considering the
benefits calculated using all four SC-CO2 estimates; the additional benefit estimates range from $ 109 million to
$1,011 million in 2025 for the finalized option. As discussed in Chapter 5, a consideration of climate benefits
calculated using discount rates below 3 percent, including 2 percent and lower, are also warranted when discounting
intergenerational impacts.
Table ES-13. Combined Health Benefits and Climate Benefits for the Final Rule and More
	and Less Stringent Alternatives for 2030 (millions of 2016$)a

Health and Climate Benefits
Climate
Benefits Onlyb
SC-CO2 Discount
Rate and Statistic
(Discount Rate Applied to Health
Benefits)

3%
7%

Final Rule
5% (average)
$470 and $2,700
$460 and $2,600
$130
3% (average)
$710 and $3,000
$700 and $2,900
$370
2.5% (average)
$870 and $3,100
$860 and $3,000
$530
3% (95th percentile)
$1,400 and $3,700
$1,430 and $3,600
$1,100
More Stringent Alternative
5% (average)
$910 and $4,900
$880 and $4,200
$320
3% (average)
$1,500 and $5,500
$1,500 and $4,800
$940
2.5% (average)
$1,900 and $5,900
$1,900 and $5,200
$1,300
3% (95th percentile)
$3,500 and $7,500
$3,500 and $6,800
$2,900
Less Stringent Alternative
5% (average)
$120 and $300
$110 and $270
$90
3% (average)
$300 and $480
$290 and $450
$270
2.5% (average)
$410 and $590
$400 and $560
$380
3% (95th percentile)
$860 and $1,040
$850 and $1,010
$830
a The two benefits estimates are separated by the word "and" to signify that they are two separate estimates. The
estimates do not represent lower- and upper-bound estimates and should not be summed.
b Climate benefits are based on changes (reductions) in CO2 emissions and are calculated using four different
estimates of the social cost of carbon (SC-CO2) (model average at 2.5 percent, 3 percent, and 5 percent discount
rates; 95th percentile at 3 percent discount rate). For purposes of this table, we show the benefits associated with the
model average at a 3 percent discount rate. However, we emphasize the importance and value of considering the
benefits calculated using all four SC-CO2 estimates; the additional benefit estimates range from $128 million to
$1,146 million in 2030 for the finalized option. As discussed in Chapter 5, a consideration of climate benefits
calculated using discount rates below 3 percent, including 2 percent and lower, are also warranted when discounting
intergenerational impacts.
ES. 6.4 Unquantified Health and Welfare Benefits Categories
Data, time, and resource limitations prevented EPA from quantifying the estimated
impacts or monetizing estimated benefits associated with exposure to NO2 (independent of the
role NO2 plays as precursors to PM2.5), as well as ecosystem effects, and visibility impairment
due to the absence of air quality modeling data for these pollutants in this analysis. Lack of
ES-23

-------
quantification does not imply that there are no benefits associated with reductions in exposures to
ozone, PM2.5, or NO2. For a qualitative description of these benefits, please see Chapter 5,
section 5.4, Table 5-13.
ES.7 Results of Benefit-Cost Analysis
Below in Table ES-14 through Table ES-16, we present the annual costs and benefits
estimates for 2021, 2025, and 2030, respectively. We present the costs and benefits for the years
2021 through 2040 at real discount rates of 3 and 7 percent in Table ES-17. This analysis uses
annual compliance costs reported above as a proxy for social costs. The net benefits of the rule
and more and less stringent alternatives reflect the benefits of implementing EGU emissions
reductions strategies for the affected 12 states via the FIPs minus the costs of those emissions
reductions. The estimated social costs to implement the rule, as described in this document, are
approximately $5 million in 2021 and $2 million in 2025 (2016$).
The estimated monetized health benefits from implementation of the rule are
approximately $230 and $1,900 million in 2021 (2016$, based on a real discount rate of 3
percent) with the two values reflecting alternative ozone and PM2.5 mortality risk estimates. For
2025, the estimated monetized health benefits from implementation of the rule are approximately
$320 and $2,400 million (2016$, based on a real discount rate of 3 percent). The estimated
monetized climate benefits are $1 million in 2021 (using a 3 percent real discount rate) and $330
million in 2025.
EPA calculates the net benefits of the rule by subtracting the estimated compliance costs
from the estimated benefits in 2021, 2025, and 2030. The benefits include those to public health
and climate. The two estimates of the benefits and net-benefits for each discount rate reflect
alternative ozone and PM2.5 mortality risk estimates. The annual net benefits of the rule in 2021
(in 2016$) are approximately $230 and $1,900 million using a 3 percent discount rate. The
annual net benefits of the rule in 2025 are approximately $650 and $2,700 using a 3 percent real
discount rate. The annual net benefits of the rule in 2030 are approximately $650 and $2,900
million using a 3 percent real discount rate. Table ES-14 presents a summary of the benefits,
costs, and net benefits of the rule and the more and less stringent alternatives for 2021. Table
ES-15 presents a summary of these impacts for the rule and the more and less stringent
ES-24

-------
alternatives for 2025, and Table ES-16 presents a summary of these impacts for the rule and the
more and less stringent alternatives for 2030.
Table ES-14. Benefits, Costs, and Net Benefits of the Final Rule and More and Less
Stringent Alternatives for 2021 for the U.S. (millions of 2016$)a'b'c
, _ .	More Stringent Less Stringent
Final Rule	... ,.6	... ,?
Alternative	Alternative
Health Benefits (3%)	$230 and $1,900	$260 and $1,900	$20 and $190
Climate Benefits (3%)	$1	$2	$1
Total Benefits	$230 and $1,900	$260 and $1,900	$20 and $190
Costs	$5	$5	$2
Net Benefits	$230 and $1,900	$260 and $1,900	$20 and $190
Health Benefits (7%)	$200 and $1,700	$200 and $1,700	$20 and $170
Climate Benefits (3%)	$1	$2	$1
Total Benefits	$200 and $1,700	$200 and $1,700	$20 and $170
Costs	$5	$5	$2
Net Benefits	$200 and $1,700	$200 and $1,700	$20 and $170
a We focus results to provide a snapshot of costs and benefits in 2021, using the best available information to
approximate social costs and social benefits recognizing uncertainties and limitations in those estimates. The two
benefits estimates are separated by the word "and" to signify that they are two separate estimates. The estimates do
not represent lower- and upper-bound estimates and should not be summed.
b Benefits include those related to public health and climate. The health benefits are associated with several point
estimates and are presented at real discount rates of 3 percent and 7 percent. Climate benefits are based on changes
(reductions) in CO2 emissions and are calculated using four different estimates of the social cost of carbon (SC-CO2)
(model average at 2.5 percent, 3 percent, and 5 percent discount rates; 95th percentile at 3 percent discount rate). For
the presentational purposes of this table, we show the benefits associated with the average SC-CO2 at a 3 percent
discount rate, but the Agency does not have a single central SC-CO2point estimate. We emphasize the importance
and value of considering the benefits calculated using all four SC-CO2 estimates; the additional benefit estimates
range from $0.24 million to $2.31 million in 2021 for the finalized option. Please see Table 5-9 for the full range of
SC-CO2 estimates. As discussed in Chapter 5, a consideration of climate benefits calculated using discount rates
below 3 percent, including 2 percent and lower, are also warranted when discounting intergenerational impacts. The
costs presented in this table are 2021 annual estimates for each alternative analyzed.
0 Rows may not appear to add correctly due to rounding.
Table ES-15. Benefits, Costs, and Net Benefits of the Final Rule and More and Less
Stringent Alternatives for 2025 for the U.S. (millions of 2016$)a'b'c
Final Rule
Health Benefits (3%)
Climate Benefits (3%)
Total Benefits
Costs
Net Benefits
Health Benefits (7%) $290 and $2,200
Climate Benefits (3%)	$330
Total Benefits $620 and $2,500
More Stringent	Less Stringent
Alternative	Alternative
$540 and $4,200	$20 and $200
$770	$250
$1,300 and $5,000	$270 and $450
$4	-$15
$280 and $460
$490 and $3,800	$20 and $170
$770	$250
$1,300 and $4,600	$270 and $420
$320 and $2,400
$330
$650 and $2,700
$2
$650 and $2,700 $1,300 and $5,000
ES-25

-------
Costs
Net Benefits
$2
$4
$620 and $2,500 $1,300 and $4,500
-$15
$280 and $430
a We focus results to provide a snapshot of costs and benefits in 2025, using the best available information to
approximate social costs and social benefits recognizing uncertainties and limitations in those estimates. The two
benefits estimates are separated by the word "and" to signify that they are two separate estimates. The estimates do
not represent lower- and upper-bound estimates and should not be summed.
b Benefits include those related to public health and climate. The health benefits are associated with several point
estimates and are presented at real discount rates of 3 and 7 percent. Climate benefits are based on changes
(reductions) in CO2 emissions and are calculated using four different estimates of the social cost of carbon (SC-CO2)
(model average at 2.5 percent, 3 percent, and 5 percent discount rates; 95th percentile at 3 percent discount rate). For
the presentational purposes of this table, we show the benefits associated with the average SC-CO2 at a 3 percent
discount rate, but the Agency does not have a single central SC-CO2point estimate. We emphasize the importance
and value of considering the benefits calculated using all four SC-CO2 estimates; the additional benefit estimates
range from $109 million to $1,011 million in 2025 for the finalized option. Please see Table 5-9 for the full range of
SC-CO2 estimates. As discussed in Chapter 5, a consideration of climate benefits calculated using discount rates
below 3 percent, including 2 percent and lower, are also warranted when discounting intergenerational impacts. The
costs presented in this table are 2025 annual estimates for each alternative analyzed.
Table ES-16. Benefits, Costs, and Net Benefits of the Final Rule and More and Less
Stringent Alternatives for 2030 for the U.S. (millions of 2016$)a'b'c
Final Rule
More Stringent
Alternative
Less Stringent
Alternative
Health Benefits (3%)
Climate Benefits (3%)
Total Benefits
Costs
$340 and $2,600
$370
$710 and $3,000
$64
$590 and $4,600
$940
$1,500 and $5,500
$32
$30 and $210
$270
$300 and $480
$67
Net Benefits
$650 and $2,900 $1,500 and $5,500 $230 and $410
Health Benefits (7%)
Climate Benefits (3%)
Total Benefits
Costs
$330 and $2,500
$370
$700 and $2,900
$64
$560 and $3,900
$940
$1500 and $4,800
$32
$20 and $180
$270
$290 and $450
$67
Net Benefits
$640 and $2,800 $1,500 and $4,800 $220 and $380
a We focus results to provide a snapshot of costs and benefits in 2030, using the best available information to
approximate social costs and social benefits recognizing uncertainties and limitations in those estimates. The two
benefits estimates are separated by the word "and" to signify that they are two separate estimates. The estimates do
not represent lower- and upper-bound estimates and should not be summed.
b Benefits include those related to public health and climate. The health benefits are associated with several point
estimates and are presented at real discount rates of 3 and 7 percent. Climate benefits are based on changes
(reductions) in CO2 emissions and are calculated using four different estimates of the social cost of carbon (SC-CO2)
(model average at 2.5 percent, 3 percent, and 5 percent discount rates; 95th percentile at 3 percent discount rate). For
the presentational purposes of this table, we show the benefits associated with the average SC-CO2 at a 3 percent
discount rate, but the Agency does not have a single central SC-CO2point estimate. We emphasize the importance
and value of considering the benefits calculated using all four SC-CO2 estimates; the additional benefit estimates
range from $ 128 million to $ 1,146 million in 2030 for the finalized option Please see Table 5-9 for the full range of
SC-CO2 estimates. As discussed in Chapter 5, a consideration of climate benefits calculated using discount rates
below 3 percent, including 2 percent and lower, are also warranted when discounting intergenerational impacts. The
costs presented in this table are 2030 annual estimates for each alternative analyzed.
0 Rows may not appear to add correctly due to rounding.
ES-26

-------
As part of fulfilling analytical guidance with respect to E.O. 12866, EPA presents
estimates of the present value (PV) of the benefits and costs over the twenty-year period 2021 to
2040. To calculate the present value of the social net-benefits of the Revised CSAPR Update,
annual benefits and costs are discounted to 2021 at 3 percent and 7 discount rates as directed by
OMB's Circular A-4. The EPA also presents the equivalent annualized value (EAV), which
represents a flow of constant annual values that, had they occurred in each year from 2021 to
2040, would yield a sum equivalent to the PV. The EAV represents the value of a typical cost or
benefit for each year of the analysis, in contrast to the year-specific estimates mentioned earlier
in the RIA.
For the twenty-year period of 2021 to 2040, the PV of the net benefits, in 2016$ and
discounted to 2021, is $8,800 and $41,000 million when using a 3 percent discount rate and
$7,300 and $29,000 million when using a 7 percent discount rate. The EAV is $590 and $2,800
million per year when using a 3 percent discount rate and $570 and $2,700 million when using a
7 percent discount rate. The comparison of benefits and costs in PV and EAV terms for the rule
can be found in Table ES-17. Estimates in the table are presented as rounded values and are
based on air quality simulations run for years 2021 and 2024.
ES-27

-------
Table ES-17. Summary of Annual Values, Present Values and Equivalent
Estimated Compliance Costs, Benefits, and Net Benefits for the
	2021)a'b'c	
Annualized Values for the 2021-2040 Timeframe for
Final Rule (millions of 2016$, discounted to
Health Benefits
Climate
Benefits0
Cost"
Net Benefits
3%
7%
3%
3%
7%
3%
7%
2021*	$230 and $1,900
2022	$230 and $2,000
2023	$230 and $2,000
2024*	$310 and $2,400
2025	$320 and $2,400
2026	$330 and $2,500
2027	$320 and $2,400
2028	$330 and $2,500
2029	$330 and $2,500
2030	$340 and $2,600
2031	$350 and $2,600
2032	$360 and $2,700
2033	$350 and $2,600
2034	$360 and $2,700
2035	$370 and $2,800
2036	$370 and $2,800
2037	$380 and $2,900
2038	$370 and $2,800
2039	$380 and $2,800
2040	$380 and $2,900
$200
$210
$210
$280
$290
$290
$300
$310
$320
$330
$340
$350
$360
$370
$380
$390
$400
$410
$430
$440
and $1,700
and $1,600
and $1,600
and $2,100
and $2,200
and $2,200
and $2,300
and $2,400
and $2,400
and $2,500
and $2,600
and $2,600
and $2,700
and $2,800
and $2,800
and $2,900
and $3,000
and $3,100
and $3,200
and $3,200
$1
$140
$290
$310
$330
$340
$350
$360
$360
$370
$350
$330
$310
$290
$270
$290
$300
$310
$330
$340
$5
$19
$19
$2
$2
$1
$0
$66
$65
$64
$64
$63
$18
$18
$18
$18
$18
$230 and $1,900
$350 and $2,100
$500 and $2,300
$620 and $2,700
$650 and $2,700
$670 and $2,800
$670 and $2,800
$620 and $2,800
$630 and $2,800
$650 and $2,900
$640 and $2,900
$630 and $3,000
$640 and $2,900
$630 and $3,000
$620 and $3,100
$640 and $3,100
$660 and $3,200
$670 and $3,100
$700 and $3,100
$710 and $3,200
$200
$330
$480
$590
$620
$630
$650
$600
$620
$640
$630
$620
$650
$640
$630
$660
$680
$710
$750
$770
and $1,700
and $1,700
and $1,900
and $2,400
and $2,500
and $2,500
and $2,700
and $2,700
and $2,700
and $2,800
and $2,900
and $2,900
and $3,000
and $3,100
and $3,100
and $3,200
and $3,300
and $3,400
and $3,500
and $3,500
PV
2021-2040
$4,800 and $37,000
$3,200 and $25,000
$4,400
$370
$260
$8,800 and $41,000
$7,300 and $29,000
EAV
2021 - 2040
$320 and $2,500
$300 and $2,400
$290
$25
$25
$590 and $2,800
$570 and $2,700
ES-28

-------
"Rows may not appear to add correctly due to rounding. The two benefits estimates are separated by the word "and" to signify that they are two separate
estimates. The estimates do not represent lower- and upper-bound estimates and should not be summed.
b The annualized present value of costs and benefits are calculated over a 20-year period from 2021 to 2040.
0 Climate benefits are based on changes (reductions) in CO2 emissions and are calculated using four different estimates of the social cost of carbon (SC-CO2)
(model average at 2.5 percent, 3 percent, and 5 percent discount rates; 95th percentile at 3 percent discount rate). For purposes of this table, we show the benefits
associated with the model average at a 3 percent discount rate. However, we emphasize the importance and value of considering the benefits calculated using all
four SC-CO2 estimates. As discussed in Chapter 5, a consideration of climate benefits calculated using discount rates below 3 percent, including 2 percent and
lower, are also warranted when discounting intergenerational impacts.
d The costs presented in this table are consistent with the costs presented in Chapter 4, Table 4-6. To estimate these annualized costs, EPA uses a conventional
and widely accepted approach that applies a capital recovery factor (CRF) multiplier to capital investments and adds that to the annual incremental operating
expenses. Annual costs were calculated using a 4.25% real discount rate consistent with the rate used in IPM's objective function for cost-minimization.
*Year in which air quality was simulated. Ozone air quality was simulated in 2021 and 2024 while the formation of PM2 5 was simulated only in 2024. Health
benefits for all other years were linearly extrapolated or interpolated from model-simulated air quality in these years. This method assumes that ozone and PM2 5
formation reaches a steady state beyond 2024 and may create increasing uncertainty in the benefits estimates the farther into the future estimates are extrapolated.
Benefits calculated as value of avoided: PM2 5-attributable deaths (quantified using a concentration-response relationship from the Di et al. 2017 study); Ozone-
attributable deaths (quantified using a concentration-response relationship from the Turner et al. 2017 study); and PM2 5 and ozone-related morbidity effects.
ES-29

-------
CHAPTER 1: INTRODUCTION AND BACKGROUND
Overview
EPA originally published the Cross-State Air Pollution Rule (CSAPR) on August 8, 2011,
to address interstate transport of ozone pollution under the 1997 ozone National Ambient Air
Quality Standards (NAAQS).1 On October 26, 2016, EPA published the CSAPR Update, which
finalized Federal Implementation Plans (FIPs) for 22 states that EPA found failed to submit a
complete good neighbor State Implementation Plan (SIP) (15 states)2 or for which EPA issued a
final rule disapproving their good neighbor SIP (7 states).3 The FIPs promulgated for these states
included new electric generating unit (EGU) oxides of nitrogen (NOx) ozone season emission
budgets to reduce interstate transport for the 2008 ozone NAAQS.4 These emission budgets took
effect in 2017 in order to assist downwind states with attainment of the 2008 ozone NAAQS by
the 2018 Moderate area attainment date. EPA acknowledged at the time that the FIPs
promulgated for 21 of the 22 states only partially addressed good neighbor obligations under the
2008 ozone NAAQS.5
This final action is taken in response to the United States Court of Appeals for the
District of Columbia Circuit's (D.C. Circuit) September 13, 2019 remand of the CSAPR
Update.6 The D.C. Circuit found that the CSAPR Update, which was a partial remedy, was
unlawful to the extent it allowed those states to continue their significant contributions to
downwind ozone problems beyond the statutory dates by which downwind states must
demonstrate their attainment of the air quality standards. This final rule will resolve 21 states'
outstanding interstate ozone transport obligations with respect to the 2008 ozone NAAQS.
1	CSAPR also addressed interstate transport of fine particulate matter (PM2 5) under the 1997 and 2006 PM2 5
NAAQS.
2	Alabama, Arkansas, Illinois, Iowa, Kansas, Maryland, Michigan, Mississippi, Missouri, New Jersey, Oklahoma,
Pennsylvania, Tennessee, Virginia, and West Virginia.
3	Indiana, Kentucky, Louisiana, New York, Ohio, Texas, and Wisconsin.
4	The 2008 ozone NAAQS is an 8-hour standard that was set at 75 parts per billion (ppb). See 73 FR 16436 (March
27, 2008).
5	In the CSAPR Update, EPA found that the finalized Tennessee emission budget fully addressed Tennessee's good
neighbor obligation with respect to the 2008 ozone NAAQS. As such, Tennessee is not considered in this rule, and
the number of states included is reduced from 22 to 21 states.
6	EPA is taking this action to address the remand of the CSAPR Update in Wisconsin v. EPA, 938 F.3d 303 (D.C.
Cir. 2019). The court remanded but did not vacate the CSAPR Update, finding that vacatur of the rule could cause
harm to public health and the environment or disrupt the trading program EPA had established and that the
obligations imposed by the rule may be appropriate and sustained on remand.
1-1

-------
This action, the Revised CSAPR Update rule, finds that for 9 of the 21 states with
remanded FIPs (Alabama, Arkansas, Iowa, Kansas, Mississippi, Missouri, Oklahoma, Texas, and
Wisconsin), their projected 2021 ozone season nitrogen oxides (NOx) emissions do not
significantly contribute to a continuing downwind nonattainment and/or maintenance problem;
therefore the CSAPR Update fully addresses their interstate ozone transport obligations with
respect to the 2008 ozone NAAQS. This action also finds that for the 12 remaining states
(Illinois, Indiana, Kentucky, Louisiana, Maryland, Michigan, New Jersey, New York, Ohio,
Pennsylvania, Virginia, and West Virginia), their projected 2021 ozone season NOx emissions
significantly contribute to downwind states' nonattainment and/or maintenance problems for the
2008 ozone NAAQS. For these 12 states, EPA amends their FIPs to revise the existing CSAPR
NOx Ozone Season Group 2 emissions budgets for EGUs and implement the revised budgets via
a new CSAPR NOx Ozone Season Group 3 Trading Program.7 EPA is implementing the revised
emission budgets starting with the 2021 ozone season (May 1 - September 30), as outlined in
section VII of the preamble.
These emission budgets represent the remaining EGU emissions after reducing those
amounts of each state's emissions that significantly contribute to nonattainment or interfere with
maintenance of the 2008 ozone NAAQS in downwind states, as required under Clean Air Act
(CAA) section 110(a)(2)(D)(i)(I). The allowance trading program is the remedy in the FIPs that
achieves the ozone season NOx emission reductions in the rule. The allowance trading program
essentially converts the EGU NOx emission budget for each of the 12 states into a limited
number of NOx allowances that, on a tonnage basis, equal the state's ozone season NOx
emission budget. EGUs covered by the FIPs can trade NOx ozone season allowances among
EGUs within their state and across state boundaries, subject to certain limits. The EGUs covered
by the FIPs and subject to the budget are fossil-fired EGUs with >25MW capacity. The 12 Group
3 states may not use allowances allocated under the CSAPR Update for compliance in 2021 and
later.8 Also, allowances allocated under the Revised CSAPR Update may not be used for
7	The CSAPR Update established a second NOx ozone season trading program for the 22 states determined to have
good neighbor obligations with respect to the 2008 ozone NAAQS - the CSAPR NOx Ozone Season Group 2
trading program.
8	EGUs can still use converted banked allowances from the CSAPR Update to comply with this rule.
1-2

-------
compliance in the 10 Group 2 states that remain subject to the budgets established in the CSAPR
Update.
Consistent with OMB Circular A-4 and EPA's Guidelines for Preparing Economic
Analyses (2010), this Regulatory Impact Analysis (RIA) presents the benefits and costs of the
rule from 2021 through 2040. The estimated benefits are those health and climate benefits
expected to arise from reduced air pollution and the estimated costs are the increased costs of
producing electricity and any state reporting requirements as a result of this rule. Unquantified
benefits and costs are described qualitatively. The RIA also provides (i) estimates of other
impacts of the rule including its effect on retail electricity prices and fuel production and (ii) an
assessment of how expected compliance with the rule will affect concentrations at nonattainment
and maintenance receptors. This chapter contains background information relevant to the rule
and an outline of the chapters of this RIA.
1.1 Background
Clean Air Act (CAA or the Act) section 110(a)(2)(D)(i)(I), which is also known as the
"good neighbor provision," requires states to prohibit emissions that will contribute significantly
to nonattainment or interfere with maintenance in any other state with respect to any primary or
secondary NAAQS. The statute vests states with the primary responsibility to address interstate
emission transport through the development of good neighbor State Implementation Plans (SIPs),
which are one component of larger SIP submittals typically required three years after EPA
promulgates a new or revised NAAQS. These larger SIPs are often referred to as "infrastructure"
SIPs or iSIPs. See CAA section 110(a)(1) and (2). EPA supports state efforts to submit good
neighbor SIPs for the 2008 ozone NAAQS and has shared information with states to facilitate
such SIP submittals. However, the CAA also requires EPA to fill a backstop role by issuing FIPs
where states fail to submit good neighbor SIPs or EPA disapproves a submitted good neighbor
SIP.
As described in the preamble for the rule, to reduce interstate emission transport under
the authority provided in CAA section 110(a)(2)(D)(i)(I), this rule further limits ozone season
(May 1 through September 30) NOx emissions from EGUs in 12 states using the same
framework used by EPA in developing the original CSAPR (the Interstate Transport
1-3

-------
Framework). The Interstate Transport Framework provides a 4-step process to address the
requirements of the good neighbor provision for ground-level ozone and fine particulate matter
(PM2.5) NAAQS: (1) identifying downwind receptors that are expected to have problems
attaining or maintaining the NAAQS; (2) determining which upwind states contribute to these
identified problems in amounts sufficient to "link" them to the downwind air quality problems
{i.e., here, a 1 percent contribution threshold); (3) for states linked to downwind air quality
problems, identifying upwind emissions that significantly contribute to downwind nonattainment
or interfere with downwind maintenance of the NAAQS; and (4) for states that are found to have
emissions that significantly contribute to nonattainment or interfere with maintenance of the
NAAQS downwind, implementing the necessary emissions reductions through enforceable
measures. Details on the methods and results of applying this process can be found in the
preamble for this rule.
1.1.1	Role of Executive Orders in the Regulatory Impact Analysis
Several statutes and executive orders apply to federal rulemakings. The analyses required
by these statutes, along with a brief discussion of several executive orders, are presented in
Chapter 6. Below we briefly discuss the requirements of Executive Orders 12866 and 13563 and
the guidelines of the Office of Management and Budget (OMB) Circular A-4 (U.S. OMB, 2003).
In accordance with Executive Orders 12866 and 13563 and the guidelines of OMB
Circular A-4, the RIA analyzes the benefits and costs associated with emissions reductions for
compliance with the Revised CSAPR Update rule. OMB Circular A-4 requires analysis of one
potential regulatory control alternative more stringent than the rule and one less stringent than
the rule. This RIA evaluates the benefits, costs, and certain impacts of a more and a less stringent
alternative to the selected alternative in this rule.
1.1.2	Alternatives Analyzed
EPA is amending FIPs for 12 states to revise the existing CSAPR NOx Ozone Season
Group 2 emissions budgets for EGUs and implement the revised budgets via a new CSAPR NOx
Ozone Season Group 3 Trading Program. Note that EGUs have flexibility in determining how
they will comply with the allowance trading program. EPA is implementing the revised emission
budgets starting with the 2021 ozone season.
1-4

-------
In response to OMB Circular A-4, this RIA analyzes the Revised CSAPR Update emission
budgets as well as a more and a less stringent alternative to the rule. The more and less stringent
alternatives differ from the Revised CSAPR Update in that they set different EGU NOx ozone
season emission budgets for the affected EGUs. The less-stringent scenario uses emission
budgets that were developed using a uniform control stringency represented by $500 per ton
(2016$). The more-stringent scenario uses emission budgets that were developed using a uniform
control stringency represented by $9,600 per ton (2016$). See Chapter 4, section 4.1 below,
section VII of the preamble, and the EGU NOx Mitigation Strategies Final Rule TSD, in the
docket for this rule9 for further details of these emission budgets.
1.1.3 The Needfor Air Quality or Emissions Regulation
OMB Circular A-4 indicates that one of the reasons a regulation may be issued is to
address a market failure. The major types of market failure include externalities, market power,
and inadequate or asymmetric information. Correcting market failures is one reason for
regulation; it is not the only reason. Other possible justifications include improving the function
of government, correcting distributional unfairness, or securing privacy or personal freedom.
Environmental problems are classic examples of externalities - uncompensated benefits
or costs imposed on another party as a result of one's actions. For example, the smoke from a
factory may adversely affect the health of local residents and soil the property in nearby
neighborhoods. Pollution emitted in one state may be transported across state lines and affect air
quality in a neighboring state. If bargaining were costless and all property rights were well
defined, people would eliminate externalities through bargaining without the need for
government regulation.
From an economics perspective, achieving emissions reductions (i.e., by establishing the
EGU NOx ozone-season emissions budgets in this rule) through a market-based mechanism is a
straightforward and cost-effective remedy to address an externality in which firms emit
pollutants, resulting in health and environmental problems without compensation for those
incurring the problems. Capping emissions through allowance allocations incentivizes those who
9 Docket ID No. EPA-HQ-OAR-2020-0272
1-5

-------
emit the pollutants to reduce their emissions, which lessens the impact on those who suffer the
health and environmental problems from higher levels of pollution.
1.2 Overview and Design of the RIA
1.2.1 Methodology for Identifying Needed Reductions
In order to apply the first and second steps of the CSAPR 4-step Interstate Transport
Framework to interstate transport for the 2008 ozone NAAQS, EPA first performed air quality
modeling coupled with ambient measurements in an interpolation technique to project ozone
concentrations at air quality monitoring sites in 2021. EPA evaluated 2021 projected ozone
concentrations at individual monitoring sites and considered current ozone monitoring data at
these sites to identify receptors that are anticipated to have problems attaining or maintaining the
2008 ozone NAAQS. In this analysis, downwind air quality problems are defined by receptors
that are projected to be unable to attain (i.e., nonattainment receptor) or maintain (i.e.,
maintenance receptor) the 2008 ozone NAAQS.
To apply the second step of the Interstate Transport Framework, EPA used air quality
modeling to quantify the contributions from upwind states to ozone concentrations in 2021 at
downwind receptors. Once quantified, EPA then evaluated these contributions relative to a
screening threshold of 1 percent of the NAAQS. States with contributions that equal or exceed 1
percent of the NAAQS are identified as warranting further analysis for significant contribution to
nonattainment or interference with maintenance.10 States with contributions below 1 percent of
the NAAQS are considered to not significantly contribute to nonattainment or interfere with
maintenance of the NAAQS in downwind states.
To apply the third step of the Interstate Transport Framework, EPA applied a multi-factor
test to evaluate cost, available emission reductions, and downwind air quality impacts to
determine the appropriate level of uniform NOx control stringency that addresses the impacts of
interstate transport on downwind nonattainment or maintenance receptors. EPA used this multi-
10 EPA assessed the magnitude of the maximum projected design value for 2021 at each receptor in relation to the
2008 ozone NAAQS. Where the value exceeds the NAAQS, EPA determined that receptor to be a maintenance
receptor for purposes of defining interference with maintenance. That is, monitoring sites with a maximum design
value that exceeds the NAAQS are projected to have a maintenance problem in 2021.
1-6

-------
factor assessment to gauge the extent to which emission reductions are needed, and to ensure any
required reductions do not result in over-control.
Using the multi-factor test, EPA identified a control strategy for EGUs at a stringency
level that maximizes cost-effective emission reductions.11 This control strategy reflects the
optimization of existing selective catalytic reduction (SCR) controls, optimization of existing
selective non-catalytic reduction (SNCR) controls, and installation of state-of-the-art NOx
combustion controls, at a representative cost of $1,800 per ton (2016$).12 It is at this control
stringency where incremental EGU NOx reduction potential and corresponding downwind ozone
air quality improvements are maximized relative to the alternative options analyzed. This
strategy maximizes the ratio of emission reductions to representative cost and the ratio of ozone
improvements to marginal cost. EPA finds that these cost-effective EGU NOx reductions will
make meaningful and timely improvements in downwind ozone air quality to address interstate
ozone transport for the 2008 ozone NAAQS, as discussed in Section VI.D. 1 of the preamble.
Further, this evaluation shows that emission budgets reflecting the $1,800 per ton cost threshold
do not over-control upwind states' emissions relative to either the downwind air quality
problems to which they are linked at step 1 or the 1 percent contribution threshold that triggers
further evaluation at step 2 of the 4-step Interstate Transport Framework for the 2008 ozone
NAAQS.
In applying the multi-factor test, EPA evaluated whether reductions resulting from the
emissions budgets for EGUs in 2021 and 2022 would resolve any downwind nonattainment or
maintenance problems. The assessment showed that the emission budgets reflecting $1,800 per
ton would change the status of one of the two nonattainment receptors (first shifting the
Stratford, Connecticut monitor to a maintenance-only receptor in 2021, then shifting that
receptor to attainment in 2022); however, no other nonattainment or maintenance problems
11	EPA's Guidelines for Preparing Economic Analysis states "[a] policy is cost-effective if it meets a given goal at
least cost, but cost-effectiveness does not encompass an evaluation of whether that goal has been set appropriately
to maximize social welfare. ... A policy is considered cost-effective when marginal abatement costs are equal across
all polluters. In other words, for any level of total abatement, each polluter has the same cost for their last unit
abated." (USEPA 2010, p 4-2). That is not the sense in which the term "cost-effective" is used in this paragraph. For
the sense of what this term means, and in particular what "maximize cost-effective reductions" means in the context
of this rulemaking, see Section VI.D. 1 of the preamble.
12	EGU NOx Mitigation Strategies Final Rule TSD, in the docket for this rule (Docket ID No. EPA-HQ-OAR-2020-
0272).
1-7

-------
would be resolved in 2021 or 2022. EPA's assessment shows that none of the 11 states are solely
linked to the Stratford receptor that is resolved at the $1,800 per ton level of control stringency in
2022. In addition, reductions resulting from the $1,800 per ton emission budgets would shift the
Houston receptor in Harris County, Texas from maintenance to attainment in 2023. These
emission reductions would also shift the last remaining nonattainment receptor (the Westport
receptor in Fairfield, Connecticut) to a maintenance-only receptor in 2024. No nonattainment or
maintenance receptors would remain after 2024.
1.2.2	States Covered by the Rule
This rule finds that the following 12 states require further ozone season NOx emission
reductions to address the good neighbor provision as to the 2008 ozone NAAQS: Illinois,
Indiana, Kentucky, Louisiana, Maryland, Michigan, New Jersey, New York, Ohio, Pennsylvania,
Virginia, and West Virginia.13 As such, EPA promulgates FIPs for these states that include new
EGU NOx ozone season emission budgets, with implementation of these emission budgets
beginning with the 2021 ozone season. EPA also adjusts states' emission budgets for each ozone
season thereafter to incentivize ongoing operation of identified emission controls to address
significant contribution, until such time that our air quality projections demonstrate anticipated
resolution of the downwind nonattainment and/or maintenance problems for the 2008 ozone
NAAQS.
1.2.3	Regulated Entities
The rule affects EGUs in these 12 states and regulates utilities (electric, natural gas, other
systems) classified as code 221112 by the North American Industry Classification System
(NAICS) and have a nameplate capacity of greater than 25 megawatts (MWe).
13 This action finds that for 9 of the 21 states with remanded FIPs (Alabama, Arkansas, Iowa, Kansas, Mississippi,
Missouri, Oklahoma, Texas, and Wisconsin), their projected 2021 ozone season NOx emissions do not significantly
contribute to a continuing downwind nonattainment and/or maintenance problem; therefore the CSAPR Update fully
addresses their interstate ozone transport obligations with respect to the 2008 ozone NAAQS. In addition, in the
CSAPR Update EPA found that the finalized Tennessee emission budget fully addressed Tennessee's good neighbor
obligation with respect to the 2008 ozone NAAQS, and Tennessee is also not considered in this rule. Allowances
allocated under the Revised CSAPR Update may not be used for compliance in these 10 Group 2 states that remain
subject to the budgets established in the CSAPR Update.
1-8

-------
1.2.4 Baseline and Analysis Years
As described in the preamble, EPA aligns implementation of this rule with relevant
attainment dates for the 2008 ozone NAAQS. EPA's final 2008 Ozone NAAQS SIP
Requirements Rule established the attainment deadline of July 20, 2021 for ozone nonattainment
areas currently designated as Serious, and EPA establishes emission budgets and implementation
of these emission budgets starting with the 2021 ozone season.
To develop and evaluate control strategies for addressing these obligations, it is important
to first establish a baseline projection of air quality in the analysis year of 2021, taking into
account currently on-the-books Federal regulations, substantial Federal regulatory proposals,
enforcement actions, state regulations, population, and where possible, economic growth.14
Establishing this baseline for the analysis then allows us to estimate the incremental costs and
benefits of the additional emissions reductions that will be achieved by the transport rule.
The baseline for this analysis does not assume states will adopt any emissions reduction
methods in and around the Air Quality Control Regions where the nonattainment and
maintenance receptors are located to reduce ozone other than those already taken into account. In
these areas that do not meet the NAAQS in the baseline that see decreased concentrations of
ozone, the states where these receptors are located may be able to avoid applying other measures
to assure NAAQS attainment. In this scenario, there would be benefits from avoided compliance
costs in these areas and the ozone and PM2.5 concentrations changes, and their associated health
and ecological benefits, would likely be lower relative to the projections in this RIA. However,
the baseline in this RIA respects that NOx reductions are required of upwind states in order to
improve air quality at the nonattainment and maintenance ozone receptors.
14 The technical support document (TSD) for the 2016vl emissions modeling platform titled Preparation of
Emissions Inventories for 2016vl North American Emissions Modeling Platform is included in the docket for this
rule. The TSD includes additional discussion on mobile source rules included in the baseline. The future year onroad
emission factors account for changes in activity data and the impact of on-the-books rules that are implemented into
MOVES2014b. These rules include the Light Duty Vehicle GHG Rule for Model-Year 2017-2025 and the Tier 3
Motor Vehicle Emission and Fuel Standards Rule. Local inspection and maintenance (I/M) and other onroad mobile
programs are included, such as California LEVIII, the National Low Emissions Vehicle (LEV) and Ozone Transport
Commission (OTC) LEV regulations, local fuel programs, and Stage II refueling control programs. Regulations
finalized after the year 2014 are not included, such as the Safer Affordable Fuel Efficient (SAFE) Vehicles Final
Rule for Model Years 2021-2026 and the Final Rule for Phase 2 Greenhouse Gas Emissions Standards and Fuel
Efficiency Standards for Medium- and Heavy-Duty Engines and Vehicles.
1-9

-------
The analysis in this RIA focuses on benefits, costs and certain impacts from 2021 through
2040. We focus on 2021 because it is by the 2021 ozone season, corresponding with the 2021
Serious area attainment date, that significant contribution from upwind states' must be
eliminated to the extent possible. In addition, impacts through 2025 are important because it is in
this period that additional NOx control technologies could potentially be installed while upwind
linkage to downwind receptors persists. EPA's analysis for the third step of the Interstate
Transport Framework indicates that by 2023 the remaining ozone receptors in the two downwind
states (Connecticut and Texas) are expected to shift from nonattainment or maintenance status to
meeting the NAAQS with application of certain EGU controls beginning in 2021, except for one
receptor in Westport, Connecticut.15 This receptor is estimated to shift from nonattainment status
to meeting the NAAQS in 2025 with the application of additional EGU controls. Presenting
benefits, costs and certain impacts in 2025 reflects the time needed to make these retrofits on a
regional scale and reflects full implementation of the policy. Additional benefits and costs are
expected to occur after 2025 as EGUs subject to this rule continue to comply with the tighter
allowance budget, which is below their baseline emissions.
1.2.5 Emissions Controls, Emissions, and Cost Analysis Approach
EPA estimated the control strategies and compliance costs of the rule using the Integrated
Planning Model (IPM) as well as certain costs that are estimated outside the model but use IPM
inputs for their estimation. These cost estimates reflect costs incurred by the power sector and
include (but are not limited to) the costs of purchasing, installing, and operating NOx control
technology, changes in fuel costs, and changes in the generation mix.16 A description of the
methodologies used to estimate the costs and economic impacts to the power sector is contained
in Chapter 4 of this RIA. This analysis also provides estimates of NOx emissions changes during
15	This RIA also provides an assessment of how expected compliance with the rule will affect concentrations at
nonattainment and maintenance receptors. See Chapter 4 for additional details.
16	Under the rule and the more stringent alternative, 10 units are projected to install state-of-the-art combustion
controls; under the less stringent alternative, no units are projected to install state-of-the-art combustion controls.
Under the rule, the less stringent alternative, and the more stringent alternative, no units are projected to install new
SCRs. Under the proposed rule, units were exogenously forced to install SCR controls in IPM. In the modeling for
the final rule, the choice to install SCR controls was endogenous to the model, and no incremental SCR installations
occurred, with the model relying on greater levels of generation shifting instead.
1-10

-------
ozone season and year-round, as well as emissions changes in carbon dioxide (CO2) due to
changes in power sector operation.
1.2.6 Benefits Analysis Approach
Implementing the Revised CSAPR Update rule is expected to reduce emissions of NOx
and provide ozone reductions, as well as consequent reductions in PM2.5 concentrations and CO2
emissions. In the proposed Revised CSAPR Update RIA EPA committed to updating its
approach for quantifying the benefits of changes in PM2.5 and ozone in this final Revised CSAPR
Update RIA. The updated approach incorporates evidence reported in the recently completed
PM2.5 and Ozone Integrated Science Assessments and accounts for recommendations from the
Science Advisory Board. Detailed descriptions of these updates are available in the Technical
Support Document (TSD) for the Final Revised Cross-State Air Pollution Rule for the 2008
Ozone NAAQS Update titled EstimatingPM2.5- and Ozone-Attributable Health Benefits. For
more details on these updates, also see Chapter 5. EPA estimated the climate benefits, and a
description of the methodologies used to estimate the climate benefits is also contained in
Chapter 5.
1.3 Organization of the Regulatory Impact Analysis
This RIA is organized into the following remaining chapters:
•	Chapter 2: Electric Power Sector Profile. This chapter describes the electric power sector
in detail.
•	Chapter 3: Emissions and Air Quality Modeling Impacts. The data, tools, and
methodology used for the air quality modeling are described in this chapter, as well as the
post-processing techniques used to produce a number of air quality metrics for input into
the analysis of benefits.
•	Chapter 4: Cost, Emissions, and Energy Impacts. The chapter summarizes the data
sources and methodology used to estimate the costs and other impacts incurred by the
power sector.
•	Chapter 5: Benefits. The chapter presents the health-related benefits of the ozone and
PM2.5-related air quality improvements as well as the climate benefits.
•	Chapter 6: Statutory and Executive Order Impact Analyses. The chapter summarizes the
Statutory and Executive Order impact analyses.
•	Chapter 7: Comparison of Benefits and Costs. The chapter compares estimates of the
total benefits with total costs and summarizes the net benefits of the three alternative
regulatory control scenarios analyzed.
1-11

-------
CHAPTER 2: ELECTRIC POWER SECTOR PROFILE
Overview
This chapter discusses important aspects of the power sector that relate to the Revised
CSAPR Update with respect to the interstate transport of emissions of nitrogen oxides (NOx)
that contribute significantly to nonattainment or interfere with maintenance of the 2008 ozone
NAAQS in downwind states. This chapter describes types of existing power-sector sources
affected by the regulation1 and provides background on the power sector and electricity
generating units (EGUs). In addition, this chapter provides some historical background on recent
trends in the power sector, as well as about existing EPA regulation of the power sector.
2.1	Background
In the past decade there have been significant structural changes in both the mix of
generating capacity and in the share of electricity generation supplied by different types of
generation. These changes are the result of multiple factors in the power sector, including normal
replacements of older generating units with new units, changes in the electricity intensity of the
U.S. economy, growth and regional changes in the U.S. population, technological improvements
in electricity generation from both existing and new units, changes in the prices and availability
of different fuels, and substantial growth in electricity generation by renewable and
unconventional methods. Many of these trends will continue to contribute to the evolution of the
power sector. The evolving economics of the power sector, specifically the increased natural gas
supply and subsequent relatively low natural gas prices, have resulted in more natural gas being
used as base load energy in addition to supplying electricity during peak load. This chapter
presents data on the evolution of the power sector from 2014 through 2018. Projections of future
power sector behavior and the impact of this rule are discussed in more detail in Chapter 4 of this
RIA.
2.2	Power Sector Overview
The production and delivery of electricity to customers consists of three distinct segments:
generation, transmission, and distribution.
1 Only coal-fired EGUs will be directly affected (i.e., have to reduce NOx emissions) by this rule.
2-1

-------
2.2.1 Generation
Electricity generation is the first process in the delivery of electricity to consumers. There
are two important aspects of electricity generation: capacity and net generation. Generating
Capacity refers to the maximum amount of production an EGU is capable of producing in a
typical hour, typically measured in megawatts (MW) for individual units, or gigawatts (1 GW =
1,000 MW) for multiple EGUs. Electricity Generation refers to the amount of electricity actually
produced by an EGU over some period of time, measured in kilowatt-hours (kWh) or gigawatt-
hours (GWh = 1 million kWh). Net Generation is the amount of electricity that is available to the
grid from the EGU (i.e., excluding the amount of electricity generated but used within the
generating station for operations). Electricity generation is most often reported as the total annual
generation (or some other period, such as seasonal). In addition to producing electricity for sale
to the grid, EGUs perform other services important to reliable electricity supply, such as
providing backup generating capacity in the event of unexpected changes in demand or
unexpected changes in the availability of other generators. Other important services provided by
generators include facilitating the regulation of the voltage of supplied generation.
Individual EGUs are not used to generate electricity 100 percent of the time. Individual
EGUs are periodically not needed to meet the regular daily and seasonal fluctuations of
electricity demand. Furthermore, EGUs relying on renewable resources such as wind, sunlight
and surface water to generate electricity are routinely constrained by the availability of adequate
wind, sunlight, or water at different times of the day and season. Units are also unavailable
during routine and unanticipated outages for maintenance. These factors result in the mix of
generating capacity types available (e.g., the share of capacity of each type of EGU) being
substantially different than the mix of the share of total electricity produced by each type of EGU
in a given season or year.
Most of the existing capacity generates electricity by creating heat to create high pressure
steam that is released to rotate turbines which, in turn, create electricity. Natural gas combined
cycle (NGCC) units have two generating components operating from a single source of heat. The
first cycle is a gas-fired turbine, which generates electricity directly from the heat of burning
natural gas. The second cycle reuses the waste heat from the first cycle to generate steam, which
is then used to generate electricity from a steam turbine. Other EGUs generate electricity by
2-2

-------
using water or wind to rotate turbines, and a variety of other methods including direct
photovoltaic generation also make up a small, but growing, share of the overall electricity
supply. The generating capacity includes fossil-fuel-fired units, nuclear units, and hydroelectric
and other renewable sources (see Table 2-1). Table 2-1 also shows the comparison between the
generating capacity in 2014 and 2018.
In 2018 the power sector consisted of over 22,000 generating units with a total capacity2 of
1,095 GW, an increase of 26 GW (or 2 percent) from the capacity in 2014 (1,068 GW). The 26
GW increase consisted primarily of natural gas fired EGUs (38 GW), and wind (30 GW) and
solar generators (22 GW), and the retirement/re-rating of 56 GW of coal capacity. Substantially
smaller net increases and decreases in other types of generating units also occurred.
Table 2-1. Total Net Summer Electricity Generating Capacity by Energy Source, 2014
and 2018

2014
2018
Change Between '14 and '18
Energy Source
Net
Summer
Capacity
(MW)
% Total
Capacity
Net
Summer
Capacity
(MW)
% Total
Capacity
%
Increase
Capacity
Change
(MW)
%of
Total
Capacity
Increase
Coal
299,094
28%
242,786
22%
-19%
-56,309
-214%
Natural Gas
432,150
40%
470,237
43%
9%
38,087
145%
Nuclear
98,569
9%
99,433
9%
0.9%
864
3.3%
Hydro
102,162
9.56%
102,702
9.38%
0.5%
540
2.1%
Petroleum
41,135
3.85%
32,218
2.94%
-22%
-8,917
-34%
Wind
64,232
6.01%
94,418
8.62%
47%
30,186
115%
Solar
10,323
0.97%
31,878
2.91%
209%
21,555
82%
Other Renewable
16,049
2%
16,178
1%
1%
129
0%
Misc
4,707
0.44%
4,891
0.45%
4%
184
1%
Total
1,068,422
100%
1,094,740
100%
2%
26,318
100%
Note: This table presents generation capacity. Actual net generation is presented in Table 2-2.
Source: EIA. Electric Power Annual 2014 and 2018, Table 4.3
The 2 percent increase in generating capacity is the net impact of newly built generating
units, retirements of generating units, and a variety of increases and decreases to the nameplate
2 This includes generating capacity at EGUs primarily operated to supply electricity to the grid and combined heat
and power facilities classified as Independent Power Producers (IPP) and excludes generating capacity at
commercial and industrial facilities that does not operate primarily as an EGU. Natural Gas information in this
chapter (unless otherwise stated) reflects data for all generating units using natural gas as the primary fossil heat
source. This includes Combined Cycle Combustion Turbine, Gas Turbine, steam, and miscellaneous (< 1 percent).
2-3

-------
capacity of individual existing units due to changes in operating equipment, changes in emission
controls, etc. During the period 2014 to 2018, a total of 98 GW of new generating capacity was
built and brought online, and 74 GW of existing units were retired. The net effect of the re-rating
of existing units reduced the total capacity by 9.4 GW. The overall net change in capacity was an
increase of 26 GW, as shown in Table 2-1.
The newly built generating capacity was primarily natural gas (44 GW), which was
partially offset by gas retirements (24 GW). Wind capacity was the second largest type of new
builds (30 GW), augmented by solar (21 GW). The largest decline was from coal retirements and
re-rating, which amounted to 56 GW over this period. The overall mix of newly built and retired
capacity, along with the net effect, is shown on Figure 2-1. The data for Figure 2-1 is from Form
EIA-860. Figure 2-1 does not show wind and solar retirements of 568 MW.
130,000
110,000
90,000
70,000
50,000
30,000
10,000
(10,000)
(30,000)
(50,000)
I Other
I Coal
I Wind & Solar
Gas
New Build
Retirement
N
et Change
Figure 2-1. National New Build and Retired Capacity (MW) by Fuel Type, 2014-2018
The information in Table 2-1 and Figure 2-1 present information about the generating
capacity in the entire U.S. The CSAPR Update, however, directly affected EGUs in 23 eastern
states (i.e., the CSAPR 2008 Ozone Region). The share of generating capacity from each major
type of generation differs between the CSAPR 2008 Ozone Region and the rest of the U.S. (non-
2-4

-------
region). Figure 2-2 shows the mix of generating capacity for each region. In 2018, the overall
capacity in the CSAPR 2008 Ozone Region is 59 percent of the national total, reflecting the
larger total population in the region. The mix of capacity is noticeably different in the two
regions. In the CSAPR 2008 Ozone Region in 2014, coal makes up a significantly larger share of
total capacity (26 percent) than it does in the rest of the country (17 percent). The share of
natural gas in the CSAPR 2008 Ozone Region is 45 percent as compared to 40 percent in the rest
of the country. The difference in the share of coal's capacity is primarily balanced by relatively
more hydro, wind, and solar capacity in the rest of country compared to the CSAPR 2008 Ozone
Region.
700,000
600,000
500,000
400,000
300,000
200,000
100,000
0
I Other
I Wind & Solar
I Hydro
I Nuclear
Gas
I Coal
In Region
Non-Region
Figure 2-2. Regional Differences in Generating Capacity (MW), 2018
Source: Form EIA-860. Note: "Other" includes petroleum, geothermal, other renewable, waste materials and
miscellaneous.
In 2018, electric generating sources produced a net 4,204 TWh to meet national electricity
demand, a 2 percent increase from 2014. As presented in Table 2-2, 62 percent of electricity in
2018 was produced through the combustion of fossil fuels, primarily coal and natural gas, with
natural gas accounting for the largest single share. Although the share of the total generation
from fossil fuels in 2018 (62 percent) was only modestly smaller than the total fossil share in
2014 (66 percent), the mix of fossil fuel generation changed substantially during that period.
Coal generation declined by 28 percent and petroleum generation by 17 percent, while natural
gas generation increased by 30 percent. This reflects both the increase in natural gas capacity
2-5

-------
during that period as well as an increase in the utilization of new and existing gas EGUs during
that period. Wind and solar generation also grew from 5 percent of the mix in 2014 to 8 percent
in 2018.
Table 2-2. Net Generation in 2014 and 2C
)18 (Trillion kWh
= TWh)



2014
2018
Change Between '14
and '18

Net
Generation Fuel Source
(TWh) Share
Net
Generation Fuel Source
(TWh) Share
Net
Generati
on
Change
(TWh)
% Change
in Net
Generation
Coal
1,582 39%
1,146
27%
-436
-440%
Natural Gas
1,127 27%
1,469
35%
342
345%
Nuclear
797 19%
807
19%
10
10%
Hydro
253 6%
287
7%
33
34%
Petroleum
30 1%
25
1%
-5
-5%
Wind
182 4%
273
6%
91
92%
Solar
18 0%
64
2%
46
47%
Other Renewable
91 2%
107
3%
16
16%
Misc
25 1%
26
1%
1
1%
Total
4,105 100%
4,204
100%
99
100%
Source: EIA 2014 and 2018 Electric Power Annual, Tables 3.2 and 3.3.
Coal-fired and nuclear generating units have historically supplied "base load" electricity,
the portion of electricity loads that are continually present and typically operate throughout all
hours of the year. The coal units meet the part of demand that is relatively constant. Although
much of the coal fleet operates as base load, there can be notable differences across various
facilities (see Table 2-3). For example, coal-fired units less than 100 megawatts (MW) in size
compose 18 percent of the total number of coal-fired units, but only 2 percent of total coal-fired
capacity. Gas-fired generation is better able to vary output and is the primary option used to meet
the variable portion of the electricity load and has historically supplied "peak" and
"intermediate" power, when there is increased demand for electricity (for example, when
businesses operate throughout the day or when people return home from work and run appliances
and heating/air-conditioning), versus late at night or very early in the morning, when demand for
electricity is reduced.
2-6

-------
Table 2-3 also shows comparable data for the capacity and age distribution of natural gas
units. Compared with the fleet of coal EGUs, the natural gas fleet of EGUs is generally smaller
and newer. While 66 percent of the coal EGU fleet capacity is over 500 MW per unit, 82 percent
of the gas fleet is between 50 and 500 MW per unit. Many of the largest gas units are gas-fired
steam-generating EGUs.
Table 2-3. Coal and Natural Gas Generating Units, by Size, Age, Capacity, and Average
Heat Rate in 2018




Avg. Net
Total Net


Unit Size



Summer
Summer

Avg. Heat
Grouping

% of All

Capacity
Capacity
% Total
Rate
(MW)
No. Units
Units
Avg. Age
(MW)
(MW)
Capacity
(Btu/kWh)
COAL
0-24
37
7%
50
12
427
0%
11,948
25-49
39
7%
34
36
1,404
1%
12,386
50-99
26
5%
39
76
1,987
1%
12,027
100 - 149
39
7%
48
122
4,757
2%
11,223
150 - 249
73
13%
50
192
14,040
7%
10,882
250 - 499
142
25%
41
364
51,748
24%
10,659
500 - 749
143
26%
39
608
87,005
40%
10,310
750 - 999
49
9%
35
827
40,521
19%
10,057
1000 - 1500
11
2%
41
1,257
13,831
6%
9,802
Total Coal
559
100%
41
386
215,720
100%
10,838
NATURAL GAS
0-24
3,910
51%
32
5
20,540
4%
14,015
25-49
931
12%
26
41
37,792
8%
11,999
50-99
1,032
14%
26
71
73,129
15%
12,315
100 - 149
418
5%
22
127
52,927
11%
9,442
150 - 249
1,018
13%
16
179
181,772
38%
8,192
250 - 499
247
3%
22
332
82,114
17%
8,296
500 - 749
38
0%
39
577
21,910
5%
10,583
750 - 1000
9
0%
44
834
7,510
2%
11,625
Total Gas
7,603
100%
28
63
477,693
100%
12,301
Source: National Electric Energy Data System (NEEDS) v.6
Note: The average heat rate reported is the mean of the heat rate of the units in each size category (as opposed to a
generation-weighted or capacity-weighted average heat rate.) A lower heat rate indicates a higher level of fuel
efficiency. Table is limited to coal-steam units in operation in 2018 or earlier and excludes those units in NEEDS
with planned retirements in 2019 or 2020.
In In terms of the age of the generating units, almost 50 percent of the total coal generating
capacity has been in service for more than 40 years, while nearly 50 percent of the natural gas
capacity has been in service less than 15 years. Figure 2-3 presents the cumulative age
distributions of the coal and gas fleets, highlighting the pronounced differences in the ages of the
fleets of these two types of fossil-fuel generating capacity. Figure 2-3 also includes the
2-7

-------
distribution of generation
100%
90%
80%
£
o
70%
+-¦
10
i—
01
c
Ol
ej
03
>¦
60%
50%
20%
10%
0%
0
10
20
30
40
50
60
70
Age of EGU (Years)
Coal Cap — — —Coal Gen	Gas Cap — — —Gas Gen
Figure 2-3. Cumulative Distribution in 2018 of Coal and Natural Gas Electricity Capacity
and Generation, by Age
Source: eGRID 2018 (March 2020 release from EPA eGRID website). Figure presents data from generators that
came online between 1949 and 2018 (inclusive); a 70-year period. Full eGrid data includes generators that came
online as far back as 1915. Full data from 1915 onward is used in calculating cumulative distributions; figure
truncation at 70 years is merely to improve visibility of diagram. Figure is limited to coal-steam units in NEEDS v6
in operation in 2018 or earlier.
The locations of existing fossil units in EPA's National Electric Energy Data System
(NEEDS) v.6 are shown in Figure 2-4.
2-8

-------
Figure 2-4. Fossil Fuel-Fired Electricity Generating Facilities, by Size
Source: National Electric Energy Data System (NEEDS) v.6
Note: This map displays fossil capacity at facilities in the NEEDS v.6 IPM frame. NEEDS v.6 reflects generating
capacity expected to be on-line at the end of 2021. This includes planned new builds already under construction and
planned retirements. In areas with a dense concentration of facilities, some facilities may be obscured.
2.2.2 Transmission
Transmission is the term used to describe the bulk transfer of electricity over a network of
high voltage lines, from electric generators to substations where power is stepped down for local
distribution. In the U.S. and Canada, there are three separate interconnected networks of high
voltage transmission lines,3 each operating synchronously. Within each of these transmission
networks, there are multiple areas where the operation of power plants is monitored and
controlled by regional organizations to ensure that electricity generation and load are kept in
balance. In some areas, the operati on of the transmission system is under the control of a single
3 These three network interconnections are the Western Interconnection, comprising the western parts of both the US
and Canada (approximately the area to the west of the Rocky Mountains), the Eastern Interconnection, comprising
the eastern parts of both the US and Canada (except those part of eastern Canada that are in the Quebec
Interconnection), and the Texas Interconnection (which encompasses the portion of the Texas electricity system
commonly known as the Electric Reliability Council of Texas (ERCOT)). See map of all NERC interconnections at
https://www.nerc.com/AboutNERC/keyplayers/PublisliingImages/NERC%20Interconnections.pdf.
2-9

-------
regional operator;4 in others, individual utilities5 coordinate the operations of their generation,
transmission, and distribution systems to balance the system across their respective service
territories.
2.2.3 Distribution
Distribution of electricity involves networks of lower voltage lines and substations that
take the higher voltage power from the transmission system and step it down to lower voltage
levels to match the needs of customers. The transmission and distribution system is the classic
example of a natural monopoly, in part because it is not practical to have more than one set of
lines running from the electricity generating sources to substations or from substations to
residences and businesses.
Over the last few decades, several jurisdictions in the United States began restructuring the
power industry to separate transmission and distribution from generation, ownership, and
operation. Historically, vertically integrated utilities established much of the existing
transmission infrastructure. However, as parts of the country have restructured the industry,
transmission infrastructure has also been developed by transmission utilities, electric
cooperatives, and merchant transmission companies, among others. Distribution, also historically
developed by vertically integrated utilities, is now often managed by a number of utilities that
purchase and sell electricity, but do not generate it. As discussed below, electricity restructuring
has focused primarily on efforts to reorganize the industry to encourage competition in the
generation segment of the industry, including ensuring open access of generation to the
transmission and distribution services needed to deliver power to consumers. In many states,
such efforts have also included separating generation assets from transmission and distribution
assets to form distinct economic entities. Transmission and distribution remain price-regulated
throughout the country based on the cost of service.
2.3 Sales, Expenses, and Prices
These electric generating sources provide electricity for ultimate commercial, industrial
and residential customers. Each of the three major ultimate categories consume roughly a quarter
4	For example, PMJ Interconnection, LLC, Western Area Power Administration (which comprises 4 sub-regions).
5	For example, Los Angeles Department of Power and Water, Florida Power and Light.
2-10

-------
to a third of the total electricity produced6 (see Table 2-4). Some of these uses are highly
variable, such as heating and air conditioning in residential and commercial buildings, while
others are relatively constant, such as industrial processes that operate 24 hours a day. The
distribution between the end use categories changed very little between 2014 and 2018.
Table 2-4. Total U.S. Electric Power Industry Retail Sales, 2014 and 2018 (billion kWh)

2014
2018


Sales/Direct

Sales/Direct



Use (Billion
Share of Total
Use (Billion
Share of Total


kWh)
End Use
kWh)
End Use

Residential
1,407
36%
1,469
37%
Sales
Commercial
1,352
35%
1,382
35%
Industrial
998
26%
1,001
25%

Transportation
8
0%
8
0%
Total
3,765
96%
3,859
96%
Direct Use
139
4%
144
4%
Total End Use
3,903
100%
4,003
100%
Source: Table 2.2, EIA Electric Power Annual, 2014 and 2018
Notes: Retail sales are not equal to net generation (Table 2-2) because net generation includes net imported
electricity and loss of electricity that occurs through transmission and distribution, along with data collection frame
differences and non-sampling error. Direct Use represents commercial and industrial facility use of onsite net
electricity generation; electricity sales or transfers to adjacent or co-located facilities; and barter transactions.
2.3.1 Electricity Prices
Electricity prices vary substantially across the United States, differing both between the
ultimate customer categories and by state and region of the country. Electricity prices are
typically highest for residential and commercial customers because of the relatively high costs of
distributing electricity to individual homes and commercial establishments. The higher prices for
residential and commercial customers are the result both of the necessary extensive distribution
network reaching to virtually every part of the country and every building, and also the fact that
generating stations are increasingly located relatively far from population centers (which
increases transmission costs). Industrial customers generally pay the lowest average prices,
reflecting both their proximity to generating stations and the fact that industrial customers
6 Transportation (primarily urban and regional electrical trains) is a fourth ultimate customer category which
accounts for less than one percent of electricity consumption.
2-11

-------
receive electricity at higher voltages (which makes transmission more efficient and less
expensive). Industrial customers frequently pay variable prices for electricity, varying by the
season and time of day, while residential and commercial prices historically have been less
variable. Overall industrial customer prices are usually considerably closer to the wholesale
marginal cost of generating electricity than residential and commercial prices.
On a state-by-state basis, all retail electricity prices vary considerably. In 2018, the national
average retail electricity price (all sectors) was 10.53 cents/KWh, with a range from 7.71 cents
(Louisiana) to 29.18 (Hawaii).7
Average national retail electricity prices decreased between 2014 and 2018 by 5 percent
in real terms (2018$).8 The amount of decrease differed for the three major end use categories
(residential, commercial and industrial). National average industrial prices decreased the most (9
percent), and residential prices decreased the least (4 percent). The real year prices for 2014
through 2019 are shown in Figure 2-5.
14
i/v
CO
o
12
10
c
a>
Q_
>-
2014
2015
• Residential
2016
• Commercial
2017	2018
Industrial —
2019
• Total
Figure 2-5. Real National Average Electricity Prices (including taxes) for Three Major
End-Use Categories
7	EIA State Electricity Profiles with Data for 2019 (http://www.eia.gov/electricity/state/)
8	All prices in this section are estimated as real 2018 prices adjusted using the GDP implicit price deflator unless
otherwise indicated.
2-12

-------
Source: EIA Monthly Energy Review (Feb 2021), Table 9.8.
Most of these electricity price decreases occurred between 2014 and 2015, when nominal
residential electricity prices followed inflation trends, while nominal commercial and industrial
electricity prices declined. The years 2016 and 2017 saw an increase in nominal commercial and
industrial electricity prices, while 2018 saw flattening of this growth. The increase in nominal
electricity prices for the major end use categories, as well as increases in the GDP price and CPI-
U indices for comparison, are shown in Figure 2-6.
7018
Residential
Commercial
Industrial
Figure 2-6. Relative Increases in Nominal National Average Electricity Prices for Major
End-Use Categories (including taxes), With Inflation Indices
Source: EIA Monthly Energy Review (Feb 2021), Table 9.8.
For a longer-term perspective, Figure 2-7 shows real9 (2018$) electricity prices for the
three major customer categories since 1960, and Figure 2-8 shows the relative change in real
electricity prices relative to the prices since 1960. As can be seen in the figures, the price for
industrial customers has always been lower than for either residential or commercial customers,
but the industrial price has been more volatile. While the industrial real price of electricity in
9 All prices in this section are estimated as real 2018 prices adjusted using the GDP implicit price deflator unless
otherwise indicated.
2-13

-------
2018 was relatively unchanged from 1960 (5 percent lower), residential and commercial real
prices are 25 percent and 33 percent lower respectively than in 1960.
Residential
Commercial
Industrial
Figure 2-7. Real National Average Electricity Prices for Three Major End-Use Categories
(including taxes), 1960-2019 (2018$)
Source: EIA Monthly Energy Review, Feb 2021, Table 9.8
2000 /205C X*
• Commercial	Industrial — —Total
60%
50%
o 40%
in
3 30%
O)
£ 20%
(?)
(v 10%
00
ro 0%
_c
^ -10%
S -20%
I—
-------
2.3.2 Prices of Fossil Fuels Usedfor Generating Electricity
Another important factor in the changes in electricity prices are the changes in delivered
fuel prices10 for the three major fossil fuels used in electricity generation: coal, natural gas and
petroleum products. Relative to real prices in 2014, the national average real price (in 2018$) of
coal delivered to EGUs in 2018 had decreased by 18 percent, while the real price of natural gas
decreased by 33 percent. The real price of delivered petroleum products also decreased by 22
percent, but with petroleum products declining as an EGU fuel (in 2018 petroleum products
generated 1 percent of electricity) the higher delivered oil prices had little overall impact in the
electricity market. The combined real delivered price of all fossil fuels in 2014 decreased by 20
percent over 2014 prices. Figure 2-9 shows the relative changes in real price of all 3 fossil fuels
between 2014 and 2019.
-10%
-20%
5P -30%
-40%
-50%
2019
-60%
¦ Coal
Oil
• Gas
> Ave rage
Figure 2-9. Relative Real Prices of Fossil Fuels for Electricity Generation; Change in
National Average Real Price per MMBtu Delivered to EGU
Source: EIA Monthly Energy Review, Feb 2021, Table 9.9.
111 Fuel prices in this section are all presented in terms of price per MMBtu to make the prices comparable.
2-15
-------
2.3.3 Changes in Electricity Intensity of the U.S. Economy from 2014 to 2018
An important aspect of the changes in electricity generation (i.e., electricity demand)
between 2014 and 2018 is that while total net generation increased by 2 percent over that period,
the demand growth for generation was lower than both the population growth (3 percent) and
real GDP growth (10 percent). Figure 2-10 shows the growth of electricity generation,
population and real GDP during this period.
Population
Generation
Figure 2-10. Relative Growth of Electricity Generation, Population and Real GDP Since
2014
Sources: Generation: U.S. EIA Monthly Energy Review, May 2020. Table 7.2a Electricity Net Generation: Total
(All Sectors). Population: U.S. Census. Real GDP: 2019 Economic Report of the President, Table B-3.
Because demand for electricity generation grew more slowly than both the population
and GDP, the relative electric intensity of the U.S. economy improved (i.e., less electricity used
per person and per real dollar of output) during 2014 to 2018. On a per capita basis, real GDP per
capita grew by 7 percent between 2014 and 2018. At the same time electricity generation per
capita decreased by 1 percent. The combined effect of these two changes improved the overall
electricity generation efficiency in the U.S. market economy. Electricity generation per dollar of
real GDP decreased 7 percent. These relative changes are shown in Figure 2-11.
2-16
-------
8%
6%
H
° 4%
-------
establishing cost-based rates for various customer classes. Deregulation and market restructuring
in the power sector involved the divestiture of generation from utilities, the formation of
organized wholesale spot energy markets with economic mechanisms for the rationing of scarce
transmission resources during periods of peak demand, the introduction of retail choice
programs, and the establishment of new forms of market oversight and coordination.
The pace of restructuring in the electric power industry slowed significantly in response to
market volatility in California and financial turmoil associated with bankruptcy filings of key
energy companies. Currently, there are 15 states plus the District of Columbia where price
deregulation of generation (restructuring) has occurred ("Active" in Figure 2-12). Power sector
restructuring is more or less at a standstill; by 2010 there were no active proposals under review
by the Federal Energy Regulatory Commission (FERC) for actions aimed at wider restructuring,
and no additional states have begun retail deregulation activity since that time.
Electricity retail choice states, 2010
States with electricity retail choice programs
States without electricity retail choice programs
Figure 2-12. Status of State Electricity Industry Restructuring Activities
Source: EIA 2010. "Status of Electricity Restructuring by State." Available online at:
.
One major effect of the restructuring and deregulation of the power sector was a
significant change in type of ownership of electricity generating units in the states that
deregulated prices. Throughout most of the 20th century electricity was supplied by vertically
integrated regulated utilities. The traditional integrated utilities provided generation, transmission
2-18
-------
and distribution in their designated areas, and prices were set by cost of service regulations set by
state government agencies (e.g., Public Utility Commissions). Deregulation and restructuring
resulted in unbundling of the vertical integration structure. Transmission and distribution
continued to operate as monopolies with cost of service regulation, while generation shifted to a
mix of ownership affiliates of traditional utility ownership and some generation owned and
operated by competitive companies known as Independent Power Producers (IPPs). The
resulting generating sector differed by state or region, as the power sector adapted to the
restructuring and deregulation requirements in each state.
By the year 2000, the major impacts of adapting to changes brought about by
deregulation and restructuring during the 1990s were nearing completion. In 2014, traditional
utilities owned 61 percent of U.S. generating capacity (MW) while IPPs11 owned 39 percent of
U.S. generating capacity, respectively. The mix of electricity generated (MWh) was more
heavily weighted towards the utilities, with a distribution in 2014 of 61 percent, and 39 percent
for IPPs.
In 2018, the share of capacity (59 percent utility, 41 percent IPPs) and generation (58
percent utility, 42 percent IPP) has remained relatively stable relative to 2014 levels.
The mix of capacity and generation for each of the ownership types is shown in Figures
2-13 (capacity) and 2-14 (generation). The capacity and generation data for commercial and
industrial owners are not shown on these figures due to the small magnitude of those ownership
11 IPP data presented in this section include both combined and non-combined heat and power plants.
2-19
-------

700

600

500
5

s.
400
>



re
300
Q.

re

<_>
200

100

0
Capacity Mix, 2014 & 2018	Generation Mix, 2014 & 2018
2,500
c I	2'000 M M
Solar	---	¦ ¦	Solar
I |
¦ot"er a i,5oo I I _
.Wind	J
¦ Hydro	2 1,000
Nuclear $	¦ Nuclear
500
Gas	H ¦ ¦ Gas
¦ Coal	q H H H H ¦ Coal
2014 2018 2014 2018	2014 2018 2014 2018
Utility	IPP	Utility	IPP
I ¦
Figures 2-13. and 2-14. Capacity and Generation Mix by Ownership Type, 2014 & 2018
types. A portion of the shift of capacity and generation is due to sales and transfers of generation
assets from traditional utilities to IPPs, rather than strictly the result of newly built units.
2-20
-------
CHAPTER 3: EMISSIONS AND AIR QUALITY IMPACTS
Overview
This Chapter describes the methods for developing spatial fields of air quality
concentrations for the baseline and regulatory control alternatives in 2021 and 2024. These
spatial fields provide the air quality inputs to potentially calculate health benefits for the Revised
CSAPR Update. The spatial fields for this rule were constructed using the method and air quality
modeling developed to support the regulatory impact analysis (RIA) for the Repeal of the Clean
Power Plan, and the Emission Guidelines for Greenhouse Gas Emissions from Existing Electric
Utility Generating Units (U.S. EPA 2019), also referred to as the Affordable Clean Energy
(ACE) rule.1
In Section 3.1 we describe the ACE air quality modeling platform; in Section 3.2 we
describe the ACE approach for processing the air quality modeling outputs to create inputs for
estimating benefits; in Section 3.3 we describe how the ACE approach was applied in the
Revised CSAPR Update, in Section 3.4 we present maps showing the impacts on ozone and
PM2.5 concentrations of each of the three regulatory control alternatives compared to the
corresponding baseline; and in Section 3.5 we identify uncertainties and limitations in the
application of the ACE approach for generating spatial fields of pollutant concentrations.
3.1 ACE Air Quality Modeling Platform
The air quality modeling for the ACE analysis utilized a 2011-based modeling platform
which included meteorology and base year emissions from 2011 and projected emissions for
2023. The air quality modeling included annual photochemical model simulations for a 2011
base year and a 2023 future year to provide hourly concentrations of ozone and primary and
secondarily formed PM2.5 component species (e.g., sulfate, nitrate, ammonium, elemental carbon
(EC), organic aerosol (OA), and crustal material2) for both years nationwide. In particular,
source apportionment modeling was performed for 2023 to quantify the contributions to ozone
and PM2.5 component species from coal-fired and non-coal electric generating units (EGUs) on a
1	Additional details on the ACE modeling and methodology for developing spatial fields of air quality for EGU
control strategies are provided in Appendix 3 A.
2	Crustal material refers to metals that are commonly found in the earth's crust such as Aluminum, Calcium, Iron,
Magnesium, Manganese, Potassium, Silicon, Titanium and the associated oxygen atoms.
3-1
-------
state-by-state or multi-state basis. As described below, the modeling results for 2011 and 2023,
in conjunction with emissions data for the baseline and regulatory control alternatives, were used
to construct the air quality spatial fields that reflect the influence of emissions changes between
the baseline and the regulatory control alternatives.
The air quality model simulations (i.e., model runs) were performed using the
Comprehensive Air Quality Model with Extensions (CAMx) (Ramboll Environ 2016). Our
CAMx nationwide modeling domain (i.e.., the geographic area included in the modeling) covers
all lower 48 states plus adjacent portions of Canada and Mexico using a horizontal grid
resolution of 12 x 12 km shown in Figure 3-1.
'Qj '1— i—


RT7C -

/
\ J _m
/ \

¦ /
\ •
Be V
V JT1
1
• *\
	r	
TV
tlusi

V ^ ^
11 ongwv. tiutMls
«11N ro» 244 t i \
V •
t
Figure 3-1. Air Quality Modeling Domain
The impact of specific emissions sources on ozone and PM2.5 in the 2023 modeled case
was tracked using a tool called "source apportionment." In general, source apportionment
modeling quantifies the air quality concentrations formed from individual, user-defined groups
of emissions sources or "tags". These source tags are tracked through the transport, dispersion,
chemical transformation, and deposition processes within the model to obtain hourly gridded3
contributions from the emissions in each individual tag to hourly modeled concentrations of
ozone and PM2.5.4 Thus, the source apportionment method provides an estimate of the effect of
3	Hourly contribution information is provided for each grid cell to provide spatial patterns of the contributions from
each tag.
4	Note that the sum of the contributions in a model grid cell from each tag for a pollutant equals the total
concentration of that pollutant in the grid cell.
3-2
-------
changes in emissions from each group of emissions sources {i.e., each tag) to changes in ozone
and PM2.5 concentrations. For this analysis we applied outputs from source apportionment
modeling for ozone and PM2.5 using the 2023 modeled case to obtain the contributions from
EGU emissions as well as other sources to ozone and to PM2.5 component species
concentrations.5 Ozone contributions were modeled using the Ozone Source Apportionment
Technique/Anthropogenic Precursor Culpability Assessment (OSAT/APCA) tool and PM2.5
component species contributions were modeled using the Particulate Source Apportionment
Technique (PSAT) tool.6 The source apportionment modeling, which was already available from
analysis performed to support the ACE rule RIA (U.S. EPA, 2019) was used to quantify the
contributions from EGU emissions on a state-by-state or, in some cases, on a multi-state basis.
For ozone, we modeled the contributions from the 2023 EGU sector emissions of NOx and VOC
to hourly ozone concentrations for the period April through October to provide data for
developing spatial fields for two seasonal ozone benefits metrics identified above (i.e., for the
May-September seasonal average of the maximum daily 8-hour average (MDA8) ozone and the
April-October seasonal average of the maximum daily 1-hour average (MDA1) ozone). For
PM2.5, we modeled the contributions from the 2023 EGU sector emissions of SO2, NOx, and
directly emitted PM2.5 for the entire year to inform the development of spatial fields of annual
mean PM2.5. For each state, or multi-state group, we separately tagged EGU emissions depending
on whether the emissions were from coal-fired units or non-coal units.7 In addition to tagging
coal-fired and non-coal EGU emissions we also tracked the ozone and PM2.5 contributions from
all other sources.
3.2. Applying Modeling Outputs to Create Spatial Fields
In this section we describe the ACE approach for creating spatial fields based on the 2011
and 2023 modeling performed for the ACE rule. The foundational data from ACE include the
ozone contributions from EGU emissions in each state based on the 2023 ACE EGU state-sector
sector contribution modeling and the 2023 emissions for coal and non-coal fired EGUs that were
5	In the source apportionment modeling for PM2 5 we tracked the source contributions from primary, but not
secondary organic aerosols (SOA). The method for treating SOA concentrations is described in U.S. EPA, 2019
chapter 8.
6	OSAT/APCA and PSAT tools are described in Ramboll Environ (2016).
7	For the purposes of this analysis non-coal fuels include emissions from natural gas, oil, biomass, municipal waste
combustion and waste coal EGUs.
3-3
-------
input to that modeling. These data are used to generate spatial fields based on ozone season EGU
NOx emissions (tons) and annual total EGU emissions of NOx, S02 and PM2.5. The inputs for
this method include emissions for each state with a breakout of emissions for coal-fired and non-
coal EGUs. The ozone season NOx emissions are used to prepare spatial fields of the May-
September seasonal average MDA8 ozone and the April-October seasonal average MDA1 ozone
concentration and the annual emissions are used to prepare spatial fields of annual PM2.5
concentrations. This method calculates the scaling ratios, described below, that are used to
prepare the air quality spatial fields.
To create the spatial fields for each future emissions scenario the 2023 state-sector source
apportionment modeling outputs from the ACE modeling described above are used in
combination with the EGU SO2, NOx, and PM2.5 emissions for each scenario. Contributions from
each state-sector contribution "tag" were scaled based on the ratio of emissions in the
year/scenario being evaluated to the emissions in the modeled ACE 2023 scenario. In this
approach, scaling ratios for PM2.5 components that are emitted directly from the source (OA, EC,
crustal) are based on relative changes in annual primary PM2.5 emissions between the modeled
ACE 2023 emissions scenario and the specific baseline or control scenario being analyzed. Also
the scaling ratios for components that are formed through chemical reactions in the atmosphere
were created as follows: scaling ratios for sulfate were based on relative changes in annual SO2
emissions; scaling ratios for nitrate were based on relative changes in annual NOx emissions; and
scaling ratios for ozone formed in NOx-limited regimes8 ("03N") were based on relative
changes in ozone season (May-September) NOx emissions. Tags representing sources other than
EGUs are held constant at 2023 ACE baseline levels for emissions scenarios analyzed by the
user. For each control scenario analyzed, the scaled contributions from all sources were summed
together to create a gridded surface of total modeled ozone or total modeled PM2.5. Finally,
spatial fields of ozone and PM2.5 are created based on "fusing" modeled data with measured
concentrations at air quality monitoring locations. The process is described in a step-by-step
manner below.
(1) The EGU annual SO2, NOx, and directly emitted PM2.5 emissions for the control scenario
of interest and the corresponding 2023 SO2, NOx, and directly emitted PM2.5 emissions
8 The CAMx model internally determines whether the ozone formation regime is NOx-limited or VOC-limited
depending on predicted ratios of indicator chemical species.
3-4
-------
used in the ACE modeling to calculate the ratio of control case emissions to the ACE
emissions for each of these pollutants for each EGU tag.
(2)	The tag-specific 2024 to 2023 EGU emissions-based scaling ratios from step (1) are
multiplied by the corresponding 365 gridded daily 24-hour average PM2.5 component
species contributions from the 2023 contribution modeling. The emissions ratios for SO2
are applied to sulfate contributions; ratios for annual NOx are applied to nitrate
contributions; and ratios for directly emitted PM2.5 are applied to the EGU contributions
to primary OA, EC and crustal material. This step results in 365 adjusted gridded daily
PM2.5 component species contributions for each EGUs tag that reflects the emissions in
the control scenario.
(3)	For each individual PM2.5 component species, the adjusted gridded contributions for each
EGU tag from step (2) are added together to produce a gridded daily EGU tag total.
(4)	The daily total EGU contributions for each PM2.5 component species from step (3) are
then combined with the species contributions from source tags representing all other
sources of PM2.5. As part of this step we also add the total secondary organic aerosol
concentrations from the 2023 ACE modeling to the net EGU contributions of primary
OA. Note that the secondary organic aerosol concentration does not change between
scenarios. This step results in 24-hour average PM2.5 component species concentrations
for the control scenario in each model grid cell, nationwide for each day in the year.
(5)	For each PM2.5 component species, the daily concentrations from step (4) are averaged
for each quarter of the year.
(6)	The quarterly average PM2.5 component species concentrations from step (5)9 are divided
by the corresponding quarterly average species concentrations from the base period air
quality model run. This step provides a Relative Response Factor (i.e., RRF) between the
base period and the control scenario for each species in each model grid cell.
(7)	The species-specific quarterly RRFs from step (6) are then multiplied by the
corresponding species-specific quarterly average concentrations from the base period
9 Ammonium concentrations are calculated assuming that the degree of neutralization of sulfate ions remains at
2011 levels (see Chapter 8 of U.S. EPA, 2019 for details).
3-5
-------
fused surfaces to produce quarterly average species concentrations for the control
scenario.
(8) The quarterly average species concentrations from step (7) are summed over the species
to produce total PM2.5 concentrations for each quarter. Finally, total PM2.5 concentrations
for the four quarters of the year are averaged to produce the spatial field of annual
average PM2.5 concentrations for the 2024 baseline.
To generate the spatial fields for each of the two ozone concentration metrics (i.e., April-
October MDA1 and May-September MDA8) we follow the steps similar to those above for
PM2.5.
(1)	The EGU May through September (i.e., Ozone Season - OS) NOx for the control scenario
and the corresponding modeled 2023 OS NOx emissions are used to calculate the ratio of
control scenario emissions to 2023 ACE emissions for each EGU tag (i.e. an ozone-
season scaling factor for each tag).
(2)	The source apportionment modeling provided separate ozone contributions for ozone
formed in VOC-limited chemical regimes (O3V) and ozone formed in NOx-limited
chemical regimes (O3N).10 The EGU OS NOx emissions for the control scenario and the
2023 ACE OS NOx baseline emissions are used to calculate the ratio of the control
scenario emissions to the 2023 ACE emission to create the EGU NOx emissions scaling
ratios. The emissions scaling ratios are multiplied by the corresponding O3N gridded
daily contributions to MDA1 and MDA8 concentrations. This step results in adjusted
gridded daily MDA1 and MDA8 contributions due to NOx changes for each EGUs tag
that reflect the emissions in the 2024 baseline.
(3)	For MDA1 and MDA8, the adjusted contributions for each EGU tag from step (2) are
added together to produce a daily adjusted EGU tag total. Since IPM does not output
VOC from EGUs, there are no predicted changes in VOC emissions in these scenarios so
the O3V contributions remain unchanged. The contributions from the unaltered O3V tags
from the 2023 ACE modeling are added to the summed adjusted O3N EGU tags.
10 Information on the treatment of ozone contributions under NOx-limited and VOC-limited chemical regimes in the
CAMx APCA source apportionment technique can be found in the CAMx v6.40 User's Guide (Ramboll, 2016).
3-6
-------
(4)	The daily total EGU contributions for MDA1 and MDA8 from step (3) are then
combined with the contributions to MDA1 and MDA8 from all other sources. This step
results in MDA1 and MDA8 concentrations for the control scenario in each model grid
cell, nationwide for each day in the ozone season.
(5)	For MDA1, we average the daily concentrations from step (4) across all the days in the
period April 1 through October 31. For MDA8, we average the daily concentrations
across all days in the period May 1 through September 30.
(6)	The seasonal mean concentrations from step (5) are divided by the corresponding
seasonal mean concentrations from the base period air quality model run. This step
provides a Relative Response Factor (i.e., RRF) between the base period and control
scenario for MDA1 and MDA8 in each model grid cell.
(7)	Finally, the RRFs for the seasonal mean metrics from step (6) are then multiplied by the
corresponding seasonal mean concentrations from the base period MDA1 and MDA8
fused surfaces to produce seasonal mean concentrations for MDA1 and MDA8 for the
control scenario that are input to BenMAP-CE.
3.3 Application of ACE Approach for the Revised CSAPR Update
In this section we describe how we applied the ACE approach to generate spatial fields of
seasonal ozone and annual PM2.5 concentrations associated with the regulatory control
alternatives (i.e., the rule and the less stringent and more stringent alternatives) in this rule RIA.
The data for creating the Revised CSAPR Update spatial fields include EGU emissions for the
2021 and 2024 baseline and the regulatory control alternatives. The EGU emissions include OS
NOx and annual NOx, S02, and PM2.5 for coal-fired and non-coal units in each state in the
continental U.S. These EGU emissions are taken from the electricity sector analysis described in
Chapter 4. In the case of the Revised CSAPR Update analysis, there are no impacts on SO2 or
PM2.5 emissions in the regulatory control scenarios compared to the 2024 baseline.
To potentially calculate ozone-related benefits in 2021 and 2024 we used the ozone season
EGU NOx emissions (tons) for the 2021 and 2024 baseline along with emissions for the rule, and
each of the two other regulatory control alternatives. These emissions were applied using the
ACE approach and source apportionment data to produce spatial fields of the May-September
3-7
-------
seasonal average MDA8 ozone and the April-October seasonal average MDA1 ozone
concentrations as described in the previous section.
In 2021, the only control measures expected to be adopted for compliance in each of the
regulatory control alternatives include optimization of existing SCRs and SNCRs beginning in
May of 2021, and these measures will operate only during the ozone season. This is relevant
because NOx reductions in the ozone season provide minimal PM2.5 reductions since PM2.5
nitrate concentrations, which result from conversion of NOx emissions to nitrate, are minimal
during the warmer temperatures during the ozone season. Conversely, the conversion of nitrates
to PM2.5 is much greater in cooler (non-ozone season) months, and thus it would be considered
worthwhile to estimate PM2.5 benefits from NOx reductions in those months (Hand et al.,
2012). In 2024, the presence of additional control measures that operate year-round and other
changes in market conditions as a result of the rule lead to notable NOx reductions in the winter
months.
To create spatial fields for PM2.5 we pre-processed the 2024 coal and non-coal fired EGU
emissions in order to obtain annual emissions of NOx, S02, and directly emissions PM2.5 in a
manner that is appropriate for assessing the impacts on annual average PM2.5 concentrations.
This additional pre-processing was needed because the vast majority of the emissions reductions
are expected to occur during the ozone season but, as noted above, PM2.5 nitrate concentrations
are lowest during that time of year. In this regard, simply treating the summer emissions
reductions as if they were abated proportionately throughout the year would overstate the
impacts of the emissions reductions on PM2.5 and therefore overstate benefits associated with
reducing exposure to PM2.5. For those states in which there are NOx emissions reductions during
the ozone season only, we reset the annual NOx emission in the regulatory alternative to be
equivalent to the corresponding baseline emissions to avoid distributing the ozone season
reductions across the entire year. That is, we assumed that there would be no impact on PM2.5
nitrate concentrations of NOx reductions in the ozone season. For those states in which there are
NOx emissions changes between the baseline and regulatory control alternative outside of the
ozone season, we accounted for those reductions by "annualizing" the EGU emissions for the
period outside the ozone season in the regulatory alternative as well as the corresponding
baseline. This method essentially applies the change in NOx tons outside the ozone season on a
3-8
-------
daily basis to changes in NOx emissions tons within the ozone season.11 With this adjustment the
impact of the regulatory control alternative on annual average PM2.5 concentrations reflects the
emissions reductions that will occur outside the ozone season when PM2.5 nitrate concentrations
are highest. The emissions of SO2 and directly emitted PM2.5 in 2024 for each of the regulatory
alternatives do not change from the 2024 baseline. That is, the regulatory control alternatives
analyzed in this RIA reduce emissions of NOx, but do not impact emissions of SO2 and directly
emitted PM2.5.
3.4 Spatial Distribution of Air Quality Impacts
Below we present the estimated impacts on May-September MDA8 ozone12 between the
baseline and each of the regulatory control alternatives for 2021 and 2024 as well as the
estimated impacts on annual mean PM2.5 concentrations between the baseline and the regulatory
control alternatives in 2024 (Figure 3-2 through Figure 3-10). The data shown in these figures
are calculated as the baseline minus the regulatory control alternative concentrations (i.e.,
positive values indicate reductions in pollutant concentrations). The spatial patterns of the
impacts of emissions reductions are a result of (1) the spatial distribution of EGU sources that are
predicted to have changes in emissions and (2) the physical or chemical processing that the
model simulates in the atmosphere.
11	In all states the actual tons reduced in the ozone season is greater than or equal to the change outside the ozone
season between the baseline and the regulatory alternatives.
12	The estimated impacts on April-October 2021 and 2024 ozone for each scenario are not shown but are similar to
May-September impacts available in Figure 12-20.
3-9
-------
i	00	IS'j	238	31?	396
M«n ¦ 0 OOE+O at {1,1k, Ma* * 0.000 at (299,169)
Figure 3-2. Map of change in May-September MDA8 ozone (ppb):
2021 baseline - less stringent regulatory alternative (scale: + 0.10 ppb)
0.50
0.40
030
020
010
000
•010
-0.20
-0.30
-0,40
-0,50
Figure 3-3. Map of change in May-September MDA8 ozone (ppb):
2021 baseline - rule (scale: + 0.50 ppb)

¦1»n = O OOE+O at [l.llt, Mac = 1 109 at (300,1341
3-10
-------
0.50
0.40
0.30
0.20
0.10
0.00
-010
¦020
-0.30
-0.40
-0.50
Figure 3-4. Map of change in May-September MDA8 ozone (ppb):
2021 baseline - more stringent regulatory alternative (scale: + 0.50 ppb)
= O OOI+O St (1.JJ, Max = 1189 at (36B,1S«»
i
n

010
0.03
0.06
004
0.02
0.00
-0.02
-0.04
•0.06
-0.08
-0.10
U.l>. MX = OOM a< <299,169)
Figure 3-5. Map of change in May-September MDA8 ozone (ppb):
2024 baseline - less stringent regulatory alternative (scale: + 0.10 ppb)
3-11
-------
Figure 3-6. Map of change in May-September MDA8 ozone (ppb):
2024 baseline - rule (scale: + 0.50 ppb)
Delta
Figure 3-7. Map of change in May-September MDA8 ozone (ppb):
2024 baseline - more stringent regulatory alternative (scale: + 0.50 ppb)
3-12
-------
0.005
0.004
0.003

0.002
1
0.001
?
0.000
-0.001
-0.002
-0.003
-0.004
-0.005
Mln = O OOt+O at (lrllr s 1 971-3 at
Figure 3-8. Map of change in annual mean PM2.5 (pg/m3):
2024 baseline - rule (scale: + 0.005 pg/m3)
3.5 Uncertainties and Limitations of ACE Approach
One limitation of the scaling methodology for creating PM2.5 surfaces associated with the
baseline and regulatory alternatives described above is that it treats air quality changes from the
tagged sources as linear and additive. It therefore does not account for nonlinear atmospheric
chemistry and does not account for interactions between emissions of different pollutants and
between emissions from different tagged sources. This is consistent with how air quality
estimations have been treated in past regulatory analyses (U.S. EPA 2012; 2019; 2020b). We
note that air quality is calculated in the same manner for the baseline and the regulatory
alternatives, so any uncertainty associated with these assumptions is carried through both sets of
scenarios in the same manner and is thus not expected to impact the air quality differences
between scenarios. In addition, emissions changes between baseline and the regulatory
alternatives are relatively small compared to modeled 2023 emissions that form the basis of the
ACE source apportionment approach. Previous studies have shown that air pollutant
concentrations generally respond linearly to small emissions changes of up to 30 percent
(Dunker et al., 2002; Cohan et al., 2005; Napelenok et al., 2006; Koo et al., 2007; Zavala et al.,
2009; Cohan and Napelenok, 2011) and that linear scaling from source apportionment can do a
reasonable job of representing impacts of 100 percent of emissions from individual sources
(Baker and Kelly 2014). Therefore, while simplistic, it is reasonable to expect that the emissions
concentration differences between the baseline and regulatory control alternatives can be
3-13
-------
adequately represented using this methodology and any uncertainty should be weighed against
the speed in which this method may be used to account for spatial differences in the effect of
EGU emissions on ozone and PM2.5 concentrations.
A second limitation is that the source apportionment PM2.5 contributions represent the
spatial and temporal distribution of the emissions from each source tag as they occur in the 2023
modeled case. Thus, the contribution modeling results do not allow us to represent any changes
to "within tag" spatial distributions. As a result, the method does not account for any changes of
spatial patterns that would result from changes in the relative magnitude of sources within a
source tag in the scenarios investigated here. As described above, the EGU tags are generally by
state and by two EGU types; one for coal-fired units and one for non-coal units.
In addition, the 2023 CAMx-modeled concentrations themselves have some uncertainty.
While all models have some level of inherent uncertainty in their formulation and inputs, the
base-year 2011 model outputs have been evaluated elsewhere against ambient measurements
(U.S. EPA 2017; 2019) and have been shown to adequately reproduce spatially and temporally
varying ozone and PM2.5 concentrations.
The regulatory alternatives lead to decreased concentrations of ozone and PM2.5, the
extent to which varies by location, relative to the baseline. However, the analysis does not
account for how interaction with NAAQS compliance would affect the benefits and costs of the
regulatory alternatives, which introduces uncertainty in the benefits and costs of the alternatives.
To the extent the Revised CSAPR Update will decrease NOx and consequentially ozone and
PM2.5, these changes may affect compliance with existing NAAQS standards and subsequently
affect the actual benefits and costs of the rule. In areas not projected to attain the 2015 ozone
NAAQS without further emissions reductions from the baseline, states may be able avoid
applying some emissions control measures to reduce emissions from local sources as a result of
this rule. If compliance behavior with the 2015 ozone NAAQS were accounted for in the
baseline in this RIA there may be additional benefits from reduced compliance costs, while the
level and spatial pattern of changes in ozone and PM2.5 concentrations, and their associated
health and ecological benefits, would differ.
3-14
-------
Similarly, the regulatory alternatives may project decreases in ozone and PM2.5
concentrations in areas attaining the NAAQS in the baseline. In practice, these potential changes
in concentrations may influence NAAQS compliance plans in these areas, which in turn would
further influence concentrations and the cost of complying with the NAAQS. However, such
behavior will be mitigated by NAAQS requirements such as Prevention of Significant
Deterioration (PSD) requirements. This RIA does not account for how interaction with NAAQS
compliance would affect the benefits and costs of the regulatory alternatives.
3.6 References
Baker, Kirk R., and James T. Kelly. 2014. "Single Source Impacts Estimated with
Photochemical Model Source Sensitivity and Apportionment Approaches." Atmospheric
Environment 96 (October): 266-74. https://doi.Org/10.1016/j.atmosenv.2014.07.042.
Cohan Daniel S., Amir Hakami, Yongtao Hu, Armistead G. Russell. 2005. "Nonlinear response
of ozone to emissions: Source apportionment and sensitivity analysis." Environmental
Science & Technology 39:6739-6748
Cohan, Daniel, and Sergey Napelenok. 2011. "Air Quality Response Modeling for Decision
Support." Atmosphere 2 (December): 407-25. https://doi.org/10.3390/atmos2030407.
Ding, Dian, Yun Zhu, Carey Jang, Che-Jen Lin, Shuxiao Wang, Joshua Fu, Jian Gao, Shuang
Deng, Junping Xie, and Xuezhen Qiu. 2016. "Evaluation of Health Benefit Using
BenMAP-CE with an Integrated Scheme of Model and Monitor Data during Guangzhou
Asian Games." Journal of Environmental Sciences 42 (April): 9-18.
https://doi.Org/10.1016/j.jes.2015.06.003.
Dunker, Alan M., Greg Yarwood, Jerome P. Ortmann, and Gary M. Wilson. 2002. "The
Decoupled Direct Method for Sensitivity Analysis in a Three-Dimensional Air Quality
Model Implementation, Accuracy, and Efficiency." Environmental Science &
Technology 36 (13): 2965-76. https://doi.org/10.1021/es0112691.
Hand, J. L., B.A. Schichtel, M. Pitchford, W.C. Malm, and N.H. Frank. 2012. "Seasonal
Composition of Remote and Urban Fine Particulate Matter in the United States. Journal
of Geophysical Research, 117, D05209, doi: 10.1029/2011JD017122.
Koo, Bonyoung, Alan M. Dunker, and Greg Yarwood. 2007. "Implementing the Decoupled
Direct Method for Sensitivity Analysis in a Particulate Matter Air Quality Model."
Environmental Science & Technology 41 (8): 2847-54.
https://doi.org/10.1021/es0619962.
3-15
-------
Napelenok, Sergey L., Daniel S. Cohan, Yongtao Hu, and Armistead G. Russell. 2006.
"Decoupled Direct 3D Sensitivity Analysis for Particulate Matter (DDM-3D/PM)."
Atmospheric Environment A0 (32): 6112-21.
https://doi.Org/10.1016/j.atmosenv.2006.05.039.
Ramboll Environ. 2016. "Comprehensive Air Quality Model with Extensions Version 6.40."
User's Guide. Novato, CA: Ramboll Environ International Corporation.
http://www.camx.com/files/camxusersguide_v6-40.pdf.
US EPA, 2012. "Regulatory Impact Analysis for the Final Revisions to the National Ambient Air
Quality Standards for Particulate Matter." EPA-452/R-12-005. Research Triangle Park,
NC: U.S. Environmental Protection Agency.
https://www3.epa.gov/ttnecasl/regdata/RIAs/finalria.pdf.
US EPA, 2017. Documentation for the EPA's Preliminary 2028 Regional Haze Modeling.
Research Triangle Park, NC
(https://www3.epa.gov/ttn/scram/reports/2028_Regional_Haze_Modeling-TSD.pdf).
US EPA, 2019. "Regulatory Impact Analysis for the Repeal of the Clean Power Plan, and the
Emission Guidelines for Greenhouse Gas Emissions from Existing Electric Utility
Generating Units." EPA-452/R-19-003. Research Triangle Park, NC: U.S. Environmental
Protection Agency, https://www.epa.gov/sites/production/files/2019-
06/documents/utilities_ria_final_cpp_repeal_and_ace_2019-06.pdf.
US EPA, 2020a. "Regulatory Impact Analysis for Revisions to the Effluent Limitations
Guidelines and Standards for the Steam Electric Power Generating Point Source
Category". EPA-821-R-20-004. Washington, DC: U.S. Environmental Protection
Agency, https://www.epa.gov/sites/production/files/2020-
08/documents/steam_electric_elg_2020_final_reconsideration_rule_regulatory_impact_a
nalysis.pdf.
US EPA, 2020b. "Analysis of Potential Costs and Benefits for the "National Emission Standards
for Hazardous Air Pollutants: Coal- and Oil-Fired Electric Utility Steam Generating
Units - Subcategory of Certain Existing Electric Utility Steam Generating Units Firing
Easter". Memo to Docket for rulemaking: "National Emission Standards for Hazardous
Air Pollutants: Coal- and Oil-Fired Electric Utility Steam Generating Units -
Subcategory of Certain Existing Electric Utility Steam Generating Units Firing Eastern
Bituminous Coal Refuse for Emissions of Acid Gas Hazardous Air Pollutants" (EPA-
HQ-OAR-2018-0794), April 8, 2020. Available at:
https://www.epa.gOv/sites/production/files/2020-04/documents/mats_coal_refuse_cost-
benefit_memo.pdf
Zavala, M., Lei, W., Molina, M.J., Molina, L.T., 2009. Modeled and observed ozone sensitivity
to mobile-source emissions in Mexico City. Atmos. Chem. Phys. 9, 39-55.
3-16
-------
APPENDIX 3A: METHODOLOGY FOR DEVELOPING AIR QUALITY SURFACES
In this appendix we describe the air quality modeling platform and methodology that was
leveraged to prepare the air quality surfaces that could inform the calculation of health benefits
of the Revised CSAPR Update final rule. The modeling and methodology described here were
developed to support the Regulatory Impact Analysis for the Repeal of the Clean Power Plan,
and the Emission Guidelines for Greenhouse Gas Emissions from Existing Electric Utility
Generating Units (U.S. EPA 2019), also referred to the Affordable Clean Energy (ACE) rule.
The foundational data in the ACE approach included the 2023 ACE baseline EGU emissions and
the 2023 ACE EGU air quality contribution data described below. To generate spatial fields for
alternative EGU scenarios, such as the scenarios analyzed for the Revised CSAPR Update, the
user provides as input EGU emissions for coal-fired and non-coal units for each state, separately.
Ozone season EGU NOx emissions (tons) are used to prepare spatial fields of the May-
September seasonal average MDA8 ozone and the April-October seasonal average MDA1 ozone
concentrations and annual total EGU emissions of NOx, S02 and PM2.5 are used to prepare
spatial fields of annual PM2.5 concentrations. Emissions scaling ratios, described below, that are
used to prepare the air quality spatial fields.
3A.1 Air Quality Modeling Platform for the ACE Rule
As part of the ACE assessment we used existing air quality modeling for 2011 and 2023
to estimate PM2.5 and ozone concentrations in the future years analyzed for the ACE final rule.
The modeling platform consists of several components including the air quality model,
meteorology, estimates of international transport, and base year and future year emissions from
anthropogenic and natural sources. An overview of each of these platform comments is provided
in the subsections below.
3A. 1.1 Air Quality Model, Meteorology and Boundary Conditions
We used the Comprehensive Air Quality Model with Extensions (CAMx version 6.40)
with the Carbon Bond chemical mechanism CB6r4 for modeling base year and future year ozone
and PM2.5 concentrations (Ramboll, 2016). CAMx is a three-dimensional grid-based
photochemical air quality model designed to simulate the formation and fate of oxidant
precursors, primary and secondary particulate matter concentrations, and deposition over
3A-1
-------
national, regional and urban spatial scales. Consideration of the different processes (e.g.,
transport and deposition) that affect primary (directly emitted) and secondary (formed by
atmospheric processes) pollutants in different locations is fundamental to understanding and
assessing the effects of emissions on air quality concentrations.
The geographic extent of the modeling domain covers the 48 contiguous states along with
the southern portions of Canada and the northern portions of Mexico as shown in Figure!. This
modeling domain contains 25 vertical layers with a top at about 17,550 meters1 and horizontal
grid resolution of 12 km x 12 km. The model simulations produce hourly air quality
concentrations for each 12-krn grid cell across the modeling domain.
1 -Vl ;
A , '
y* —
fK
—.Nkf. M
j)
m* \
1 1 \ > ~
¦4

, 1 I
z?M
¦w
]• !	
¦ A
\ i ~ -
1 ^v»4, (
\ \ .a
\ m
• V V i
	 _ /
- \
tJW lUMto "
tiw^K JnjOOOm t'txcMoi /¦WV
Ml JM roo-Mt 1 \ V '¦
' f

y'
Figure 3A-1. Air Quality Modeling Domain
The 2011 meteorological data for air quality modeling were derived from running
Version 3.4 of the Weather Research Forecasting Model (WRF) (Skamarock, et a!., 2008). The
meteorological outputs from WRF include hourly-varying horizontal wind components (i.e.,
speed and direction), temperature, moisture, vertical diffusion rates, and rainfall rates for each
vertical layer in each grid cell. The 2011 meteorology was used for both the 2011 base year and
2023 future year air quality modeling. Details of the annual 2011 meteorological model
simulation and evaluation are provided in a separate technical support document (US EPA,
1 Since the model top is defined based on atmospheric pressure, the actual height of the model top varies somewhat
with time and location.
3A-2
-------
2014a) which can be obtained at:
http ://www. epa.gov/ttn/ scram/reports/MET_T SD_2011 _final_l 1 -26-14. pdf
The lateral boundary and initial species condition concentrations are provided by a three-
dimensional global atmospheric chemistry model, GEOS-Chem (Yantosca, 2004) standard
version 8-03-02 with 8-02-01 chemistry. The global GEOS-Chem model simulates atmospheric
chemical and physical processes driven by assimilated meteorological observations from the
NASA's Goddard Earth Observing System (GEOS-5).2 GEOS-Chem was run for 2011 with a
grid resolution of 2.0 degrees x 2.5 degrees (latitude-longitude). The predictions were used to
provide one-way dynamic boundary condition concentrations at three-hour intervals and an
initial concentration field for the CAMx simulations. The 2011 boundary concentrations from
GEOS-Chem were used for both the 2011 and 2023 model simulations. The procedures for
translating GEOS-Chem predictions to initial and boundary concentrations are described
elsewhere (Henderson, 2014). More information about the GEOS-Chem model and other
applications using this tool is available at: http://www-as.harvard.edu/chemistry/trop/geos.
3A.1.2 2011 and 2023 Emissions
The purpose of the 2011 base year modeling is to represent the year 2011 in a manner
consistent with the methods used in the 2023 future year base case. The emissions data in this
platform are primarily based on the 2011 National Emissions Inventory (NEI) v2 for point
sources, nonpoint sources, commercial marine vessels, nonroad mobile sources and fires.3 The
onroad mobile source emissions are similar to those in the 2011 NEIv2, but were generated using
the 2014a version of the Motor Vehicle Emissions Simulator (MOVES2014a)
(http://www.epa.gov/otaq/models/moves/). The 2011 and 2023 emission inventories incorporate
revisions implemented based on comments received on the Notice of Data Availability (NOD A)
2	Additional information is available at:
http://gmao.gsfc.nasa.gov/GEOS/ and http://wiki.seas.harvard.edu/geos-chem/index.php/GEOS-5).
3	Note that EPA used a more recent 2016-based emissions platform for air quality modeling to provide the
foundational data needed to identify receptors and interstate contributions for the rule. The 2016-based mobile
emissions platform data were based on MOVES2014b. The 2016-based emissions platform is described in the
Emissions Modeling Technical Support Document available at: https://www.epa.gov/air-emissions-
modeling/2016vl-platform. Although the modeling data in the ACE approach are based on the 2011 platform (and
the 2011-based platform mobile emissions data were developed using MOVES2014a), the state-EGU contribution
modeling data, as described in this appendix, provide a means to develop spatial fields of air quality for the 2021
and 2025 baseline and the rule and alternative control scenarios analyzed in this RIA.
3A-3
-------
issued in January 2017 "Preliminary Interstate Ozone Transport Modeling Data for the 2015
Ozone National Ambient Air Quality Standard" (82 FR 1733), along with revisions made from
prior notices and rulemakings on earlier versions of the 2011 platform. The preparation of the
emission inventories for air quality modeling is described in the Technical Support Document
(TSD) Additional Updates to Emissions Inventories for the Version 6.3, 2011 Emissions
Modeling Platform for the Year 2023 (US EPA, 2017a). Electronic copies of the emission
inventories and ancillary data used to produce the emissions inputs to the air quality model are
available from the 2011 en and 2023 en section of the EPA Air Emissions Modeling website for
the 2011v6.3 emissions modeling platform: https://www.epa.gov/air-emissions-modeling/2011-
version-63-platform.
The emission inventories for the 2023 ACE future year were developed using projection
methods that are specific to the type of emission source. Future emissions are projected from the
2011 current year either by running models to estimate future year emissions from specific types
of emission sources (e.g., EGUs, and onroad and nonroad mobile sources)4, or for other types of
sources by adjusting the base year emissions according to the best estimate of changes expected
to occur in the intervening years. For sectors which depend strongly on meteorology (such as
biogenic and fires), the same emissions are used in the base and future years to be consistent with
the 2011 meteorology used when modeling 2023. For the remaining sectors, rules and specific
legal obligations that go into effect in the intervening years, along with changes in activity for
the sector, are considered when possible. Emissions inventories for neighboring countries used in
our modeling are included in this platform, specifically 2011 and 2023 emissions inventories for
Mexico, and 2013 and 2025 emissions inventories for Canada. The meteorological data used to
create and temporalize emissions for the future year cases is held constant and represents the
year 2011. The same ancillary data files5 are used to prepare the future year emissions
inventories for air quality modeling as were used to prepare the 2011 base year inventories with
the exception of chemical speciation profiles for mobile sources and temporal profiles for EGUs.
4	California provided emissions for the modeling platform. As such, onroad mobile source emissions for California
were consistent with the emissions provided by the state.
5	Ancillary data files include temporal, spatial, and VOC/PM2 5 chemical speciation surrogates.
3A-4
-------
The projected EGU emissions reflect the emissions reductions expected due to the Final
Mercury and Air Toxics (MATS) rule announced on December 21, 2011, the Cross-State Air
Pollution Rule (CSAPR) issued July 6, 2011, and the CSAPR Update issued October 26, 2016.
The 2023 EGU projected inventory was developed using an engineering analysis approach. EPA
started with 2016 reported, seasonal, historical emissions for each unit. The emissions data for
NOx and SO2 for units that report data under either the Acid Rain Program (ARP) and/or the
CSAPR were aggregated to the summer/ozone season period (May-September) and winter/non-
ozone period (January-April and October-December).6 Adjustments to 2016 levels were made to
account for retirements, coal to gas conversion, retrofits, state-of-the-art combustion controls,
along with other unit-specific adjustments. Details and these adjustments, and information about
handling for units not reporting under Part 75 and pollutants other than NOx and SO2 are
described in the emissions modeling TSD (US EPA, 2017a).
The 2023 non-EGU stationary source emissions inventory includes impacts from
enforceable national rules and programs including the Reciprocating Internal Combustion
Engines (RICE) and cement manufacturing National Emissions Standards for Hazardous Air
Pollutants (NESHAPs) and Boiler Maximum Achievable Control Technology (MACT)
reconsideration reductions. Projection factors and percent reductions for non-EGU point sources
reflect comments received by EPA in response to the January 2017 NOD A, along with emissions
reductions due to national and local rules, control programs, plant closures, consent decrees and
settlements. Growth and control factors provided by states and by regional organizations on
behalf of states were applied. Reductions to criteria air pollutant (CAP) emissions from
stationary engines resulting from the Reciprocating Internal Combustion Engines (RICE)
National Emission Standard for Hazardous Air Pollutants (NESHAP) are included. Reductions
due to the New Source Performance Standards (NSPS) VOC controls for oil and gas sources, and
the NSPS for process heaters, internal combustion engines, and natural gas turbines were also
included.
6 EPA notes that historical state-level ozone season EGU NOx emission rates are publicly available and quality
assured data. They are monitored using continuous emissions monitors (CEMs) data and are reported to EPA
directly by power sector sources. They are reported under Part 75 of the CAA.
3A-5
-------
For point and nonpoint oil and gas sources, state projection factors were generated using
state-specific historical oil and gas production data available from EIA for 2011 to 2015 and
information from regional factors based AEO 2017 to project the emission to the year 2023.
Emission reductions of stationary engines CAP reductions (RICE NESHAP) and controls from
New Source Performance Standards (NSPS) are reflected for select source categories. Mid-
Atlantic Regional Air Management Association (MARAMA) factors for the year 2023 were used
where applicable. Projection factors for other nonpoint sources such as stationary source fuel
combustion, industrial processes, solvent utilization, and waste disposal, reflect emissions
reductions due to control programs along with comments on the growth and control of these
sources as a result of the January 2017 NOD A and information gathered from prior rulemakings
and outreach to states on emission inventories.
The MOVES2014a-based 2023 onroad mobile source emissions account for changes in
activity data and the impact of on-the-books national rules including: the Tier 3 Vehicle
Emission and Fuel Standards Program, the 2017 and Later Model Year Light-Duty Vehicle
Greenhouse Gas Emissions and Corporate Average Fuel Economy Standards (LD GHG), the
Renewable Fuel Standard (RFS2), the Mobile Source Air Toxics Rule, the Light Duty Green
House Gas/Corporate Average Fuel Efficiency (CAFE) standards for 2012-2016, the Greenhouse
Gas Emissions Standards and Fuel Efficiency Standards for Medium- and Heavy-Duty Engines
and Vehicles, the Light-Duty Vehicle Tier 2 Rule, and the Heavy-Duty Diesel Rule. The
MOVES-based emissions also include state rules related to the adoption of LEV standards,
inspection and maintenance programs, Stage II refueling controls, and local fuel restrictions.
The nonroad mobile 2023 emissions, including railroads and commercial marine vessel
emissions also include all national control programs. These control programs include the Clean
Air Nonroad Diesel Rule - Tier 4, the Nonroad Spark Ignition rules, and the Locomotive-Marine
Engine rule. For ocean-going vessels (Class 3 marine), the emissions data reflect the 2005
voluntary Vessel Speed Reduction (VSR) within 20 nautical miles, the 2007 and 2008 auxiliary
engine rules, the 40 nautical mile VSR program, the 2009 Low Sulfur Fuel regulation, the 2009-
2018 cold ironing regulation, the use of 1 percent sulfur fuel in the Emissions Control Area
(ECA) zone, the 2012-2015 Tier 2 NOx controls, the 2016 0.1 percent sulfur fuel regulation in
EC A zone, and the 2016 International Marine Organization (IMO) Tier 3 NOx controls. Non-
3A-6
-------
U.S. and U.S. category 3 commercial marine emissions were projected to 2025 using consistent
methods that incorporated controls based on ECA and IMO global NOx and SO2 controls.
3A.1.3 2011 Model Evaluation for Ozone and PM2.5
An operational model performance evaluation was conducted to examine the ability of
the 2011 base year model run to simulate the corresponding 2011 measured ozone and PM2.5
concentrations. This evaluation focused on four statistical metrics comparing model predictions
to the corresponding observations. The performance statistics include mean bias, mean error,
normalized mean bias, and normalized mean error. Mean bias (MB) is the sum of the difference
(predicted - observed) divided by the total number of replicates (n). Mean bias is given in units
of ppb and is defined as:
Where:
•	Pis the model-predicted concentration;
•	O is the observed concentrations; and
•	n is the total number of observations
Mean error (ME) calculates the sum of the absolute value of the difference (predicted -
observed) divided by the total number of replicates (n). Mean error is given in units of ppb and is
defined as:
Normalized mean bias (NMB) is the sum of the difference (predicted - observed) over the
sum of observed values. NMB is a useful model performance indicator because it avoids over
inflating the observed range of values, especially at low concentrations. Normalized mean bias is
given in percentage units and is defined as:
MB = -Zi(P-O)
n
(Eq-l)
ME = i£I|P-0|
(Eq-2)
™B=ilir*100
(Eq-3)
3A-7
-------
Normalized mean error (NME) is the sum of the absolute value of the difference
(predicted - observed) divided by the sum of observed values. Normalized mean error is given in
percentage units and is defined as:
NME=^f1*100	
-------
U.S. Climate Regions
cT	\f
Jr	'h
Figure 3A-2. NO A A Climate Regions
Model performance statistics for PM2.5 for each region are provided in Table 3A.1. These
data indicate that over the year as a whole, PM2.5 is over predicted in the Northeast, Ohio Valley,
Upper Midwest, Southeast, and Northwest regions and under predicted in the South and
Southwest regions. Normalized mean bias is within ±30 percent in all regions except the
Northwest which has somewhat larger model over-predictions. Model performance for PM2.5 for
the 2011 modeling platform is similar to the model performance results for other contemporary,
state of the science photochemical model applications (Simon et al., 2012). Additi onal details on
PM2.5 model performance for the 2011 base year model run can be found in the Technical
Support Document for EPA's preliminary regional haze modeling (US EPA, 2017b).
3 A-9
-------
Table 3A.1. Model Performance Statistics by Region for PM2.5
Region
Network
No. of Obs
MB
frig/m3)
ME
frig/m3)
NMB
(%)
NME
(%)
Northeast
IMPROVE
1577
0.87
2.21
17.70
44.90
CSN
2788
0.97
4.04
9.70
40.40
Ohio Valley
IMPROVE
680
0.10
2.96
1.20
35.50
CSN
2475
0.13
3.85
1.10
32.80
Upper Midwest
IMPROVE
CSN
700
1343
0.83
1.37
2.37
3.66
14.20
13.60
40.40
36.30
Southeast
IMPROVE
1172
0.52
3.54
6.30
43.20
CSN
1813
0.19
3.92
1.70
34.20
South
IMPROVE
933
-0.47
2.69
-6.50
37.40
CSN
962
-0.08
4.48
-0.75
39.50
Southwest
IMPROVE
3695
-1.12
1.86
-28.00
46.30
CSN
746
-0.08
3.93
-1.00
47.10
N. Rockies/
IMPROVE
1952
0.07
1.39
2.40
44.90
Plains
CSN
275
-2.07
4.18
-21.80
43.90
Northwest
IMPROVE
1901
1.19
2.28
43.20
82.90
CSN
668
5.77
7.25
69.90
87.90
West
IMPROVE
1782
-1.08
2.08
-25.30
48.50
CSN
936
-2.92
5.08
-23.10
40.30
Model performance statistics for May through September MDA8 ozone concentrations for
each region are provided in Table 3 A.2. Overall, measured ozone is under predicted in most
regions, except for the Northeast and Southeast where over prediction is found. Normalized
mean bias is within ±15 percent in all regions. Model performance for ozone for the 2011
modeling platform is similar to the model performance results for other contemporary, state of
the science photochemical model applications (Simon et al., 2012). Additional details on ozone
model performance for the 2011 base year model run can be found in the Air Quality Technical
Support Document for EPA's preliminary interstate ozone transport modeling for the 2015 ozone
National Ambient Air Quality Standard (US EPA, 2017c).
3 A-10
-------
Table 3A.2. Model Performance Statistics by Region for Ozone on Days Above 60 ppb
(May-Sep)
Region
No. of Obs
MB
(ppb)
ME
(ppb)
NMB
(%)
NME
(%)
Northeast
4085
1.20
7.30
1.80
10.70
Ohio Valley
6325
-0.60
7.50
-0.90
11.10
Upper Midwest
1162
-4.00
7.60
-5.90
11.10
Southeast
4840
2.30
6.80
3.40
10.20
South
5694
-5.30
8.40
-7.60
12.20
Southwest
6033
-6.20
8.50
-9.40
12.90
N. Rockies/Plains
380
-7.20
8.40
-11.40
13.40
Northwest
79
-5.60
9.00
-8.70
14.00
West
8655
-8.60
10.30
-12.20
14.50
Thus, the model performance results demonstrate the scientific credibility of our 2011
modeling platform for predicting PM2.5 and ozone concentrations. These results provide
confidence in the ability of the modeling platform to provide a reasonable projection of expected
future year ozone concentrations and contributions.
3A.2 Source Apportionment Tags
CAMx source apportionment modeling was used to track ozone and PM2.5 component
species impacts from pre-defined groups of emissions sources (source tags). Separate tags were
created for state-level EGUs split by fuel type (coal units versus non-coal units13). For some
states with low EGU emissions, EGUs are grouped with nearby states that also have low EGU
emissions. In addition, there are no coal EGUs operating in the 2023 emissions case for the
following states: Idaho, Oregon, and Washington. Therefore, there is no coal EGU tag for those
states. Similarly, there were no EGUs (coal or non-coal) in Washington D.C. in the 2023
emissions scenario, so there were no EGU tags for Washington D.C. There were also several
domain-wide tags for sources other than EGUs. Table 3 A.3 provides a full list of the emissions
group tags that were tracked in the source apportionment modeling.
13 For the purposes of this analysis non-coal fuels include emissions from natural gas, oil, biomass, and waste coal-
fired EGUs.
3 A-11
-------
Table 3A.3. Source Apportionment Tags
Coal-fired EGU tags
Non-coal EGU tags
Domain-wide tags
Alabama
Arizona
Arkansas
California
Colorado
Connecticut + Rhode Island
Delaware + New Jersey
Florida
Georgia
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine + Mass. + New Hamp. +
Vermont
Maryland
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Mexico
New York
North Carolina
North Dakota + South Dakota
Ohio
Oklahoma
Pennsylvania
South Carolina
Tennessee
Texas
Utah
Virginia
West Virginia
Wisconsin
Wyoming
Tribal Data*
Alabama
•
EGU retirements
Arizona

through 2025
Arkansas
•
EGU retirements
California

2026-2030
Colorado
•
All U.S.
Connecticut + Rhode Island

anthropogenic
Delaware + New Jersey

emissions from
Florida

source sectors
Georgia

other than EGUs
Idaho + Oregon + Washington
Illinois
•
International
within-domain
Indiana
Iowa

emissions

(sources
Kansas

occurring in
Canada, Mexico,
Kentucky

and from
Louisiana

offshore marine
Maine + Mass. + New Hamp. +

vessels and
Vermont

drilling
Maryland

platforms)
Michigan
•
Fires (wildfires
Minnesota

and prescribed
Mississippi

fires)
Missouri
•
Biogenic sources
Montana
•
Boundary
Nebraska

conditions
Nevada
New Mexico
New York
North Carolina
North Dakota + South Dakota
Ohio
Oklahoma
Pennsylvania
South Carolina
Tennessee
Texas
Utah
Virginia
West Virginia
Wisconsin
Wyoming
Tribal Data
14
14 EGUs operating on tribal lands were tracked together in a single tag. There are EGUs on tribal land in the
following states: Utah (coal), New Mexico (coal), Arizona (coal and non-coal), Idaho (non-coal). EGU emissions
occurring on tribal lands were not included in the state-level EGU source tags.
3A-12
-------
The contributions represent the spatial and temporal distribution of the emissions within
each source tag. Thus, the contribution modeling results do not allow us to represent any changes
to any "within tag" spatial distributions. For example, the location of coal-fired EGUs in
Michigan are held in place based on locations in the 2023 emissions. Additionally, the relative
magnitude of sources within a source tag do not change from what was modeled with the 2023
emissions inventory.
3A.3 Applying Source Apportionment Contributions to Create Air Quality Fields
We created air quality surfaces for the ACE future year baseline and illustrative policy
scenarios by scaling the EGU sector tagged contributions from the 2023 modeling based on
relative changes in EGU emissions associated with each tagged category between the 2023
emissions case and the ACE scenarios. Below, we provide equations used to apply these scaling
ratios along with tables of the ratios.
3A.3.2 Scaling Ratio Applied to Source Apportionment Tags
Scaling ratios for PM2.5 components that are emitted directly from the source (OA, EC,
crustal) were based on relative changes in annual primary PM2.5 emissions between the 2023
emissions case and the ACE baseline and the illustrative policy scenario. Scaling ratios for
components that are formed through chemical reactions in the atmosphere were created as
follows: scaling ratios for sulfate were based on relative changes in annual SO2 emissions;
scaling ratios for nitrate were based on relative changes annual NOx emissions; and scaling
ratios for ozone formed in NOx-limited regimes15 ("03N") were based on relative changes in
ozone season (May-September) NOx emissions. The scaling ratios that were determined based
on emissions provided for each scenario.
Scaling ratios were applied to create air quality surfaces for ozone using equation (9):
15 The CAMx model internally determines whether the ozone formation regime is NOx-limited or VOC-limited
depending on predicted ratios of indicator chemical species.
3 A-13
-------
Ozonemg d iy — C-in,g,d,BC "I" Cm,g,d,int Cm,g,d,bio ^m,g,d,fires
T
Cm,g,d,USanthro Cm,g,d,y,EGUret / CVOC,m,g,d,t
I'
^ ' CNOx,m,g,d,t^t,i,y
t=1
t=l
T
(Eq-9)
where:
•	Ozonem g d i y is the estimated ozone for metric, "m" (MDA8 or MDA1), grid-
cell, "g", day, "d", scenario, "i", and year, "y";
•	Cm,g,d,Bc's the total ozone contribution from the modeled boundary inflow;
Cm,g,d,int is the total ozone contribution from international emissions within the
model domain;
•	Cmg d bio is the total ozone contribution from biogenic emissions;
•	Cm,g,d,fireS's the total ozone contribution from fires;
•	Cmg d uSanthro is the total ozone contribution from U.S. anthropogenic sources
other than EGUs;
•	Cm,g,d,y,EGUret is the total ozone contribution from retiring EGUs after year, "y"
(this term is equal to 0 in 2030 and 2035);
•	CV0C mg d t is the ozone contribution from EGU emissions of VOCs from tag, "t";
•	CNOx,m,g,d,t is the ozone contribution from EGU emissions of NOx from tag, "t";
and
•	St i y is the ozone scaling ratio for tag, "t", scenario, "i", and year, "y".
Scaling ratios were applied to create air quality surfaces for PM2.5 species using equation
(10) (for sulfate, nitrate, EC or crustal material) or using equation (11) (for OA):
3A-14
-------
s,g,d,USanthro
's,g,d,y,EGUret
(Eq-10)
OAgdiy — CpQA g age + CpoA,g,d,int C po A,g,cl,bio CpOA,g,d,fires
CpOA,g,d,USanthro CpOA,g,d,y,EGUret SOAg C[
T
(Eq-11)
^ ' CpOA,g,d,tSpri,t,i,y
t=1
PMSig:d,i,y is the estimated concentration for species, "s" (sulfate, nitrate, EC, or crustal
material), grid-cell, "g", day, "d", scenario, "i", and year, "y";
Cs,g,d,BC is the species contribution from the modeled boundary inflow;
Cs,g,d,int is the species contribution from international emissions within the model
domain;
Cs,g,d,bio is the species contribution from biogenic emissions;
Cs,g,d,fires is the species contribution from fires;
Cs,g,d,usanthro is the species contribution from U.S. anthropogenic sources other than
EGUs;
Cs,g,d,y,EGUret is the species contribution from retiring EGUs after year, "y" (this term is
equal to 0 in 2030 and 2035);
Cs,g,d,t is the species contribution from EGU emissions from tag, "t"; and
Ss,t,i,y is the scaling ratio for species, "s", tag, "t", scenario, "i", and year, "y".
3 A-15
-------
Similarly, for Equation (11):
•	0Ag d i y is the estimated OA concentration for grid-cell, "g", day, "d", scenario, "i",
and year, "y";
•	Each of the contribution terms refers to the contribution to primary OA (POA); and
•	SOAgd represents the modeled secondary organic aerosol concentration for gird-
cell, "g", and day, "d", which does not change among scenarios
3A.4 Creating Fused Fields Based on Observations and Model Surfaces
In this section we describe steps taken to estimate PM2.5 and ozone gridded surfaces
associated with the baseline and the illustrative policy scenario for every year. For PM2.5, (daily
gridded PM2.5 species were processed into annual average surfaces which combine observed
values with model predictions using the enhanced Veronoi Neighbor Average (eVNA) method
(Gold et al., 1997; US EPA, 2007; Ding et al., 2015). These steps were performed using EPA's
software package, Software for the Modeled Attainment Test - Community Edition (SMAT-
CE)16 and have been previously documented both in the user's guide for the predecessor
software (Abt, 2014) and in EPA's modeling guidance document (U.S. EPA, 2014b). First, we
create a 2011 eVNA surface for each PM component species. To create the 2011 eVNA surface,
SMAT-CE first calculates quarterly average values (January-March; April-June; July-September;
October-December) for each PM2.5 component species at each monitoring site with available
measured data. For this calculation we used 3 years of monitoring data (2010-2012)17. SMAT-
CE then creates an interpolated field of the quarterly-average observed data for each PM2.5
component species using inverse distance squared weighting resulting in a separate 3-year
average interpolated observed field for each PM2.5 species and each quarter. The interpolated
observed fields are then adjusted to match the spatial gradients from the modeled data. These two
steps can be calculated using Equation (12):
16	Software download and documentation available at https://www.epa.gov/scram/photochemical-modeling-tools.
17	Three years of ambient data is used to provide a more representative picture of air pollution concentrations.
3 A-16
-------
eVNAg,s,q, 2011 = Z VK eightxMonitorxsq2010_2012 fl,s,g,2°11	(Eq-12)
X jSjCfj'Z, Oil
Where:
•	eVNAg s q current is the gradient adjusted quarterly-average eVNA value at grid-
cell, g, for PM component species, s, during quarter, q for the year 2011;
•	Weightx is the inverse distance weight for monitor x at the location of grid-cell,
g;
•	Monitorxsq,2010-2012 is the 3-year (2010-2012) average of the quarterly
monitored concentration for species, s, at monitor, x, during quarter, q;
•	Modelg s q 2011 is the 2011 modeled quarterly-average concentrations of species,
s, at grid cell, g, during quarter, q; and
•	Modelx s q 2011 is the 2011 modeled quarterly-average concentration of species, s,
at the location of monitor, x, during quarter q.
The 2011 eVNA field serves as the starting point for future-year projections. To create a
gridded future-year eVNA surfaces for the baseline and ACE illustrative policy, we take the ratio
of the modeled future year18 quarterly average concentration to the modeled 2011 concentration
in each grid cell and multiply that by the corresponding 2011 eVNA quarterly PM2.5 component
species value in that grid cell (Equation 13).
eVNAgSqjuture = (eVNAgsq 2011) x Mode^^20ll	(Eq-13)
This results in a gridded future-year projection which accounts for adjustments to match
observations in the 2011 modeled data.
Finally, particulate ammonium concentrations are impacted both by emissions of
precursor ammonia gas as well as ambient concentrations of particulate sulfate and nitrate.
18 In this analysis the "future yeaf' modeled concentration is the result of Equations 9, 10, or 11 that represents
either the ACE scenarios.
3 A-17
-------
Because of uncertainties in ammonium speciation measurements combined with sparse
ammonium measurements in rural areas, the SMAT-CE default is to calculate ammonium values
using the degree of sulfate neutralization (i.e., the relative molar mass of ammonium to sulfate
with the assumption that all nitrate is fully neutralized). Degree of neutralization values are
mainly available in urban areas while sulfate measurements are available in both urban and rural
areas. Ammonium is thus calculated by multiplying the interpolated degree of neutralization
value by the interpolated sulfate value at each grid-cell location which allows the ammonium
fields to be informed by rural sulfate measurements in locations where no rural ammonium
measurements are available. The degree of neutralization is not permitted to exceed the
maximum theoretical molar ratio of 2:1 for ammonium sulfate. When creating the future year
surface for particulate ammonium, we use the default SMAT-CE assumption that the degree of
neutralization for the aerosol remains at 2011 levels.
A similar method for creating future-year eVNA surfaces is followed for the two ozone
metrics with a few key differences. First, while PM2.5 is split into quarterly averages and then
averaged up to an annual value, we look at ozone as a summer-season average using definitions
that match metrics from epidemiology studies (May-Sep for MDA8 and Apr-Oct for MDA1).
The other main difference in the SMAT-CE calculation for ozone is that the spatial interpolation
of observations uses an inverse distance weighting rather than an inverse distance squared
weighting. This results in interpolated observational fields that better replicate the more gradual
spatial gradients observed in ozone compared to PM2.5.
3 A-18
-------
3A.5 References
82 FR 1733. Notice of Availability of the Environmental Protection Agency's Preliminary
Interstate Ozone Transport Modeling Data for the 2015 Ozone National Ambient Air
Quality Standard (NAAQS), (January 6, 2017).
Abt Associates, 2014. User's Guide: Modeled Attainment Test Software.
http://www.epa.gov/scram001/modelingapps_mats.htm
Cohan, D.S., Hakami, A., Hu, Y.T., Russell, A.G., 2005. Nonlinear response of ozone to
emissions: Source apportionment and sensitivity analysis. Environ. Sci. Technol. 39,
6739-6748.
Cohan, D.S., Napelenok, S.L., 2011. Air Quality Response Modeling for Decision Support.
Atmosphere. 2, 407-425.
Ding, D., Zhu, Y., Jang, C., Lin, C., Wang, S., Fu, J., Gao, J., Deng, S., Xie, J., Qui, X., 2015.
Evaluation of heath benefit using BenMAP-CE with an integrated scheme of model and
monitor data during Guangzhou Asian Games. Journal of Environmental Science. 29,
178-188.
Dunker, A.M., Yarwood, G., Ortmann, J.P., Wilson, G.M., 2002. The decoupled direct method
for sensitivity analysis in a three-dimensional air quality model—Implementation,
accuracy, and efficiency. Environ. Sci. Technol. 36, 2965-2976.
Gold C, Remmele P.R., Roos T., 1997. In: Algorithmic Foundation of Geographic Information
Systems. In: Lecture Notes in Computer Science, Vol. 1340 (van Kereveld M, Nievergelt
J, Roos T, Widmayer P, eds) Berlin, Germany: Springer-Verlag. Voronoi methods in
GIS. pp. 21-35.
Henderson, B.H., Akhtar, F., Pye, H.O.T., Napelenok, S.T., Hutzell, W.T., 2014. A Database and
Tool for Boundary Conditions for Regional Air Quality Modeling: Description and
Evaluations, Geoscientific Model Development. 7, 339-360.
Koo, B., Dunker, A.M., Yarwood, G., 2007. Implementing the decoupled direct method for
sensitivity analysis in a particulate matter air quality model. Environ. Sci. Technol. 41,
2847-2854.
Napelenok, S.L., Cohan, D.S., Hu, Y., Russell, A.G., 2006. Decoupled direct 3D sensitivity
analysis for particulate matter (DDM-3D/PM). Atmospheric Environment. 40, 6112-
6121.
Ramboll Environ, 2016. User's Guide Comprehensive Air Quality Model with Extensions
version 6.40. Ramboll Environ International Corporation, Novato, CA.
Simon, H., Baker, K. R., Phillips, S., 2012. Compilation and interpretation of photochemical
model performance statistics published between 2006 and 2012, Atmos. Environ. 61,
124-139.
3 A-19
-------
Skamarock, W.C., Klemp, J.B., Dudhia,J., Gill, D.O., Barker, D.M., Duda, M.G., Huang, X.-Y.,
Wang, W., Powers, J.G., 2008. A Description of the Advanced Research WRF Version 3.
NCAR Tech. Note NCAR/TN-475+STR.
(http://wwww.mmm.ucar.edu/wrf/users/docs/arw_v3.pdf).
US EPA, 2007, Technical Report on Ozone Exposure, Risk, and Impact Assessments for
Vegetation. EPA 452/R-07-002. Prepared by Abt Associates Inc. for U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards, Health and
Environmental Impacts Division. Research Triangle Park, NC.
(https://www3.epa.gOv/ttn/naaqs/standards/ozone/data/2007_01_environmental_tsd.pdf).
US EPA, 2014a. Meteorological Model Performance for Annual 2011 Simulation WRF v3.4,
Research Triangle Park, NC. (http://www.epa.gov/scram001/).
US EPA, 2014b, Modeling Guidance for Demonstrating Attainment of Air Quality Goals for
Ozone, PM2.5, and Regional Haze- December 2014 DRAFT, Research Triangle Park,
NC. (https://www3.epa.gov/ttn/scram/guidance/guide/Draft_03-PM-
RH_Modeling_Guidance-2014.pdf).
US EPA, 2015, Regulatory Impact Analysis of the Final Revisions to the National Ambient Air
Quality Standards for Ground-Level Ozone, EPA-452/R-15-07, Research Triangle Park,
NC. (https://www.epa.gov/sites/production/files/2016-02/documents/20151001ria.pdf).
US EPA, 2017a, Technical Support Document (TSD) Additional Updates to Emissions
Inventories for the Version 6.3, 2011 Emissions Modeling Platform for the Year 2023,
Research Triangle Park, NC. (https://www.epa.gov/sites/production/files/2017-
11/documents/ 201 Iv6.3_2023en_update_emismod_tsd_oct2017.pdf).
US EPA, 2017b. Documentation for the EPA's Preliminary 2028 Regional Haze Modeling.
Research Triangle Park, NC
(https://www3.epa.gov/ttn/scram/reports/2028_Regional_Haze_Modeling-TSD.pdf).
US EPA, 2017c. Air Quality Modeling Technical Support Document for the 2015 Ozone
NAAQS Preliminary Interstate Transport Assessment. Research Triangle Park, NC
(https://www.epa.gov/airmarkets/notice-data-availability-preliminary-interstate-ozone-
transport-modeling-data-2015-ozone).
Yantosca, B. 2004. GEOS-CHEMv7-01-02 User's Guide, Atmospheric Chemistry Modeling
Group, Harvard University, Cambridge, MA.
Zavala, M., Lei, W., Molina, M.J., Molina, L.T., 2009. Modeled and observed ozone sensitivity
to mobile-source emissions in Mexico City. Atmos. Chem. Phys. 9, 39-55.
3A-20
-------
CHAPTER 4: COST, EMISSIONS, AND ENERGY IMPACTS
Overview
This chapter reports the compliance costs, emissions, and energy analyses performed for
the Revised CSAPR Update final rule. EPA used the Integrated Planning Model (IPM) to
conduct most of the analysis discussed in this chapter. As explained in detail below, this chapter
presents analysis for three regulatory control alternatives that differ in the level of electric
generating units (EGU) nitrogen oxides (NOx) ozone season emissions budgets in the 12 states
subject to this action.1 These regulatory control alternatives impose different budget levels based
on different NOx mitigation technologies.
The chapter is organized as follows: following a summary of the regulatory control
alternatives analyzed and a summary of EPA's methodology, we present estimates of compliance
costs, as well as estimated impacts on emissions, generation, capacity, fuel use, fuel price, and
retail electricity price.
4.1 Regulatory Control Alternatives
Of the 22 states currently covered by the Cross-State Air Pollution Rule (CSAPR) NOx
Ozone Season Group 2 trading program, EPA is establishing revised budgets for 12 states.
Therefore, EPA is creating an additional geographic group and ozone season trading program
comprised of these 12 upwind states with remaining linkages to downwind air quality problems
in 2021. This new group, Group 3, will be covered by a new CSAPR NOx Ozone Season Group
3 trading program and will no longer be subject to Group 2 budgets. Aside from the removal of
the 12 covered states from the current Group 2 program, this rule leaves unchanged the budget
stringency and geography of the existing CSAPR NOx Ozone Season Group 1 and Group 2
trading programs. The EGUs covered by the FIPs and subject to the budget are fossil-fired EGUs
with >25 megawatt (MW) capacity.
This regulatory impact analysis (RIA) evaluates the benefits, costs and certain impacts of
compliance with three regulatory control alternatives: the final Revised CSAPR Update, a less-
1 The 12 states for which EPA is promulgating FIPs to reduce interstate ozone transport for the 2008 ozone NAAQS
are listed in Table I.A-2 of the preamble and are Illinois, Indiana, Kentucky, Louisiana, Maryland, Michigan, New
York, New Jersey, Ohio, Pennsylvania, Virginia, and West Virginia.
4-1
-------
stringent alternative, and a more-stringent alternative. For details on the derivation of these
budgets, please see Section VI of the preamble. Aside from the difference in emission budgets,
other key regulatory features of the allowance trading program, such as the ability to bank
allowances for future use, are the same across all the three different sets of NOx emissions
budgets analyzed.
4.1.2 Regulatory Control Alternatives Analyzed
In accordance with Executive Orders 12866 and 13563, the guidelines of OMB Circular A-
4, and EPA's Guidelines for Preparing Economic Analyses, this RIA analyzes the benefits and
costs associated with complying with the final Revised CSAPR Update. The illustrative Revised
CSAPR Update emission budgets in this RIA represent EGU NOx ozone season emission
budgets for each state that were developed using uniform control stringency represented by
$1,800 per ton of NOx (2016$).2 This RIA analyzes the illustrative Revised CSAPR Update
emission budgets, as well as a more and a less stringent alternative to the final Revised CSAPR
Update. The more and less stringent alternatives differ from the final Revised CSAPR Update in
that they set different NOx ozone season emission budgets for the affected EGUs. The less-
stringent scenario uses emission budgets that were developed using uniform control stringency
represented by $500 per ton of NOx (2016$). The more-stringent scenario uses emission budgets
that were developed using uniform control stringency represented by $9,600 per ton of NOx
(2016$). For details, please see EGU NOx Mitigation Strategies Final Rule TSD, in the docket
for this rule.3
All three scenarios are illustrative in nature, and the budgets included in the Revised
CSAPR Update scenario differ slightly from the budgets finalized in this rule. That is because
subsequent to completion of the analysis of these three scenarios, EPA made minor updates to
budgets. In particular, the modeling presented in the RIA assumes that SNCR optimization is
available in 2022, whereas the final budgets assume that SNCR optimization is available in 2021.
The estimated incremental emission reductions would be 1,163 tons if EPA had used the actual
2021 budgets, or a 1.1% tightening in the modeled budget across the Group 3 states. The choice
2	The budget setting process is described in section VII of the preamble and in detail in the Ozone Transport Policy
Analysis Final Rule Technical Support Document (TSD).
3	Docket ID No. EPA-HQ-OAR-2020-0272
4-2
-------
of 2021 or 2022 for the initial year of SNCR optimization is an exogenous input into the model.
EPA finds that the three illustrative regulatory control alternatives presented in this RIA provide
a reasonable approximation of the impacts of the final rule, as well as an evaluation of the
relative impacts of two regulatory alternatives. This finding is supported by an analysis of the
costs and impacts (but not the benefits) of the final Revised CSAPR Update emission budgets
assuming 2021 for the initial year of SNCR optimization and provided in the docket for this
rulemaking.
Table 4-1 reports the illustrative EGU NOx ozone season emission budgets that are
evaluated in this RIA. As described above, starting in 2021, emissions from affected EGUs in the
12 states cannot exceed the sum of emissions budgets but for the ability to use banked
allowances from previous years for compliance. No further reductions in budgets occur after
2024, and budgets remain in place for future years. Furthermore, emissions from affected EGUs
in a particular state are subject to the CSAPR assurance provisions, which require additional
allowance surrender penalties (a total of 3 allowances per ton of emissions) on emissions that
exceed a state's CSAPR NOx ozone season assurance level, or 121 percent of the emissions
budget. Similar to the approach taken in the CSAPR Update, EPA is implementing a one-time
conversion of banked Group 2 allowances according to a formula. The size of the initial bank
would be set at a level that would ensure that the use of these converted allowances, in addition
to the allowances provided in the states' emissions budgets under the Group 3 trading program,
would not authorize emissions in the trading program region in the first year of the program to
exceed the sum of the states' budgets by more than the sum of the states' variability limits. The
CSAPR NOx ozone season allowance trading program is described in further detail in Section
VII of the preamble.
4-3
-------
Table 4-1. Illustrative NOx Ozone Season Emission Budgets (Tons) Evaluated
	Revised CSAPR Update	
State
2021
2022
2023
2024
2025
Illinois
9,198
9,102
8,179
8,059
8,059
Indiana
13,085
12,582
12,553
9,564
9,564
Kentucky
15,307
14,051
14,051
14,051
14,051
Louisiana
15,389
14,818
14,818
14,818
14,818
Maryland
1,499
1,266
1,266
1,348
1,348
Michigan
12,732
12,290
9,975
9,786
9,786
New Jersey
1,253
1,253
1,253
1,253
1,253
New York
3,416
3,416
3,421
3,403
3,403
Ohio
9,690
9,773
9,773
9,773
9,773
Pennsylvania
8,379
8,373
8,373
8,373
8,373
Virginia
4,614
3,897
3,980
3,663
3,663
West Virginia
13,686
12,884
12,884
12,884
12,884
Total
108,248
103,703
100,525
96,974
96,974
Less-Stringent Alternative
State
2021
2022
2023
2024
2025
Illinois
9,348
9,348
8,393
8,272
8,272
Indiana
15,677
15,206
15,179
12,083
12,083
Kentucky
15,606
15,606
15,606
15,606
15,606
Louisiana
15,430
15,430
15,430
15,430
15,430
Maryland
1,501
1,267
1,267
1,350
1,350
Michigan
13,126
12,688
10,386
10,188
10,188
New Jersey
1,346
1,346
1,346
1,346
1,346
New York
3,463
3,463
3,468
3,450
3,450
Ohio
15,487
15,569
15,569
15,569
15,569
Pennsylvania
11,807
11,806
11,806
11,806
11,806
Virginia
4,661
4,270
4,357
4,021
4,021
West Virginia
15,017
15,017
15,017
15,017
15,017
Total
122,468
121,016
117,822
114,138
114,138
More-Stringent Alternative4
State
2021
2022
2023
2024
2025
Illinois
9,198
9,102
8,179
6,891
6,891
Indiana
13,085
12,582
12,553
8,430
8,430
Kentucky
15,307
14,051
14,051
9,775
9,775
Louisiana
15,389
14,818
14,818
12,622
12,622
Maryland
1,499
1,266
1,266
1,168
1,168
Michigan
12,732
12,290
9,975
7,344
7,344
New Jersey
1,253
1,253
1,253
1,257
1,257
New York
3,416
3,416
3,421
3,297
3,297
Ohio
9,690
9,773
9,773
9,222
9,222
Pennsylvania
8,379
8,373
8,373
7,851
7,851
Virginia
4,614
3,897
3,980
3,184
3,184
West Virginia
13,686
12,884
12,884
10,568
10,568
Total
108,248
103,703
100,525
81,609
81,609
Note that EGUs have flexibility in determining how they will comply with the allowance
trading program. As discussed below, the way that they comply may differ from the methods
4-4
-------
forecast in the modeling for this RIA. See Section 4.3 for further discussion of the modeling
approach used in the analysis presented below.
4.2 Power Sector Modeling Framework
IPM is a state-of-the-art, peer-reviewed, dynamic linear programming model that can be
used to project power sector behavior under future business-as-usual conditions and to examine
prospective air pollution control policies throughout the contiguous United States for the entire
electric power system. EPA used IPM to project likely future electricity market conditions with
and without the Revised CSAPR Update.
IPM, developed by ICF, is a multi-regional, dynamic, deterministic linear programming
model of the contiguous U.S. electric power sector. It provides estimates of least cost capacity
expansion, electricity dispatch, and emissions control strategies while meeting energy demand
and environmental, transmission, dispatch, and reliability constraints.5 EPA has used IPM for
almost three decades to better understand power sector behavior under future business-as-usual
conditions and to evaluate the economic and emissions impacts of prospective environmental
policies. The model is designed to reflect electricity markets as accurately as possible. EPA uses
the best available information from utilities, industry experts, gas and coal market experts,
financial institutions, and government statistics as the basis for the detailed power sector
modeling in IPM. The model documentation provides additional information on the assumptions
discussed here as well as all other model assumptions and inputs.6
The model incorporates a detailed representation of the fossil-fuel supply system that is
used to estimate equilibrium fuel prices. The model uses natural gas fuel supply curves and
regional gas delivery costs (basis differentials) to simulate the fuel price associated with a given
4	For the illustrative purposes in this RIA, EPA's analytical technique for assessing the more stringent alternative
presents emission reduction values incremental to the final rule's stringency prior to 2025. This does not reflect a
determination that new SCR controls could be installed on a fleetwide basis before the 2025 ozone season. See
sections VLB. 1, C. 1, and D. 1 of the preamble for further discussion.
5	Due to the compliance timing for the Revised CSAPR Update, EPA does not allow IPM to build certain new
capital investments such as new, unplanned natural gas or renewable capacity or new SCR or SNCR through 2025 in
response to the state emission budgets. EPA's compliance modeling does allow for new combustion controls, which
represent the most likely potential capital expenditure in the 2022 analysis year.
6	Detailed information and documentation of EPA's Baseline run using IPM (v6), including all the underlying
assumptions, data sources, and architecture parameters can be found on EPA's website at:
http://www.epa.gov/airmarkets/powersectormodeling.html.
4-5
-------
level of gas consumption within the system. These inputs are derived using ICF's Gas Market
Model (GMM), a supply/demand equilibrium model of the North American gas market.7
IPM also endogenously models the partial equilibrium of coal supply and EGU coal
demand levels throughout the contiguous U.S., taking into account assumed non-power sector
demand and imports/exports. IPM reflects 36 coal supply regions, 14 coal grades, and the coal
transport network, which consists of over four thousand linkages representing rail, barge, and
truck and conveyer linkages. The coal supply curves in IPM were developed during a thorough
bottom-up, mine-by-mine approach that depicts the coal choices and associated supply costs that
power plants would face if selecting that coal over the modeling time horizon. The IPM
documentation outlines the methods and data used to quantify the economically recoverable coal
reserves, characterize their cost, and build the 36 coal regions' supply curves.8
To estimate the annualized costs of additional capital investments in the power sector, EPA
uses a conventional and widely accepted approach that applies a capital recovery factor (CRF)
multiplier to capital investments and adds that to the annual incremental operating expenses. The
CRF is derived from estimates of the power sector's cost of capital (i.e., private discount rate),
the amount of insurance coverage required, local property taxes, and the life of capital.9 It is
important to note that there is no single CRF factor applied in the model; rather, the CRF varies
across technologies, book life of the capital investments, and regions in the model in order to
better simulate power sector decision-making.
EPA has used IPM extensively over the past three decades to analyze options for reducing
power sector emissions. Previously, the model has been used to estimate the costs, emission
changes, and power sector impacts for the Clean Air Interstate Rule (U.S. EPA, 2005), the
original Cross-State Air Pollution Rule (U.S. EPA, 2011), the Mercury and Air Toxics Standards
(MATS) (U.S. EPA, 201 la), the Clean Power Plan (CPP) for Existing Power Plants (U.S. EPA,
2015), the Carbon Pollution Standards for New Power Plants (U.S. EPA, 2015), the Affordable
7	See Chapter 8 of EPA's Baseline run using IPM v6 documentation, available at:
https://www.epa.gov/airmarkets/power-sector-modeling-platform-v6-may-2019.
8	See Chapter 7 of the IPM v.6 documentation. The documentation for EPA's Baseline run v.6 using IPM consists of
a comprehensive document for the November 2018 release of IPM v. 6, and incremental update documents for
subsequent releases: http://www.epa.gov/airmarkets/powersectormodeling.html.
9	See Chapter 10 of EPA's Baseline run using IPM (v6) documentation, available at:
http://www.epa.gov/airmarkets/powersectormodeling.html
4-6
-------
Clean Energy Rule (U.S. EPA, 2019), and the Clean Power Plan Repeal (U.S. EPA, 2019). EPA
has also used IPM to estimate the air pollution reductions and power sector impacts of water and
waste regulations affecting EGUs, including Cooling Water Intakes (316(b)) Rule (U.S. EPA,
2014), Disposal of Coal Combustion Residuals from Electric Utilities (CCR) (U.S. EPA, 2015b)
and Steam Electric Effluent Limitation Guidelines (ELG) (U.S. EPA, 2015c).
The model and EPA's input assumptions undergo periodic formal peer review. The
rulemaking process also provides opportunity for expert review and comment by a variety of
stakeholders, including owners and operators of capacity in the electricity sector that is
represented by the model, public interest groups, and other developers of U.S. electricity sector
models. The feedback that the Agency receives provides a highly detailed review of key input
assumptions, model representation, and modeling results. IPM has received extensive review by
energy and environmental modeling experts in a variety of contexts. For example, in October
2014 U.S. EPA commissioned a peer review10 of EPA Baseline run version 5.13 using the
Integrated Planning Model. Additionally, and in the late 1990s, the Science Advisory Board
reviewed IPM as part of the CAA Amendments Section 812 prospective studies11 that are
periodically conducted. The Agency has also used the model in a number of comparative
modeling exercises sponsored by Stanford University's Energy Modeling Forum over the past 15
years. IPM has also been employed by states (e.g., for the Regional Greenhouse Gas Initiative,
the Western Regional Air Partnership, Ozone Transport Assessment Group), other Federal and
state agencies, environmental groups, and industry.
4.3 EPA's Power Sector Modeling of the Baseline Run and Three Regulatory Control
Alternatives
The IPM "baseline run" for any regulatory impact analysis is a business-as-usual scenario
that represents expected behavior in the electricity sector under market and regulatory conditions
in the absence of a regulatory action. As such, an IPM baseline run represents an element of the
baseline for this RIA.12 EPA frequently updates the IPM baseline run to reflect the latest
10	See Response and Peer Review Report EPA Baseline run Version 5.13 Using IPM, available at:
https://www.epa.gov/airmarkets/response-and-peer-review-report-epa-base-case-version-513-using-ipm.
11	http://www2.epa.gov/clean-air-act-overview/benefits-and-costs-clean-air-act
12	As described in Chapter 5 of EPA's Guidelines for Preparing Economic Analyses, the baseline "should
incorporate assumptions about exogenous changes in the economy that may affect relevant benefits and costs (e.g.,
4-7
-------
available electricity demand forecasts from the U.S. Energy Information Administration (EIA) as
well as expected costs and availability of new and existing generating resources, fuels, emission
control technologies, and regulatory requirements.
4.3.1 EPA 's IPM Baseline Run v. 6
For our analysis of the Revised CSAPR Update, EPA used the January 2020 release of
IPM version 6 to provide power sector emissions data for air quality modeling, as well as a
companion updated database of EGU units (the National Electricity Energy Data System, or
NEEDS v.6 rev: 1-8-202013) that is used in EPA's modeling applications of IPM. The IPM
Baseline run includes both the CSAPR and CSAPR Update. EPA's Affordable Clean Energy
(ACE) rule was vacated by the United States Court of Appeals for the District of Columbia
Circuit on January 19, 2021. However, since the analysis for this rulemaking was far along by
that date, the model includes the ACE rule's effects consistent with the RIA for the final ACE
rule. Given the very modest impacts anticipated from the ACE rule on the power sector and
associated emissions levels in the relevant years analyzed for this final rule, EPA does not view
the vacatur of ACE as sufficiently important to warrant separate analysis. Further, since the rule
is included in both the baseline run as well as the three policy alternatives, inclusion of the ACE
rule is not likely to significantly affect cost and benefit estimates for this rulemaking. The
baseline run includes the 2015 Effluent Limitation Guidelines (ELG) and the 2015 Coal
Combustion Residuals (CCR), but does not include the recently finalized 2020 ELG and CCR
rules.14 The analysis of cost and impacts presented in this chapter is based on a single IPM
baseline run, and represents incremental impacts projected solely as a result of compliance with
the emissions budgets presented in Table 4-1 above.
4.3.2. Methodology for Evaluating the Regulatory Control Alternatives
To estimate the costs, benefits, and economic and energy market impacts of the Revised
CSAPR Update, EPA conducted quantitative analysis of the three regulatory control alternatives:
changes in demographics, economic activity, consumer preferences, and technology), industry compliance rates,
other regulations promulgated by EPA or other government entities, and behavioral responses to the rule by firms
and the public." (USEPA, 2010).
13	https://www.epa.gov/airmarkets/national-electric-energy-data-system-needs-v6
14	For a full list of modeled policy parameters, please see:
https://www.epa.gov/sites/production/files/2020-
02/documents/incremental_documentation_for_epa_v6 January _2020_reference_case.pdf
4-8
-------
the final Revised CSAPR Update emission budgets and a more and a less stringent alternative.
Details about these regulatory control alternatives, including state-specific EGUNOx ozone-
season emissions budgets for each alternative as analyzed in this RIA, are provided above in
Section 4.1.
Before undertaking power sector analysis to evaluate compliance with the regulatory
control alternatives, EPA first considered available EGUNOx mitigation strategies that could be
implemented for the upcoming ozone season (i.e., the 2021 ozone season). EPA considered all
widely-used EGU NOx control strategies: optimizing NOx removal by existing, operational
selective catalytic reduction (SCRs) and turning on and optimizing existing idled SCRs;
optimizing existing idled selective non-catalytic reduction (SNCRs); installation of (or upgrading
to) state-of-the-art NOx combustion controls; shifting generation to units with lower NOx
emission rates; and installing new SCRs and SNCRs. EPA determined that affected EGUs within
the 12 states could implement all of these NOx mitigation strategies, except installation of new
SCRs or SNCRs and state-of-the-art combustion controls, for the 2021 ozone season.15 After
assessing the available NOx mitigation methods for complying with the annual budgets, this RIA
projects that the system-wide least-cost strategies for compliance with the illustrative Revised
CSAPR Update budgets presented above and the more and less stringent regulatory alternatives
lead to the application of the same controls at the same sources as in the analysis used to
calculate the budgets for these alternatives. As a consequence, the sectoral analyses used to
establish the budgets are the same analyses used to estimate the compliance cost, benefits, and
impacts of the Revised CSAPR Update and the more stringent and less stringent alternatives. In
the analysis of the rule presented in this RIA, in each year of the analysis period (2021-2040) and
in each of the 12 states subject to tighter seasonal NOx budgets, seasonal NOx emissions from
the sources subject to the rule equal the seasonal NOx budget. For more details on these
assessments, including the assessment of EGU NOx mitigation costs and feasibility, please refer
to the EGU NOx Mitigation Strategies Final Rule TSD, in the docket for this rule.16
15	The modeling presented in this RIA assumes that SNCR optimization is available starting in 2022. This choice
reflects an exogenous input into the model.
16	Docket ID No. EPA-HQ-OAR-2020-0272
4-9
-------
These mitigation strategies are primarily captured within the model. However, due to
limitations on model size, IPM v.6 does not have the ability to endogenously determine whether
or not to operate existing EGU post-combustion NOx controls (i.e., SCR or SNCR) in response
to a regulatory emissions requirement.17 The operating status of existing post-combustion NOx
controls at a particular EGU in a model scenario is determined by the model user. In order to
evaluate compliance with the regulatory alternatives, EPA determined outside of IPM whether or
not operation of existing controls that are idle in the baseline would be reasonably expected for
compliance with each of the evaluated regulatory alternatives and for which model years they
can feasibly be applied. IPM includes optimization and perfect foresight in solving for least cost
dispatch. Given that the final rule will become effective soon after the start of the 2021 ozone
season, to avoid overstating optimization and dispatch decisions that are not possible in that short
time frame, EPA complemented the projected IPM EGU outlook with historical (e.g.,
engineering analytics) perspective based on historical data that only factors in known changes to
the fleet. This analysis forms the basis for the benefits calculations presented in this RIA.
EPA considers a unit to have optimized use of an SCR if emissions rates are equal to (or
below) the "widely achievable" rate of 0.08 lbs/MMBtu.18 Within IPM, units with extant SCRs
are defined as SCR-equipped units with ozone season NOx emission rates less than 0.20 lbs/
MMBtu in the baseline run. These units had their emission rates lowered to the lower of their
mode 419 NOx rate in NEEDS and the "widely achievable" optimized emissions rate of 0.08 lbs/
MMBtu in the final Revised CSAPR Update. Units equipped with SCRs with an emissions rate
exceeding 0.20 lbs/ MMBtu were considered to have idled SCRs. These units had their emission
rates lowered to the lower of their mode 4 NOx rate in NEEDS and the "widely achievable"
optimized emissions rate of 0.08 lbs/ MMBtu in the Revised final Revised CSAPR Update.20
These control options (optimizing partially operating SCR controls or turning on idled SCR
17	EGUs with idled SCR or SNCR in the baseline run represent a small percentage (less than 10 percent) of the EGU
fleet that is equipped with NOx post-combustion controls.
18	For details on the derivation of this standard, please see preamble Section VII.B. 1.
19	NEEDS includes four possible states of NOx control operations, designated Modes 1-4. For details, please see
Chapter 3.9.3 of IPM v6 documentation available at:
https://www.epa.gov/sites/production/files/2018-08/documents/epa_platform_v6_documentation_-
_all_chapters_august_23_2018_updated_table_6-2.pdf.
20	EPA updated the total emission reduction potential for each technology based on information provided by
commenters where EPA determined that information to be credible. Further details are provided in the Response to
Comment document included in the docket and in the Ozone Transport Policy Analysis Final Rule TSD.
4-10
-------
controls) are achievable in 2021 and were associated with a uniform control cost of $800 per ton
and $1,600 per ton respectively. No further adjustments were made to the variable and fixed
operating cost of these units, and their heat rates were also not adjusted to reflect energy
requirements from increasing SCR removal efficiency within IPM. Under the proposed rule, 47
units are projected to fully run existing SCR controls, while 4 units are projected to turn on idled
SCR controls.
EPA considers a unit to have optimized use of an SNCR if NOx emissions rates are equal
to or less than the mode 2 rate from the NEEDS database (June 2020). As described in EPA's
power sector IPM Modeling Documentation (Chapter 3), these unit-specific NOx mode rates are
calculated from historical data and reflect operation of existing post-combustion controls. Mode
2 for SNCR-controlled coal units is intended to reflect the operation of that unit's post
combustion control based on prior years when that unit operated its control. Hence any units with
existing SNCRs with NOx emission rates greater than their mode 2 rates in the 12-state region
had their rates lowered to their mode 2 rates. The illustrative budgets assume these control
options as being achievable in 2022 and were associated with a uniform control cost of $1,800
per ton.21 No further adjustments were made to the variable and fixed operating cost of these
units, and their heat rates were also not adjusted to reflect energy requirements from increasing
SNCR removal efficiency within IPM. Units with potential SNCR optimization-based emission
reductions and their corresponding emissions, heat input, and emission rate are shown in
Appendix A of the Ozone Transport Policy Analysis Final Rule TSD.
Finally, unit combustion control configurations listed in NEEDS were compared against
Table 3-11 in the Documentation for EPA Baseline run v.5.13 Using the Integrated Planning
Model IPM v.6, which lists state-of-the-art combustion control configurations based on unit
firing type. This allowed EPA to identify units that would receive state-of-the-art combustion
control upgrades in IPM. EPA then followed the procedure in the EGU NOx Mitigation
Strategies Final Rule TSD to calculate each of these unit's new NOx emission rate. These
upgrades were assumed to occur in 2022 and were assigned a uniform control cost of $1,600 per
21 EPA notes that the final rule implements state emissions budgets reflecting SNCR optimization starting in 2021.
The Agency conducted additional sensitivity analysis using IPM demonstrating that the 2021 implementation made
no significant difference in the cost implications described in the body of the RIA. EPA provides that IPM scenario
in the docket for this rulemaking.
4-11
-------
ton. No further adjustments were made to the variable and fixed operating cost of these units, and
their heat rates were also not adjusted to reflect increased energy input requirements at a given
load from the use of additional combustion controls, within IPM. Under the final Revised
CSAPR Update, 10 units are projected to install state-of-the-art combustion controls.
The EGU NOx mitigation strategies that are assumed to operate or are available to reduce
NOx in response to each of the regulatory control alternatives are shown in Table 4-2; more
information about the estimated costs of these controls can be found in the EGU NOx Mitigation
Strategies Final Rule TSD.
Table 4-2. NOx Mitigation Strategies Represented in Modeling of the Regulatory Control
	Alternatives	
Regulatory Control i\qx Controls Implemented
	Alternative	_	
Less Stringent Alternative	(1) Shift generation to minimize costs (costs estimated within IPM)	
(All controls above)
(2)	Fully operating existing SCRs to achieve 0.08 lb/MMBtu NOx emission rate
RevkeH CSAPR TTnHste	(C0StS estimated outside IPM)
(3)	Turn on idled SCRs (costs estimated outside IPM) and fully operate akin to
(2)
(4)	Fully operate existing SNCRs (costs estimated outside IPM)
	(5) Install state-of-the-art combustion controls.	
(All controls above)
(6) In 2025, impose state emission limits commensurate with installation of new
SCRs on units without such controls. However, additional SCRs controls are
not a least-cost compliance strategy, (costs estimated within IPM)	
More Stringent Alternative
For the NOx controls identified in Table 4-2, under the final rule and the more stringent
alternative, 47 units, not already doing so in 2019, are projected to fully operate existing SCRs
and 4 units are projected to turn on idled SCRs. Under the less stringent alternative, no units are
projected to either fully operate existing SCRs or turn on idled SCRs. Under the final rule and
the more stringent alternative, 29 units are projected to fully operate existing SNCRs, and under
the less stringent alternative no units are projected to fully operate SNCR controls. Under the
final rule and the more stringent alternative, 10 units are projected to install state-of-the-art
combustion controls, and under the less stringent alternative no units are projected to install
state-of-the-art combustion controls. The book-life of the controls is assumed to be 15 years.
4-12
-------
Under the final rule, the more stringent alternative, and the less stringent alternative, no units are
projected to install new SCRs.22 The book-life of the new SCRs is assumed to be 15 years.
In addition to the limitation on ozone season NOx emissions required by the EGU
emissions budgets for the 12 states, there are four important features of the allowance trading
program represented in the model that may influence the level and location of NOx emissions
from affected EGUs, including: the ability of affected EGUs to buy and sell NOx ozone season
allowances from one another for compliance purposes; the ability of affected EGUs to bank NOx
ozone season allowances for future use; the effect of limits on the total ozone season NOx
emissions from affected EGUs in each state required by the assurance provisions; and the
treatment of banked pre-2021 vintage NOx ozone season allowances issued under the CSAPR
Update program now being revised under this final rule. Each of these features of the ozone
season allowance trading program is described below.
Affected EGUs are expected to choose the least-cost method of complying with the
requirements of the allowance trading program, and the distribution of ozone season NOx
emissions across affected EGUs is generally governed by this cost-minimizing behavior in the
analysis. The total ozone season NOx emissions from affected EGUs in this analysis are limited
to the amount allowed by the sum of the NOx budgets across the 12 states. Furthermore,
allowances may be banked for future use. The number of banked allowances is influenced by the
determination, outside the model, of whether (i) existing controls that are idle in the baseline run
are turned on and (ii) it is less costly to abate ozone season NOx emissions in a current ozone
season than to abate emissions in a later ozone season. Affected EGUs are expected to bank NOx
ozone season allowances in the 2021 ozone season for use in a later ozone season. The model
starts with an assumed bank level in 2021 and endogenously determines the bank in each
subsequent year. Based on observation, EPA believes that this is a reasonable compliance path
for EGUs, even though there may be other non-economic reasons, such as being prepared for
future variability in power sector operations, that can potentially influence this decision.
22 Under the proposed rule, units were exogenously forced to install SCR controls in IPM. In the modeling for the
final rule, the choice to install SCR controls was endogenous to the model, and no incremental SCR installations
occurred, with the model relying on greater levels of generation shifting instead.
4-13
-------
While there are no explicit limits on the exchange of allowances between affected EGUs
and on the banking of 2021 and future-year vintage NOx ozone season allowances, the assurance
provisions limit the amount of seasonal NOx emissions by affected EGUs in each of the 12
states. The assurance level limits affected EGU emissions over an ozone season to the state's
NOx ozone season emissions budget plus an increment equal to 21 percent of each state's
emissions budget. This increment is called the variability limit. See Section VII.C.4 of the
preamble for a discussion of the purpose of the assurance provision and further detail about how
the variability limits and assurance levels are determined. If a state exceeds its assurance level in
a given year, sources within that state are assessed a 3-to-l allowance surrender penalty on the
excess tons. Section VII.C.4 of the preamble also explains how EPA then determines which
EGUs are subject to this surrender requirement. In the modeling, the assurance provisions are
represented by a limit on the total ozone season NOx emissions that may be emitted by affected
EGUs in each state, and thus the modeling does not permit affected EGUs to emit beyond the
assurance levels and thus incur penalties.
As described in Section VII.D.4 of the preamble, the rule allows pre-2021 vintage NOx
ozone season allowances (that had been issued under the CSAPR Update program now being
revised under this rule) to be used for compliance with this rule, following a one-time conversion
that reduces the overall quantity of banked allowances from that time period. Based on EPA's
expectation of the size of the NOx allowance bank after the one-time conversion carried out
pursuant to the terms of this rule, the treatment of these banked allowances is represented in the
modeling as an additional 22,737 tons of NOx allowances, the equivalent of one year of the
variability limit associated with the emission budgets, that may be used by affected EGUs during
the 2021 ozone season or in later ozone seasons under the final Revised CSAPR Update. Under
the more stringent and less stringent alternatives an additional 22,732 tons and 25,718 tons
respectively may be used by affected EGUs during the 2021 ozone season or in later ozone
seasons.
4.3.3 Methodology for Estimating Compliance Costs
This section describes EPA's approach to quantify estimated compliance costs associated
with the three illustrative regulatory control alternatives. These compliance costs include
estimates projected directly by the model as well as calculations performed outside of the model
4-14
-------
that use IPM model inputs and methods. The model projections capture the costs associated with
shifting generation to lower-NOx emitting EGUs. The costs of increasing the use and optimizing
the performance of existing and operating SCRs and SNCRs,23 and for installing or upgrading
NOx combustion controls, were estimated outside of the model. The costs for these three NOx
mitigation strategies are calculated based on engineering analytics emissions projections and use
the same NOx control cost equations used in IPM. Therefore, this estimate is consistent with
modeled projections and provides the best available quantification of the costs of these NOx
mitigation strategies.
The following steps summarize EPA's methodology for estimating the component of
compliance costs that are calculated outside of the model for the final rule alternative in 202124:
(1)	In the model projections, identify all EGUs in the 12 states that can adopt the following
NOx mitigation strategies:
•	Fully operating existing SCRs
•	Fully operating existing SNCRs
•	Installing state-of-the-art combustion controls
(2)	Estimate the total NOx reductions that are attributable to each of these strategies:25
•	Fully operating existing SCRs (SCRs operating in baseline run): 7,447 tons
•	Fully operating existing SCRs (SCRs not operating in baseline run): 5,870 tons
•	Fully operating existing SNCRs (not available in 2021 in the illustrative
scenarios modeled): 0 tons26
•	Installing state-of-the-art combustion controls (not available in 2021): 0 tons
23	This includes optimizing the performance of SCRs that were not operating.
24	For more information on the derivation of costs and useful life of combustion controls, please see EGU NOx
Mitigation Strategies Final Rule TSD.
25	For more information on how NOx reductions were attributed to strategies, see the Ozone Transport Policy
Analysis Final Rule TSD.
26	The estimated emission reductions would be 1,163 tons if EPA had used the actual 2021 budgets. (The selection
of 2022 rather than 2021 was an exogenous input to the model.)
4-15
-------
(3)	Estimate the average cost associated with each of these strategies:27
•	Fully operating existing SCRs (SCRs operating in baseline run): $800/ton
•	Fully operating existing SCRs (SCRs not operating in baseline run): $l,600/ton
•	Fully operating existing SNCRs: $l,800/ton28
•	Installing state-of-the-art combustion controls: $l,600/ton
(4)	Multiply (2) by (3) to estimate the total cost associated with each of these strategies.
Table 4-3 summarizes the results of this methodology for the final rule alternative in 2021.
Table 4-3. Summary of Methodology for Calculating Compliance Costs Estimated Outside
of IPM for Revised CSAPR Update Final Rule, 2021 (2016$)

NOx Ozone Season



Emissions
Average Cost
Total Cost
NOx Mitigation Strategy
(tons)
($/ton)
($MM)
Optimize existing SCRs
7,447
$800
$6
Operate existing SCRs
5,870
$1,600
$9
EPA exogenously updated the emissions rates for the identified EGUs within the 12 states
consistent with the set of controls determined for 2021-2025 within IPM. The model was
updated to incorporate the emissions budgets identified for each case, and the first-year bank
adjustment as outlined in Section 4.3.2. The Group 2 regional trading program was updated to
exclude the 12-state Group 3 regional trading program, and budgets for the remaining Group 2
states were left otherwise unchanged. The change in the reported power system production cost
between this model run and the baseline run was used to capture the cost of generation shifting.
The total costs of compliance with the regulatory control alternatives are estimated as the sum of
the costs that are modeled within IPM and the costs that are calculated outside the model.
27	See EGU NOx Mitigation Strategies Final Rule TSD for derivation of cost-per-ton estimates for fully operating
SCRs and upgrading to state-of-the-art combustion controls.
28	If SNCRs were fully operated in 2021 an additional $2 million (2016$) cost would be incurred.
4-16
-------
4.4 Estimated Impacts of the Regulatory Control Alternatives
4.4.1 Emission Reduction Assessment
As discussed in Chapter 1, EPA determined that NOx emissions in 12 eastern states affect
the ability of downwind states to attain and maintain the 2008 ozone NAAQS. For these 12
eastern states, EPA is issuing Federal Implementation Plans (FIPs) that update the existing
CSAPR Update NOx ozone-season emission budgets for EGUs and implement these budgets via
the CSAPR NOx ozone-season allowance trading program.
As indicated in Chapter 1, the NOx emissions reductions are presented in this RIA from
2021 through 2040. The 2021 emissions estimates are based on IPM projections for 2021, and
adjustments to account for historical data. For more information on these and other adjustments,
see the Ozone Transport Policy Analysis Final Rule TSD.
Table 4-4 presents the estimated reduction in power sector NOx emissions resulting from
compliance with the evaluated regulatory control alternatives (i.e., emissions budgets) in the 12
states, as well as the impact on other states. The emission reductions follow an expected pattern:
the less stringent alternative produces substantially smaller emissions reductions than the final
rule emissions budgets, and the more stringent alternative results in slightly more NOx emissions
reductions.
4-17
-------
Table 4-4. EGU Ozone Season NOx Emissions and Emissions Changes (thousand tons) for
	the Baseline Run and the Regulatory Control Alternatives	
Total Emissions
Ozone Season NOx
(thousand tons)
Baselin
e Run
Revised
CSAPR
Update
Less-
Stringent
Alternative
More-
Stringent
Alternative
Revised
CSAPR
Update
Less-
Stringent
Alternative
More-
Stringent
Alternative

12 States
124
108
122
108
-16
-2
-16
2021
Other States
236
236
236
236
0
0
0

Total
360
344
358
344
-16
-2
-16

12 States
123
104
121
104
-19
-2
-19
2022
Other States
233
233
233
233
0
0
0

Total
355
337
354
337
-19
-2
-19

12 States
119
101
118
101
-19
-2
-19
2023
Other States
222
222
222
222
0
0
0

Total
341
322
340
322
-19
-2
-19

12 States
116
97
114
82
-19
-2
-34
2024
Other States
218
218
218
218
0
0
0

Total
334
315
332
300
-19
-2
-34

12 States
116
97
114
82
-19
-2
-34
2025
Other States
218
218
218
218
0
0
0

Total
334
315
332
300
-19
-2
-34

12 States
91
78
89
66
-13
-2
-25
2030
Other States
207
207
207
207
0
0
0

Total
298
285
297
274
-13
-2
-25

12 States
90
77
89
66
-13
-2
-24
2035
Other States
207
207
207
207
0
0
0

Total
298
285
296
273
-13
-2
-24

12 States
90
77
88
66
-13
-2
-24
2040
Other States
207
207
207
207
0
0
0

Total
298
285
296
274
-13
-2
-24
Change from Baseline Run
The results of EPA's analysis show that, with respect to compliance with the EGU NOx
emission budgets in 2021, maximizing the use of existing operating SCRs provides the largest
amount of ozone season NOx emission reductions (47 percent, affecting 47 units), and turning on
idled SCRs produces an additional 37 percent (affecting 4 units) of the total ozone season NOx
reductions. Generation shifting primarily from coal to gas generation (16 percent) makes up the
remainder of the ozone season NOx reductions. Based on this analysis of how EGUs are
expected to comply with the final Revised CSAPR Update combining IPM model projections
4-18
-------
with historical (e.g., engineering analytics) perspective based on historical data that only factors
in known changes to the fleet, none of the Group 3 states are projected to hit their variability
limits, nor bank or withdraw a substantial additional number of allowances above the starting
bank during the analysis period (2021-2025).29
If EPA were to have included SNCR optimization in 2021, with respect to compliance
with the EGU NOx emission budgets in 2021, maximizing the use of existing operating SCRs
provides the largest amount of ozone season NOx emission reductions (44 percent, affecting 47
units), turning on idled SCRs produces an additional 35 percent (affecting 4 units) of the total
ozone season NOx reductions, and fully operating existing SNCRs produces an additional 6
percent of reductions (affecting 29 units). Generation shifting primarily from coal to gas
generation (15 percent) makes up the remainder of the ozone season NOx reductions.
In addition to the ozone season NOx reductions, there will also be reductions of other air
emissions associated with EGUs burning fossil fuels (i.e., co-pollutants). These other emissions
include the annual total changes in emissions of NOx and CO2 as well as partially-estimated
changes in emissions for SO2 and direct PM2.5 emissions. EPA relied on Engineering Analysis to
account for changes in NOx (annual and ozone season), SO2, and direct PM. While this approach
captures the impact of generation shifting for NOx emissions, it does not fully capture the impact
of generation shifting for SO2 and direct PM in complying with the Revised CSAPR Update
budgets. EPA did not analyze changes in emissions due to generation shifting for either SO2 or
PM2.5 because the majority of the reductions associated with this rulemaking are tied to the
operation of SCR and SNCR controls (which do not affect SO2 emission rates) and state of the
art combustion controls (which have a minimal impact on SO2 emission rates). Additionally in
order to meet the court-ordered timeline for this rulemaking EPA prioritized fully capturing the
impact of reductions from generation shifting on NOx and CO2, but did not account for the
relatively small amount of SO2 and primary PM emissions reductions that would likely occur due
to generation shifting. Hence total benefits could be higher than those reported in this RIA. EPA
relied on IPM estimates to capture changes in CO2 emissions, which fully account for the impact
29 As shown in Table 4-4, in 2021 and 2025 seasonal NOx emissions from affected EGUs in the Group 3 states are
projected to emit at levels equal to the seasonal budget, and therefore (i) will not bank additional allowances, or (ii)
on net, use any banked allowances available at the end of the previous year or, in the case of 2021, from the starting
bank.
4-19
-------
of generation shifting. The emissions reductions are presented in Table 4-5. Consistent with the
limited impact of generation shifting, there were de minimis emissions changes of CO, mercury,
and HC1.
4-20
-------
Table 4-5. EGU Annual Emissions and Emissions Changes for NOx, SO2, PM2.5, and CO2 for the Regulatory Control
Alternatives



Total Emissions

Change from Baseline Run
Annual NOx
(thousand tons)
Baseline
Run
Revised
CSAPR
Update
Less-
Stringent
Alternative
More-
Stringent
Alternative
Revised
CSAPR
Update
Less-
Stringent
Alternative
More-Stringent
Alternative

12 States
291
275
289
275
-16
-2

-16
2021
Other States
524
524
524
524
0
0

0

Total
815
799
814
799
-16
-2

-16

12 States
287
265
285
265
-22
-2

-22
2022
Other States
517
517
517
517
0
0

0

Total
804
782
802
782
-22
-2

-22

12 States
280
258
278
258
-22
-2

-22
2023
Other States
490
490
490
490
0
0

0

Total
769
748
768
748
-22
-2

-22

12 States
271
249
269
234
-21
-2

-37
2024
Other States
482
482
482
482
0
0

0

Total
753
731
751
716
-21
-2

-37

12 States
271
249
269
234
-21
-2

-37
2025
Other States
482
482
482
482
0
0

0

Total
753
731
751
716
-21
-2

-37

12 States
209
193
207
182
-16
-2

-27
2030
Other States
458
458
458
458
0
0

0

Total
666
651
665
639
-16
-2

-27

12 States
207
193
206
181
-15
-2

-26
2035
Other States
458
458
458
458
0
0

0

Total
665
650
663
639
-15
-2

-26

12 States
207
193
205
182
-14
-2

-25
2040
Other States
458
458
458
458
0
0

0

Total
665
651
663
640
-14
-2

-25
4-21
-------



Total Emissions

Change from Baseline Run
Annual SO2*
(thousand tons)
Baseline
Run
Revised
CSAPR
Update
Less-
Stringent
Alternative
More-
Stringent
Alternative
Revised
CSAPR
Update
Less-
Stringent
Alternative
More-Stringent
Alternative

12 States
374
374
374
374
0
0
0
2021
Other States
556
556
556
556
0
0
0

Total
930
930
930
930
0
0
0

12 States
371
371
371
371
0
0
0
2022
Other States
545
545
545
545
0
0
0

Total
916
916
916
916
0
0
0

12 States
347
347
347
347
0
0
0
2023
Other States
528
528
528
528
0
0
0

Total
874
874
874
874
0
0
0

12 States
340
340
340
340
0
0
0
2024
Other States
518
518
518
518
0
0
0

Total
858
858
858
858
0
0
0

12 States
340
340
340
340
0
0
0
2025
Other States
518
518
518
518
0
0
0

Total
858
858
858
858
0
0
0

12 States
234
234
234
234
0
0
0
2030
Other States
503
503
503
503
0
0
0

Total
737
737
737
737
0
0
0

12 States
233
233
233
233
0
0
0
2035
Other States
503
503
503
503
0
0
0

Total
736
736
736
736
0
0
0

12 States
231
231
231
231
0
0
0
2040
Other States
503
503
503
503
0
0
0

Total
734
734
734
734
0
0
0
4-22
-------



Total Emissions

Change from Baseline Run
Annual Direct PM2.5*
(thousand tons)
Baseline
Run
Revised
CSAPR
Update
Less-
Stringent
Alternative
More-
Stringent
Alternative
Revised
CSAPR
Update
Less-
Stringent More-Stringent
Alternative Alternative

12 States
50
50
50
50
0
0 0
2021
Other States
76
76
76
76
0
0 0

Total
126
126
126
126
0
0 0

12 States
50
50
50
50
0
0 0
2022
Other States
76
76
76
76
0
0 0

Total
126
126
126
126
0
0 0

12 States
50
50
50
50
0
0 0
2023
Other States
73
73
73
73
0
0 0

Total
122
122
122
122
0
0 0

12 States
48
48
48
48
0
0 0
2024
Other States
72
72
72
72
0
0 0

Total
120
120
120
120
0
0 0

12 States
48
48
48
48
0
0 0
2025
Other States
72
72
72
72
0
0 0

Total
120
120
120
120
0
0 0

12 States
40
40
40
40
0
0 0
2030
Other States
70
70
70
70
0
0 0

Total
110
110
110
110
0
0 0

12 States
41
41
41
41
0
0 0
2035
Other States
70
70
70
70
0
0 0

Total
111
111
111
111
0
0 0

12 States
42
42
42
42
0
0 0
2040
Other States
70
70
70
70
0
0 0

Total
112
112
112
112
0
0 0
4-23
-------



Total Emissions

Change from Baseline Run
Annual CO2

Revised
Less-
More-
Revised
Less-


(million tons)
Baseline
CSAPR
Stringent
Stringent
CSAPR
Stringent
More-Stringent


Run
Update
Alternative
Alternative
Update
Alternative
Alternative

12 States
478
478
478
478
0
0

0
2021
Other States
959
959
959
959
0
0

0

Total
1,437
1,437
1,437
1,437
0
0

0

12 States
507
505
505
503
-3
-2

-4
2022
Other States
985
986
985
986
0
0

0

Total
1,493
1,490
1,491
1,489
-2
-2

-4

12 States
537
532
533
528
-5
-4

-9
2023
Other States
1,011
1,012
1,011
1,012
0
0

1

Total
1,548
1,543
1,544
1,540
-5
-4

-8

12 States
532
527
528
520
-5
-4

-11
2024
Other States
1,004
1,004
1,004
1,005
0
0

1

Total
1,536
1,530
1,531
1,525
-5
-4

-10

12 States
526
521
523
512
-5
-4

-14
2025
Other States
996
996
996
998
0
0

2

Total
1,523
1,518
1,519
1,511
-5
-4

-12

12 States
534
526
528
515
-8
-6

-19
2030
Other States
920
922
921
925
2
2

5

Total
1,454
1,448
1,450
1,440
-5
-4

-14

12 States
504
500
502
492
-4
-3

-13
2035
Other States
903
903
903
903
0
0

1

Total
1,407
1,403
1,404
1,395
-4
-3

-12

12 States
512
507
508
496
-4
-3

-15
2040
Other States
928
928
929
931
0
0

2

Total
1,440
1,436
1,437
1,427
-4
-3

-13
4-24
-------
* There are no annual SO2 and PM2 5 emissions reductions that come from turning on SCRs and SNCRs assuming that nothing else changes, but EPA did not
analyze the effects on SO2 and direct PM that may come from shifting power generation, for example from coal-fired power plants to gas-fired or other types of
power plants. EPA does expect some changes in SO2 and PM2 5 emissions due to shifting of power generation.
4-25
-------
4.4.2	Impact of Emissions Reductions on Maintenance and Nonattainment Monitors
In 2021, there are two nonattainment receptors and two maintenance receptors (see section
V.C of the preamble for additional discussion). EPA evaluated the air quality improvements at
the four receptors for the three EGU emission control technologies that are available in the near-
term and that comprise the selected control stringency of the final rule. EPA determined that the
average air quality improvement at the four receptors relative to the engineering analytics
baseline run was 0.165 ppb for optimization of existing SCRs and LNB upgrades, and 0.17 ppb
when also including optimization of existing SNCRs (see Table VI.D.1-1 in the preamble for
additional discussion). EPA found that one of the receptors (Westport, Connecticut receptor)
remains nonattainment across these control stringencies in 2021, another receptor (Stratford,
Connecticut receptor) switches from nonattainment to maintenance with the optimization of
existing SCRs (i.e., its average design value (DV)1 falls below the standard but its maximum
DV remains above the NAAQS), while a third receptor (Houston receptor) remains maintenance
across these control stringencies.2
EPA observes these control stringencies result in all downwind air quality problems for the
2008 ozone NAAQS being resolved after 2024 (one year earlier than the baseline run). There are
also projected changes in receptor status (from projected nonattainment to maintenance-only) for
the Stratford and Westport receptors (the first in 2021, the second in 2024). In addition, the
Houston receptor changes from maintenance to attainment in 2023.
4.4.3	Compliance Cost Assessment
The estimates of the changes in the cost of supplying electricity for the regulatory control
alternatives are presented in Table 4-6. Since the rule does not result in any additional
recordkeeping, monitoring or reporting requirements, the costs associated with compliance,
monitoring, recordkeeping, and reporting requirements are not included within the estimates in
this table and can be found in preamble Section VII.C.6.
1	The DV is calculated as the 3-year average of the annual 4th highest daily maximum 8-hour ozone concentration in
parts per billion, with decimals truncated. The D V is a metric compared to the standard level to determine whether a
monitor is violating the NAAQS.
2	The fourth receptor was clean in the engineering baseline run, which is the starting point for a Step 3 analysis.
4-26
-------
Table 4-6. National Compliance Cost Estimates (millions of 2016$) for the Regulatory
Control Alternatives

Revised CSAPR
Update
More-Stringent
Alternative
Less-Stringent
Alternative
2021-2025 (Annualized)
$10.0
$41.4
$(2.9)
2021-2040 (Annualized)
$24.8
$28.5
$19.6
2021 (Annual)
$5.1
$5.2
$1.6
2022 (Annual)
$19.2
$61.5
$5.9
2023 (Annual)
$19.2
$61.5
$5.9
2024 (Annual)
$2.1
$4.5
$(14.9)
2025 (Annual)
$1.6
$4.0
$(14.9)
2030 (Annual)
$63.6
$32.3
$67.0
2035 (Annual)
$18.2
$41.2
$14.3
2040 (Annual)
$8.8
$134.0
$ 18.9
"2021-2025 (Annualized)" reflects total estimated annual compliance costs levelized over the period 2021 through
2025 and discounted using a 4.25 real discount rate.3 This does not include compliance costs beyond 2025. "2021-
2040 (Annualized)" reflects total estimated annual compliance costs levelized over the period 2021 through 2040
and discounted using a 4.25 real discount rate. This does not include compliance costs beyond 2040. "2021
(Annual)" through "2040 (Annual)" costs reflect annual estimates in each of those years.4
There are several notable aspects of the results presented in Table 4-6. The most notable
result in Table 4-6 is that the estimated annual compliance costs for the less stringent alternative
is negative (i.e., a cost reduction) in 2024 and 2025, although this regulatory control alternative
reduces NOx emissions by 2,000 tons as shown in Table 4-5. While seemingly counterintuitive,
estimating negative compliance costs in a single year is possible given the assumption of perfect
foresight. IPM's objective function is to minimize the discounted net present value (NPV) of a
stream of annual total cost of generation over a multi-decadal time period.5 For example, with
the assumption of perfect foresight it is possible that on a national basis within the model the
least-cost compliance strategy may be to delay a new investment or retirement that was projected
to occur sooner in the baseline run. Such a delay could result in a lowering of annual cost in an
3	This table reports compliance costs consistent with expected electricity sector economic conditions. An NPV of
costs was calculated using a 4.25% real discount rate consistent with the rate used in IPM's objective function for
cost-minimization. The NPV of costs was then used to calculate the levelized annual value over a 5-year period
(2021-2025) and a 20-year period (2021-2040) using the 4.25% rate as well. Tables ES-14 and 7-4 report the NPV
of the annual stream of costs from 2021-2040 using 3% and 7% consistent with OMB guidance.
4	Cost estimates include financing charges on capital expenditures that would reflect a transfer and would not
typically be considered part of total social costs. Exclusion of these costs would have a minimal impact on the
results.
5	For more information, please see Chapter 2 of the IPM documentation.
4-27
-------
early time period and increase it in later time periods. Since the less-stringent alternative is
designed to include only generation shifting, it does not necessitate full operation of existing
controls, nor installation of new controls, leading to a negative total cost point estimate in 2025,
reflecting the decision to delay retirements until later in the forecast period. Under the final rule,
fully operating existing SCR controls provides a large share of the total emissions reductions.
These options are selected in 2021, while upgrading to state-of-the-art combustion controls and
fully operating existing SNCRs are assumed to begin in 2022. Generation shifting costs are
positive in 2021 and 2023, but negative in 2025. The result is that the costs in 2021-23 are higher
than costs in 2025. Projected costs for the illustrative Revised CSAPR Update peak in 2030 at
$63.6 million (2016$) and annualized costs for the 2021-40 period are $24.8 million (2016$).
These trends are also consistent with IPM base case projections: total costs increase most rapidly
between the 2025 and 2030 run years as the system responds to projected demand growth and
baseload capacity retirements by adding capacity, shifting the generation mix and altering inter-
regional transmission flows. Hence under the Revised CSAPR Update rule and the less stringent
alternative, incremental costs peak in 2030, driven by IPM projected generation shifting costs.
Under the more stringent scenario the tighter cap modeled results in higher costs towards the end
of the analysis period as well, as the system alters generation and build patterns to accommodate
the tighter modeled budgets. To put these costs into context, the incremental 2030 projected cost
constitutes 0.04 percent of total projected baseline run system production costs. For comparison,
the compliance cost for the Final CSAPR Update was estimated at $154 million (2016$) in 2020.
Under the more stringent alternative, while 2021 includes the same set of controls as under
the Revised CSAPR Update, a wider range of technologies is considered in subsequent years.
This, combined with a more stringent cap driving generation shifting costs positive in every year,
results in costs that grow over the 2021-25 period.
As part of the IPM model runs, the Group 2 regional trading program was updated to
exclude the 12-state Group 3 regional trading program, and budgets for the remaining Group 2
states were left otherwise unchanged. The Group 2 states did not exhibit significant changes in
projected allowance prices and level and location of Group 2 NOx emissions between the
baseline and regulatory alternatives as a result of this update.
4-28
-------
In addition to evaluating annual compliance cost impacts, EPA believes that a full
understanding of these three regulatory control alternatives benefits from an evaluation of
annualized costs over the 2021-2025 timeframe. Starting with the estimated annual cost time
series, it is possible to estimate the net present value of that stream, and then estimate a levelized
annual cost associated with compliance with each regulatory control alternative.6 For this
analysis we first calculated the NPV of the stream of costs from 2021 through 20257 using a 4.25
percent discount rate. EPA typically uses a 3 and a 7 percent discount rate to discount future year
social benefits and social costs in regulatory impact analyses (USEPA, 2010). In this cost
annualization we use a 4.25 percent discount rate, which is consistent with the rate used in IPM's
objective function for minimizing the NPV of the stream of total costs of electricity generation.
This discount rate is meant to capture the observed equilibrium market rate at which investors
are willing to sacrifice present consumption for future consumption and is based on a Weighted
Average Cost of Capital (WACC).8 After calculating the NPV of the cost streams, the same 4.25
percent discount rate and 2021-2025 time period are used to calculate the levelized annual (i.e.,
annualized) cost estimates shown in Table 4-6.9 The same approach was used to develop the
annualized cost estimates for the 2021-2040 timeframe.
Additionally, note that the 2021-2025 and 2021-2040 equivalent annualized compliance
cost estimates have the expected relationship to each other; the annualized costs are lowest for
the less stringent alternative, and highest for the more stringent alternative.
4.4.4 Impacts on Fuel Use, Prices and Generation Mix
While the Revised CSAPR Update is expected to result in significant NOx emissions
reductions, it is estimated to result in relatively modest impacts to the power sector. While these
impacts are relatively small in percentage terms, consideration of these potential impacts is an
6	The XNPV() function in Microsoft Excel 2013 was used to calculate the NPV of the variable stream of costs, and
the PMT() function in Microsoft Excel 2013 is used to calculate the level annualized cost from the estimated NPV.
7	Consistent with the relationship between IPM run years and calendar years, EPA assigned 2023 compliance cost
estimates to both 2022 and 2023 in the calculation of NPV, and 2025 compliance cost to 2024 and 2025. For more
information, see Chapter 7 of the IPM Documentation.
8	The IPM Baseline run documentation (Section 10.4.1 Introduction to Discount Rate Calculations) states "The real
discount rate for all expenditures (capital, fuel, variable operations and maintenance, and fixed operations and
maintenance costs) in the EPA Platform v6 is 4.25%."
9	The PMT() function in Microsoft Excel 2013 is used to calculate the level annualized cost from the estimated
NPV.
4-29
-------
important component of assessing the relative impact of the regulatory control alternatives. In
this section we discuss the estimated changes in fuel use, fuel prices, generation by fuel type,
capacity by fuel type, and retail electricity prices.
Table 4-7 and Table 4-8 present the percentage changes in national coal and natural gas
usage by EGUs in 2021. These fuel use estimates reflect a modest shift to natural gas from coal.
The projected impacts in 2025 are similarly small.
Table 4-7.2021 Projected U.S. Power Sector Coal Use for the Baseline Run and the
Regulatory Control Alternatives
Million Tons
Percent Change from Baseline run


Revised
Less-

Revised
Less-
More-

Baseline
CSAPR
Stringent
More-
CSAPR
Stringent
Stringent

Run
Update
Alt.
Stringent Alt.
Update
Alt.
Alt.
Appalachia
85
85
85
85
0.14%
0.09%
0.14%
Interior
115
115
115
115
0.01%
0.01%
0.00%
Waste Coal
0
0
0
0
0.00%
0.00%
0.00%
West
287
286
286
286
-0.07%
-0.05%
-0.09%
Total
487
487
487
487
-0.01%
-0.01%
-0.03%
Table 4-8. 2021 Projected U.S. Power Sector Natural Gas Use for the Baseline Run and the
Regulatory Control Alternatives
Trillion Cubic Feet
Percent Change from Baseline run

Revised
More-
Revised
More-
Baseline
CSAPR Less-Stringent
Stringent
CSAPR Less-Stringent
Stringent
Run
Update Alternative
Alternative
Update Alternative
Alternative
11
11 11
11
0.00% 0.00%
0.00%
Table 4-9 and Table 4-10 present the projected coal and natural gas prices in 2021, as well
as the percent change from the baseline run projected as a result of the regulatory control
alternatives. These minor impacts in 2021 are consistent with the small changes in fuel use
summarized above. The projected impacts in 2025 are similarly very small.
4-30
-------
Table 4-9. 2021 Projected Minemouth and Power Sector Delivered Coal Price for the
	Baseline Run and the Regulatory Control Alternatives	
$/MMBtu
Percent Change from Baseline run

Baseline
Run
Revised Less-
CSAPR Stringent
Update Alternative
More-
Stringent
Alternative
Revised Less- More-
CSAPR Stringent Stringent
Update Alternative Alternative
Minemouth
1.21
1.21 1.21
1.21
0.02% 0.03% 0.06%
Delivered
1.87
1.87 1.87
1.87
-0.01% 0.01% 0.00%
Table 4-10. 2021 Projected Henry Hub and Power Sector Delivered Natural Gas Price for
	the Baseline Run and the Regulatory Control Alternatives	
$/MMBtu
Percent Change from Baseline run


Revised
Less-
More-
Revised
Less-
More-

Baseline
CSAPR
Stringent
Stringent
CSAPR
Stringent
Stringent

Run
Update
Alternative
Alternative
Update
Alternative
Alternative
Henry Hub
3.19
3.19
3.19
3.19
-0.02%
0.00%
-0.05%
Delivered
3.24
3.24
3.24
3.24
-0.02%
0.00%
-0.05%
Table 4-11 presents the projected percentage changes in the amount of electricity
generation in 2021 by fuel type. Consistent with the fuel use projections and emissions trends
above, EPA projects a small overall shift from coal to gas. The projected impact in 2025 is
similarly small.
Table 4-11. 2021 Projected U.S. Generation by Fuel Type for the Baseline Run and the
	Regulatory Control Alternatives	
Generation (TWh)
Percent Change from Baseline run


Revised
Less-
More-
Revised
Less-
More-

Baseline
CSAPR
Stringent
Stringent
CSAPR
Stringent
Stringent

Run
Update
Alternative
Alternative
Update
Alternative
Alternative
Coal
797
797
797
797
0.00%
0.00%
0.00%
Natural Gas
1,582
1,582
1,582
1,582
0.00%
0.00%
-0.01%
Nuclear
740
740
740
740
0.00%
0.00%
0.00%
Hydro
304
304
304
304
0.00%
0.00%
0.00%
Non-Hydro RE
536
536
536
536
0.00%
0.00%
0.00%
Oil\Gas Steam
58
58
58
58
0.01%
0.00%
0.07%
Other
34
34
34
34
-0.02%
-0.03%
0.02%
Total
4,051
4,051
4,051
4,051
0.00%
0.00%
0.00%
Note: In this table, "Non-Hydro RE" includes biomass, geothermal, landfill gas, solar, and wind.
4-31
-------
Table 4-12 presents the projected percentage changes in the amount of generating capacity
in 2021 and 2025 by primary fuel type. As explained above, none of the regulatory control
alternatives are expected to have a net impact on overall capacity by primary fuel type in 2021,
and the model was specified accordingly. By 2030 the Revised CSAPR Update is projected to
result in 594 MW of incremental coal retirements nationwide relative to the Baseline run,
constituting a reduction of 0.4% of national coal capacity. The majority of this capacity reflects
early retirement, i.e. retirement by 2025 of units that retired by 2035 in the baseline.
Table 4-12. 2021 and 2025 Projected U.S. Capacity by Fuel Type for the Baseline Run and
	the Regulatory Control Alternatives	
Capacity (GW)
Percent Change from Baseline run


Revised
Less-
More-
Revised
Less-
More-
2021
Baseline
CSAPR
Stringent
Stringent
CSAPR
Stringent
Stringent

Run
Update
Alternative
Alternative
Update
Alternative
Alternative
Coal
216
216
216
216
0.0%
0.0%
0.0%
Natural Gas
421
421
421
421
0.0%
0.0%
0.0%
Nuclear
94
94
94
94
0.0%
0.0%
0.0%
Hydro
107
107
107
107
0.0%
0.0%
0.0%
Non-Hydro RE
184
184
184
184
0.0%
0.0%
0.0%
Oil\Gas Steam
74
74
74
74
0.0%
0.0%
0.0%
Other
8
8
8
8
0.0%
0.0%
0.0%
Total
1,106
1,106
1,106
1,106
0.0%
0.0%
0.0%
Note: In this table,
'Non-Hydro RE
" includes biomass, geothermal, landfill gas, solar, and wind

Capacity (GW)
Percent Change from Baseline Run


Revised
Less-
More-
Revised
Less-
More-
2025
Baseline
CSAPR
Stringent
Stringent
CSAPR
Stringent
Stringent

Run
Update
Alternative
Alternative
Update
Alternative
Alternative
Coal
167
166
166
166
-0.4%
-0.3%
-0.6%
Natural Gas
418
418
418
418
0.0%
0.0%
0.1%
Nuclear
77
77
77
77
0.5%
0.4%
1.1%
Hydro
110
110
110
110
0.0%
0.0%
0.0%
Non-Hydro RE
231
231
231
231
0.0%
0.0%
0.0%
Oil\Gas Steam
67
67
67
67
0.1%
0.0%
-0.1%
Other
11
11
11
11
0.0%
0.0%
0.0%
Total
1,081
1,081
1,081
1,081
0.0%
0.0%
0.0%
Note: In this table, "Non-Hydro RE" includes biomass, geothermal, landfill gas, solar, and wind
4-32
-------
EPA estimated the change in the retail price of electricity (2016$) using the Retail Price
Model (RPM).10 The RPM was developed by ICF for EPA, and uses the IPM estimates of
changes in the cost of generating electricity to estimate the changes in average retail electricity
prices. The prices are average prices over consumer classes (i.e., consumer, commercial, and
industrial) and regions, weighted by the amount of electricity used by each class and in each
region. The RPM combines the IPM annual cost estimates in each of the 64 IPM regions with
EIA electricity market data for each of the 22 electricity supply regions (shown in Figure 4-1) in
the electricity market module of the National Energy Modeling System (NEMS).11
Table 4-13 and Table 4-14 present the projected percentage changes in the retail price of
electricity for the three regulatory control alternatives in 2021 and 2025, respectively. Consistent
with other projected impacts presented above, average retail electricity prices at both the national
and regional level are projected to be small. By 2025, EPA estimates that this rule will result in a
0.02 percent increase in national average retail electricity price, or by about 0.02 mills/kWh.
Table 4-13. Average Retail Electricity Price by Region for the Baseline Run and the
Regulatory Control Alternatives, 2021
All Sector
2021 Average Retail Electricity Price
(2016 mills/kWh)
Percent Change from Baseline Run

Baseline
Run
Revised
Less-
More-
Revised
Less-
More-
Region
CSAPR
Update
Stringent
Alternative
Stringent
Alternative
CSAPR
Update
Stringent
Alternative
Stringent
Alternative
MROE
117
117
117
117
0%
0%
0%
NYCW
166
166
166
166
0%
0%
0%
NYLI
134
134
134
134
0%
0%
0%
NYUP
109
109
109
109
0%
0%
0%
RFCE
115
115
115
115
0%
0%
0%
RFCM
91
91
91
91
0%
0%
0%
RFCW
93
93
93
93
0%
0%
0%
SRDA
83
83
83
83
0%
0%
0%
SRGW
87
87
87
87
0%
0%
0%
SRCE
85
85
85
85
0%
0%
0%
SRVC
99
99
99
99
0%
0%
0%
SPSO
88
88
88
88
0%
0%
0%
NATIONAL
99
99
99
99
0%
0%
0%
10	See documentation available at: https://www.epa.gov/airmarkets/retail-price-model
11	See documentation available at:
http://www.eia.gov/forecasts/aeo/nems/documentation/electricity /pdf/m068(2014).pdf
4-33
-------
Residential
Sector
2021 Average Retail Electricity Price
(2016 mills/kWh)
Percent Change from Baseline Run

Baseline
Run
Revised
Less-
More-
Revised
Less-
More-
Region
CSAPR
Update
Stringent
Alternative
Stringent
Alternative
CSAPR
Update
Stringent
Alternative
Stringent
Alternative
MROE
153
153
153
153
0%
0%
0%
NYCW
401
401
401
401
0%
0%
0%
NYLI
174
174
174
174
0%
0%
0%
NYUP
137
137
137
137
0%
0%
0%
RFCE
152
152
152
153
0%
0%
1%
RFCM
146
146
146
146
0%
0%
0%
RFCW
126
126
126
126
0%
0%
0%
SRDA
97
97
97
97
0%
0%
0%
SRGW
107
107
107
107
0%
0%
0%
SRCE
97
97
97
97
0%
0%
0%
SRVC
114
114
114
114
0%
0%
0%
SPSO
103
103
103
103
0%
0%
0%
NATIONAL
122
122
122
122
0%
0%
0%

Commercial
Sector
2021 Average Retail Electricity Price
(2016 mills/kWh)
Percent Change from Baseline Run

Baseline
Run
Revised
Less-
More-
Revised
Less-
More-
Region
CSAPR
Update
Stringent
Alternative
Stringent
Alternative
CSAPR
Update
Stringent
Alternative
Stringent
Alternative
MROE
119
119
119
119
0%
0%
0%
NYCW
105
105
105
105
0%
0%
0%
NYLI
115
115
115
115
0%
0%
0%
NYUP
98
98
98
98
0%
0%
0%
RFCE
103
103
103
103
0%
0%
0%
RFCM
70
70
70
70
0%
0%
0%
RFCW
88
88
88
88
0%
0%
0%
SRDA
81
81
81
81
0%
0%
0%
SRGW
83
83
83
83
0%
0%
0%
SRCE
81
81
81
81
0%
0%
0%
SRVC
93
93
93
93
0%
0%
0%
SPSO
85
85
85
85
0%
0%
0%
NATIONAL
93
93
93
93
0%
0%
0%

Industrial
Sector
2021 Average Retail Electricity Price
(2016 mills/kWh)
Percent Change from Baseline Run

Baseline
Run
Revised
Less-
More-
Revised
Less-
More-
Region
CSAPR
Update
Stringent
Alternative
Stringent
Alternative
CSAPR
Update
Stringent
Alternative
Stringent
Alternative
4-34
-------
MROE
86
86
86
86
0%
0%
0%
NYCW
94
94
94
94
0%
0%
0%
NYLI
53
53
53
53
0%
0%
0%
NYUP
84
84
84
84
0%
0%
0%
RFCE
72
72
72
72
0%
0%
0%
RFCM
61
61
61
61
0%
0%
0%
RFCW
68
68
68
68
0%
0%
0%
SRDA
68
68
68
68
0%
0%
0%
SRGW
54
54
54
54
0%
0%
0%
SRCE
71
71
71
71
0%
0%
0%
SRVC
83
83
83
83
0%
0%
0%
SPSO
73
73
73
73
0%
0%
0%
NATIONAL
75
75
75
75
0%
0%
0%
Table 4-14. Average Retail Electricity Price by Region for the Baseline Run and the
Regulatory Control Alternatives, 2025
All Sector
2025 Average Retail Electricity Price
(2016 mills/kWh)
Percent Change from Baseline Run

Baseline
Run
Revised
Less-
More-
Revised
Less-
More-
Region
CSAPR
Update
Stringent
Alternative
Stringent
Alternative
CSAPR
Update
Stringent
Alternative
Stringent
Alternative
MROE
115
115
115
115
0%
0%
0%
NYCW
198
198
198
198
0%
0%
0%
NYLI
159
159
159
159
0%
0%
0%
NYUP
135
135
134
134
0%
0%
0%
RFCE
132
132
132
132
0%
0%
0%
RFCM
105
105
105
105
0%
0%
0%
RFCW
104
104
104
104
0%
0%
0%
SRDA
83
83
83
83
0%
0%
0%
SRGW
97
97
97
97
0%
0%
0%
SRCE
83
83
83
83
0%
0%
0%
SRVC
100
100
100
101
0%
0%
0%
SPSO
93
93
93
93
0%
0%
0%
NATIONAL
105
105
105
105
0%
0%
0%
Residential
Sector
2025 Average Retail Electricity Price
(2016 mills/kWh)
Percent Change from Baseline Run

Baseline
Run
Revised
Less-
More-
Revised
Less-
More-
Region
CSAPR
Update
Stringent
Alternative
Stringent
Alternative
CSAPR
Update
Stringent
Alternative
Stringent
Alternative
MROE
148
148
148
148
0%
0%
0%
NYCW
477
477
477
477
0%
0%
0%
NYLI
210
210
210
210
0%
0%
0%
NYUP
172
172
172
172
0%
0%
0%
4-35
-------
RFCE
178
178
178
178
0%
0%
0%
RFCM
171
172
172
172
0%
0%
0%
RFCW
144
144
144
144
0%
0%
0%
SRDA
97
97
97
97
0%
0%
0%
SRGW
119
119
119
119
0%
0%
0%
SRCE
95
95
95
95
0%
0%
0%
SRVC
115
115
115
115
0%
0%
0%
SPSO
108
108
108
108
0%
0%
0%
NATIONAL
130
130
130
130
0%
0%
0%

Commercial
2025 Average Retail Electricity Price
(2016 mills/kWh)
Percent Change from Baseline Run

Baseline
Run
Revised
Less-
More-
Revised
Less-
More-
Region
CSAPR
Update
Stringent
Alternative
Stringent
Alternative
CSAPR
Update
Stringent
Alternative
Stringent
Alternative
MROE
117
117
117
117
0%
0%
0%
NYCW
130
130
130
130
0%
0%
0%
NYLI
138
138
138
138
0%
0%
0%
NYUP
120
120
120
120
0%
0%
0%
RFCE
116
116
116
116
0%
0%
0%
RFCM
85
85
85
85
0%
0%
1%
RFCW
101
101
101
101
0%
0%
0%
SRDA
81
81
81
81
0%
0%
0%
SRGW
94
94
94
94
0%
0%
0%
SRCE
80
80
80
80
0%
0%
0%
SRVC
95
95
95
95
0%
0%
0%
SPSO
90
90
90
90
0%
0%
0%
NATIONAL
99
99
99
99
0%
0%
0%

Industrial
Sector
2025 Average Retail Electricity Price
(2016 mills/kWh)
Percent Change from Baseline Run

Baseline
Run
Revised
Less-
More-
Revised
Less-
More-
Region
CSAPR
Update
Stringent
Alternative
Stringent
Alternative
CSAPR
Update
Stringent
Alternative
Stringent
Alternative
MROE
86
86
86
86
0%
0%
0%
NYCW
119
119
119
119
0%
0%
0%
NYLI
80
80
80
80
0%
0%
0%
NYUP
106
106
106
106
0%
0%
0%
RFCE
90
89
89
90
0%
0%
0%
RFCM
68
69
69
69
0%
0%
1%
RFCW
75
75
75
76
0%
0%
0%
SRDA
68
68
68
68
0%
0%
0%
SRGW
65
65
65
65
0%
0%
0%
SRCE
70
70
70
70
0%
0%
0%
SRVC
83
83
83
83
0%
0%
0%
SPSO
77
77
77
77
4-36
0%
0%
0%
-------
NATIONAL 80	80	80
80
0%	0%	0%
.NEWE,
MYUP
NYCW
WOW
AZNM.
SRDA
Figure 4-1. Electricity Market Module Regions
Source: EIA (http://www.eia.gov/forecasts/aeo/pdf/nerc_map.pdf)
4.5 Social Costs
As discussed in EPA's Guidelines for Preparing Economic Analyses, social costs are the
total economic burden of a regulatory action (USEPA, 2010). This burden is the sum of all
opportunity costs incurred due to the regulatory action, where an opportunity cost is the value
lost to society of any goods and services that will not be produced and consumed as a result of
reallocating some resources towards pollution mitigation. Estimates of social costs may be
compared to the social benefits expected as a result of a regulation to assess its net impact on
society. The social costs of a regulatory action will not necessarily be equal to the expenditures
by the electricity sector to comply with the rule. Nonetheless, here we use compliance costs as a
proxy for social costs.
4-37
-------
The compliance cost estimates for the rule and more or less stringent regulatory control
alternatives presented in this chapter are the change in expenditures by the electricity generating
sector required by the power sector for compliance under each alternative. The change in the
expenditures required by the power sector to maintain compliance reflect the changes in
electricity production costs resulting from application of NOx control strategies, including
changes in expenditures resulting from changes in the mix of fuels used for generation, necessary
to comply with the emissions budgets. Ultimately, part of the compliance costs may be borne by
electricity consumers through higher electricity prices. As discussed above, the electricity and
fossil fuel price impacts from this rule are expected to be small.
4.6 Limitations
EPA's modeling is based on expert judgment of various input assumptions for variables
whose outcomes are uncertain. As a general matter, the Agency reviews the best available
information from engineering studies of air pollution controls and new capacity construction
costs to support a reasonable modeling framework for analyzing the cost, emission changes, and
other impacts of regulatory actions.
As described in the preamble at section VII.C.4.C, EPA has introduced a safety valve
which gives Group 3 sources the option, in February 2022 only, to purchase additional 2021
Group 3 allowances by surrendering 18 banked 2017-2020 Group 2 allowances per Group 3
allowance. This safety valve was not implemented in the modeling presented in this RIA.
However, since the analysis shows that 2021-22 costs remain modest and variability limits are
not hit, inclusion of the safety valve would not have changed the modeling results because this
option would not have been selected in the model.
The IPM-projected annualized cost estimates of private compliance costs provided in this
analysis are meant to show the increase in production (generating) costs to the power sector in
response to the rule. To estimate these annualized costs, EPA uses a conventional and widely
accepted approach that applies a capital recovery factor (CRF) multiplier to capital investments
and adds that to the annual incremental operating expenses. The CRF is derived from estimates
of the cost of capital (private discount rate), the amount of insurance coverage required, local
4-38
-------
property taxes, and the life of capital. The private compliance costs presented earlier are EPA's
best estimate of the direct private compliance costs of the rule.
As discussed in section 4.3.2, IPM v6 does not have the capacity to endogenously
determine whether or not to maximize the use of existing EGU post-combustion NOx controls
(i.e., SCR), or install/upgrade combustion controls in response to a regulatory control
requirement. These decisions were imposed exogenously on the model, as documented in section
4.3.2 and Ozone Transport Policy Analysis Final Rule TSD. While the emissions projections
reflect operation of these controls, the projected compliance costs were supplemented with
exogenously estimated costs of optimizing SCR operation, optimizing SNCR operation, and
installing/upgrading combustion controls (see section 4.3.3). As a result of this modeling
approach, the dispatch decisions made within the model do not take into consideration the
additional operating costs associated with these three types of compliance strategies (the
operating costs of the units on which these strategies are imposed do not reflect the additional
costs of these strategies). The effect of changes in facility and system-wide emissions from these
changes in operating costs are also not accounted for in the spatial fields for the regulatory
alternatives described in Chapter 3. These additional costs and their influence on projected
changes in emissions and the level and location of ozone and PM2.5 concentration patterns from
the regulatory alternatives are relatively minor, and do not have a significant impact on the
overall finding that the economic impacts of this rule are minimal.
Additionally, the modeling includes two emission reduction strategies that are exogenously
imposed where applicable: turning on idled SCRs and SNCRs (Revised CSAPR Update and
more-stringent alternative). While these strategies are exogenously imposed, the costs and
emissions reductions are estimated within IPM. Since the costs of these strategies are accounted
for within the model, they are able to influence the projected behavior of the EGUs within the
model.
The annualized cost of the final rule, as quantified here, is EPA's best assessment of the
cost of implementing the rule. These costs are generated from rigorous economic modeling of
changes in the power sector due to implementation of the Revised CSAPR Update.
4-39
-------
4.7 References
U.S. Energy Information Administration (EIA). 2014. The Electricity Market Module of the
National Energy Modeling System: Model Documentation 2014. Available at:

Accessed 9/17/2015.
U.S. EPA, 2015. Carbon Pollution Emission Guidelines for Existing Stationary Sources: Electric
Utility Generating Units (FinalRule), http://www2.epa.gov/cleanpowerplan/clean-
power-plan-existing-power-plants.
U.S. EPA, 2015 a. Standards of Performance for Greenhouse Gas Emissions from New,
Modified, and Reconstructed Stationary Sources: Electric Utility Generating Units (Final
Rule), http://www2.epa.gov/cleanpowerplan/carbon-pollution-standards-new-modified-
and-reconstructed-power-plants.
U.S. EPA, 2015b. Disposal of Coal Combustion Residuals from Electric Utilities (Final Rule),
http://www2.epa.gov/coalash/coal-ash-rule.
U.S. EPA, 2015c. Steam Electric Power Generating Effluent Guidelines (Final Rule),
http://www2.epa.gov/eg/steam-electric-power-generating-effluent-guidelines-2015-final-
rule.
U.S. EPA, 2014. Final Rule for Existing Power Plants and Factories,
http://www2.epa.gov/cooling-water-intakes.
U.S. EPA, 2011. Cross-State Air Pollution Rule,
http://www3.epa.gov/airtransport/CSAPR/index.html
U.S. EPA, 201 la. Mercury and Air Toxics Standards (MATS), http://www3.epa.gov/mats/.
U.S. EPA. 2010. EPA Guidelines for Preparing Economic Analyses. Available at:
. Accessed 9/21/2015.
U.S. EPA. 2010a. Regulatory Impact Analysis for the Proposed Federal Transport Rule
Analyses. Available at: < http://www3.epa.gov/ttnecasl/ria.html>. Accessed 9/21/2015.
U.S. EPA, 2005. Clean Air Interstate Rule,
http://archive.epa.gov/airmarkets/programs/cair/web/html/index.html.
4-40
-------
CHAPTER 5: BENEFITS
Overview
This final action to revise the Cross-State Air Pollution Rule (CSAPR) Update reduces
the emissions of nitrogen oxides (NOx) transported from states that contribute significantly to
nonattainment or interfere with maintenance of the 2008 ozone National Ambient Air Quality
Standard (NAAQS) in downwind states. Implementing the Final Revised CSAPR Update is
expected to reduce emissions of NOx which will in turn reduce ozone and fine particle (PM2.5)
concentrations; the rule will also reduce carbon dioxide (CO2) emissions. This chapter reports the
estimated monetized health benefits from reducing concentrations of ozone and PM2.5 and global
climate benefits associated with emission reductions for the three regulatory control alternatives
across several discount rates.
This chapter describes the methods used to estimate health benefits from reducing
concentrations of ozone and PM2.5. While the steps to performing an air pollution benefits
analysis remains unchanged, both the data used to quantify each health endpoint, and the
endpoints quantified, have been updated to reflect more recent scientific evidence. EPA
committed to the update in the proposal Revised CASPR Update RIA; however, the timing of the
final rule has prevented the peer review of the updated approach prior to EPA issuing this final
RIA. Nevertheless, this update uses information from the recent PM2.5 and ozone ISAs (U.S.
EPA, 2019a, U.S. EPA, 2020b); these ISAs were reviewed by CAS AC and the public. These
updates are summarized below and detailed in a technical support document (TSD) for the Final
Revised CSAPR for the 2008 Ozone NAAQS Update titled Estimating PM2.5- and Ozone-
Attributable Health Benefits. The chapter also describes the methods used to estimate the climate
benefits from reductions of CO2 emissions. Data, resource, and methodological limitations
prevent EPA from monetizing health benefits of reducing direct exposure to NO2, ecosystem
effects and visibility impairment as well as benefits from reductions in other pollutants, such as
hazardous air pollutants (HAP). We qualitatively discuss these unquantified benefits in this
chapter.
5-1
-------
5.1 Estimated Human Health Benefits
The final rule is expected to reduce ozone season and annual NOx emissions. In the
presence of sunlight, NOx and volatile organic compounds (VOCs) can undergo a chemical
reaction in the atmosphere to form ozone. Reducing NOx emissions generally reduces human
exposure to ozone and the incidence of ozone-related health effects, though the degree to which
ozone is reduced will depend in part on local levels of VOCs. The final rule would also reduce
emissions of NOx throughout the year. Because NOx is also a precursor to formation of ambient
PM2.5, reducing these emissions would reduce human exposure to ambient PM2.5 throughout the
year and would reduce the incidence of PIVh.s-attributable health effects.1 Reducing emissions of
NOx would also reduce ambient exposure to NO2 and its associated health effects, though we do
not quantify these effects due to lack of data.
In the proposed Revised CSAPR Update regulatory impact analysis (RIA) EPA committed
to updating its approach for quantifying the benefits of changes in PM2.5 and ozone in this final
Revised CSAPR Update RIA (U.S. EPA 2020c). The updated approach incorporates evidence
reported in the recently completed PM2.5 and Ozone ISAs and accounts for recommendations
from the Science Advisory Board (U.S. EPA 2019, U.S. EPA 2020b, U.S. EPA-SAB 2019, U.S.
EPA-SAB 2020). When updating each health endpoint EPA considered: (1) the extent to which
the science supports the existence of a causal relationship between that pollutant and the adverse
effect; (2) whether suitable epidemiologic studies exist to support quantifying health impacts; (3)
and whether robust economic approaches are available for estimating the value of the impact of
reducing human exposure to the pollutant. Our approach for updating the endpoints and to
identify suitable epidemiologic studies, baseline incidence rates, population demographics, and
valuation estimates is summarized below. Detailed descriptions of these updates are available in
the TSD for the Final Revised Cross-State Air Pollution Rule for the 2008 Ozone NAAQS
Update titled Estimating PM2.5- and Ozone-Attributable Health Benefits.
EPA followed a five-step approach for updating its methodology for quantifying and
monetizing ozone and PM2.5 attributable health endpoints:
1 This RIA does not quantify PM2 5-related benefits associated with direct PM2 5 and SO2 emission reductions. As
discussed in Chapter 4, EPA does not estimate significant direct PM2 5 and SO2 emission reductions as a result of
this rule.
5-2
-------
1.	Establish criteria for identifying studies and risk estimates most appropriate to inform a
PM2.5 and O3 benefit analysis for an RIA. Study criteria, such as study design, location,
population characteristics, and other attributes, were used to identify the most suitable
estimates. This step precedes health endpoint identification to ensure impartial health
endpoint identification and prevent identification of non-quantifiable endpoints.
2.	Identify pollutant-attributable health effects for which the ISA reports strong evidence
and that may be quantified in a benefits assessment. EPA considered new evidence
reported in the recent IS As (U.S. EPA, 2019a, U.S. EPA, 2020b) and clinically
significant outcomes (e.g. premature mortality and hospital admissions) for which
endpoint-specific baseline incidence data is available. While ISAs form casual
determinations for broad endpoint categories (e.g., respiratory effects), which are
generally preferred over specific health endpoints (e.g., hay fever symptoms) for
comprehensive benefits assessments, they do not make causal determinations for each
specific health endpoints. Instead, the ISAs provide information on the strength and
consistency of the evidence supporting more specific endpoints within each broad
category. The strength and consistency of evidence supporting relationships with specific
health endpoints, together with the broad category causality determinations, are used
when identifying specific health endpoints for inclusion in benefits assessments. New
ISA evidence was considered sufficient for inclusion in the benefits assessment if the ISA
determine the broad heath endpoint category was causally related to pollutant exposure
(TSD section 2.2.1.1), the ISA determined that the specific health endpoint is a
biologically plausible health effect of exposure (TSD section 2.2.1.2 of TSD), and the
ISA found strong and consistent support relating the specific health endpoint with
pollutant exposure (TSD section 2.3).
3.	Collect baseline incidence and prevalence estimates and demographic information. EPA
develops either daily or annual baseline incidence and prevalence rates at the most
geographically- and age-specific levels feasible for each health endpoint assessed. EPA
uses population projections based on economic forecasting models developed by Woods
and Poole, Inc. (Woods & Poole, 2015). The Woods and Poole (WP) database contains
county-level projections of population by age, sex, and race out to 2050, relative to a
baseline using the 2010 Decennial Census.
5-3
-------
4.	Develop economic unit values. To directly compare benefits estimates associated with a
rulemaking to cost estimates, the number of instances of each air pollution-attributable
health impact must be converted to a monetary value. This requires a valuation estimate
for each unique health endpoint, and potentially also discounting if the benefits are
expected to accrue over more than a single year. EPA develops valuation estimates at the
most age-refined level feasible for each health endpoint assessed.
5.	Characterize uncertainty associated with quantified benefits estimates. Building on
EPA's current methods for characterizing uncertainty, these approaches will include,
among others, reporting confidence intervals calculated from risk estimates and separate
quantification using multiple studies and risk estimates for particularly influential
endpoints (e.g., mortality risk), and approaches for aggregating and representing the
results of multiple studies evaluating a particular health endpoint.2
The Estimating PM2.5- and Ozone-Attributable Health Benefits TSD provides a full
discussion of the Agency's updated approach for quantifying the number and value of estimated
air pollution-related impacts. In this document the reader can find the rationale for selecting
health endpoints to quantify; the demographic, health and economic data used; modeling
assumptions; and our techniques for quantifying uncertainty.3
Implementing the final rule will affect the distribution of ozone and PM2.5 concentrations
in much of the U.S.; this includes locations both meeting and exceeding the NAAQS for ozone
and PM. This RIA estimates avoided ozone- and PIVh.s-related health impacts that are distinct
from those reported in the RIAs for both NAAQS (U.S. EPA 2012, 2015e). The ozone and PM2.5
NAAQS RIAs hypothesize, but do not predict, the benefits and costs of strategies that States may
choose to enact when implementing a revised NAAQS; these costs and benefits are illustrative
and cannot be added to the costs and benefits of policies that prescribe specific emission control
measures. This RIA estimates the benefits (and costs) of specific emissions control measures.
2	We consider study quality, inter-study heterogeneity, and redundancy where epidemiologic risk estimates are
combined or aggregated.
3	Updated information has been incorporated into BenMAP-CE version 1.5.7.0, which will be publicly released by
March 15, 2021.
5-4
-------
We project how the levels of ozone and PM2.5 may increase and decrease over the U.S.
compared to the baseline. Some portion of the air quality and health benefits from the regulatory
control alternatives will occur in areas not attaining the Ozone or PM2.5 NAAQS. However, we
do not simulate how states would account for this rule when complying with the NAAQS; this
affects the estimated benefits (and costs) of the final rule and more and less stringent alternatives,
which introduces uncertainty in the estimated benefits (and costs).
5.2.1 Health Impact Assessment for Ozone and PM2.5
The benefits analysis presented in this chapter incorporates an array of science-policy and
technical changes that the Agency has adopted since the previous reviews of the PM2.5 standards
in 2012 and the ozone standards in 2015, as well as since publishing the most recent major
benefits analysis, documented in the benefits chapter of the RIA accompanying the proposed
Revised CSAPR Update (U.S. EPA, 2020c). Below we note aspects of this analysis that differ
from the Revised CSAPR Update proposal RIA. The rationale for these choices is detailed in the
Estimating PM2.5- and Ozone-Attributable Health Benefits TSD:
1. Updated mortality studies and endpoints
a. Removed short-term all-cause ozone mortality endpoint and include respiratory
ozone mortality. The recent ozone ISA downgraded the causal determination between
short-term ozone exposure and total mortality to "suggestive of, but not sufficient to
infer, a causal relationship." As the 2020 ozone ISA retained the long-term ozone
respiratory determination of "likely to be causal" we selected a risk estimate for this
health endpoint from Turner et al. (2016). As such, respiratory ozone mortality
endpoints were added (for short-term), and all-cause (or non-accidental) ozone
mortality based upon (Smith et al. 2009) was removed. Qualitatively, the updated risk
estimate of long-term ozone exposure-attributable respiratory mortality is larger than
the Jerrett et al. (2009) estimate used previously. The two identified risk estimates of
short-term ozone exposure -attributable respiratory mortality are not directly
comparable to previous estimates of short-term exposure-related total/nonaccidental
mortality as the different baseline incidence rates associated with each health
endpoint also impact the benefits estimates.
5-5
-------
b.	Removed the Harvard Six Cities (HSC) PM2.5 mortality study. Several recent
epidemiological analyses of PM2.5 and total mortality have increased geographical
coverage and better characterize recent exposures as compared to the HCS cohort, so
Lepeule et al. (2012) was removed.
c.	Incorporated newer long-term exposure PM2.5 mortality studies. We selected risk
estimates from an extended analysis of the American Cancer Society (ACS) cohort
analysis Turner et al. (2016) and a recent analysis of the Medicare cohort Di et al.
(2017). Compared to the ACS study it replaces (Krewski et al. 2009), Turner et al.
(2016) incorporates additional years of cohort follow-up and PM2.5 air quality data.
We also selected a risk estimate from an analysis of the Medicare cohort (Di et al.
2017), which incorporates a large and representative population, improved exposure
assessment methods, and recent air quality information. Qualitatively, the two
updated risk estimates are both very similar to the previous Krewski et al. (2009)
estimate of mortality derived from the ACS cohort.
2.	Updated morbidity endpoints. Incorporated new PM2.5 and ozone morbidity endpoints. Upon
careful evaluation of this new literature identified in the 2019 PM ISA and 2020 Ozone ISA
(U.S. EPA, 2019, U.S. EPA, 2020) we added several new morbidity endpoints to our health
impact assessment. New endpoints include the association between exposure and
cardiovascular emergency department visits, lung cancer, allergic rhinitis (hay fever), asthma
onset, Alzheimer's disease hospital admissions, Parkinson's disease hospital admission, and
stroke.
3.	Incorporated new demographic data. We updated the baseline population demographic data
to reflect an updated projection of the 2010 Census data to future years (Woods & Poole,
2015). These data replace the earlier demographic projection data from Woods & Poole
(2015).
4.	Incorporated new valuations estimates. Economic valuation estimates were developed or
identified for new morbidity health endpoints.
5.	Incorporated updated uncertainty analyses. Advancements in epidemiology allow for updated
uncertainty and sensitivity analyses, such as those comparing different techniques for
estimating exposure, are included in the TSD. Where possible, we quantitatively assess
5-6
-------
uncertainties associated with PM2.5 benefit estimates (TSD section 6.1), ozone benefit
estimates (TSD section 6.2), baseline incidence rate estimates (TSD section 6.3), and
economic valuation estimates (TSD section 6.4). We focus on input parameters that are likely
to most greatly influence the size of the estimated health impacts, such as those related to
mortality impacts. When quantitative analysis is not possible, we characterize the sensitivity
of the results to alternative plausible input parameters when feasible.
6. Replaced medical charges with medical costs, when possible. Based on newly available data
from the Healthcare Cost and Utilization Project (HCUP) National Inpatient Sample (NIS),
we replaced several hospital admissions charge estimates with cost estimates, allowing for a
more appropriate estimate of the costs borne by patients and insurance companies.4 Using the
hospital admission cost-to-charge ratio, we also developed emergency department cost
estimates, which replaced emergency department charge estimates.
Estimating the health benefits of reductions in PM2.5 and O3 exposure begins with
estimating the change in exposure for each individual and then estimating the change in each
individual's risks for those health outcomes affected by exposure. The benefit of the reduction in
each health risk is based on the exposed individual's WTP for the risk change, assuming that
each outcome is independent of one another. The greater the magnitude of the risk reduction
from a given change in concentration, the greater the individual's WTP, all else equal. The social
benefit of the change in health risks equals the sum of the individual WTP estimates across all of
the affected individuals.5 We conduct this analysis by adapting primary research—specifically,
air pollution epidemiology studies and economic value studies—from similar contexts. This
approach is sometimes referred to as "benefits transfer." Below we describe the procedure we
follow for: (1) selecting air pollution health endpoints to quantify; (2) calculating counts of air
pollution effects using a health impact function; (3) specifying the health impact function with
concentration-response parameters drawn from the epidemiological literature.
4	Willingness-to-pay (WTP) is the preferred welfare estimate for avoiding the risk of health effects. However, for
more endpoints WTP estimates are not available and therefore cost-of-illness (COI) estimates are used as a proxy.
Usually COI estimates are an underestimate of WTP. Cost estimates are a better estimate of COI than charge
estimates as they reflect the actual cost to society of providing care for these endpoints.
5	This RIA also reports the change in the sum of the risk, or the change in the total incidence, of a health outcome
across the population. If the benefit per unit of risk is invariant across individuals, the total expected change in the
incidence of the health outcome across the population can be multiplied by the benefit per unit of risk to estimate the
social benefit of the total expected change in the incidence of the health outcome.
5-7
-------
5.2.1.1 Selecting Air Pollution Health Endpoints to Quantify
As a first step in quantifying ozone and PM2.5-related human health impacts, the Agency
consults the Integrated Science Assessment for Ozone and Related Photochemical Oxidants
(Ozone ISA) (U.S. EPA 2020b) and the Integrated Science Assessment for Particulate Matter
(PM ISA) (U.S. EPA 2019a). These two documents synthesize the toxicological, clinical and
epidemiological evidence to determine whether each pollutant is causally related to an array of
adverse human health outcomes associated with either acute (i.e., hours or days-long) or chronic
(i.e. years-long) exposure; for each outcome, the ISA reports this relationship to be causal, likely
to be causal, suggestive of a causal relationship, inadequate to infer a causal relationship or not
likely to be a causal relationship. The Agency estimates the incidence of air pollution effects for
those health endpoints above where the ISA classified as either causal or likely-to-be-causal. In
brief, the ISA for ozone found short-term (less than one month) exposures to ozone to be
causally related to respiratory effects, a "likely to be causal" relationship with metabolic effects
and a "suggestive of, but not sufficient to infer, a causal relationship" for central nervous system
effects, cardiovascular effects, and total mortality. The ISA reported that long-term exposures
(one month or longer) to ozone are "likely to be causal" for respiratory effects including
respiratory mortality, and a "suggestive of, but not sufficient to infer, a causal relationship" for
cardiovascular effects, reproductive effects, central nervous system effects, metabolic effects,
and total mortality. The PM ISA found short-term exposure to PM2.5 to be causally related to
cardiovascular effects and mortality (i.e., premature death), respiratory effects as likely-to-be-
causally related, and a suggestive relationship for metabolic effects and nervous system effects.
The ISA identified cardiovascular effects and total mortality as being causally related to long-
term exposure to PM2.5. A likely-to-be-causal relationship was determined between long-term
PM2.5 exposures and respiratory effects, nervous system effects, and cancer effects were; and the
evidence was suggestive of a causal relationship for male and female reproduction and fertility
effects, pregnancy and birth outcomes, and metabolic effects. Table 5-1 reports the effects we
quantified and those we did not quantify in this RIA. The list of benefit categories not quantified
is not exhaustive. And, among the effects quantified, it might not have been possible to quantify
completely either the full range of human health impacts or economic values. The table below
omits health effects associated with NO2 exposure, and any welfare effects such as acidification
5-8
-------
and nutrient enrichment; these effects are described in the Ozone and PM NAAQS RIA (U.S.
EPA 2015e, 2012) and summarized later in this appendix.
Consistent with economic theory, the willingness to pay (WTP) for reductions in
exposure to environmental hazard will depend on the expected impact of those reductions on
human health and other outcomes. All else equal, WTP is expected to be higher when there is
stronger evidence of a causal relationship between exposure to the contaminant and changes in a
health outcome (McGartland et al., 2017). For example, in the case where there is no evidence of
a potential relationship the WTP would be expected to be zero and the effect should be excluded
from the analysis. Alternatively, when there is some evidence of a relationship between exposure
and the health outcome, but that evidence is insufficient to definitively conclude that there is a
causal relationship, individuals may have a positive WTP for a reduction in exposure to that
hazard (U.S. EPA-SAB 2020b, Kivi and Shogren, 2010). Lastly, the WTP for reductions in
exposure to pollutants with strong evidence of a relationship between exposure and effect are
likely positive and larger than for endpoints where evidence is weak, all else equal.
Unfortunately, the economic literature currently lacks a settled approach for accounting for how
WTP may vary with uncertainty about causal relationships.
Given these challenges, the Agency draws its assessment of the strength of evidence on
the relationship between exposure to PM2.5 or O3 and potential health endpoints from the ISAs
that are developed for the NAAQS process as discussed above. The focus on categories
identified as having a "causal" or "likely to be causal" relationship with the pollutant of interest
is to estimate the pollutant-attributable human health benefits in which we are most confident.6
All else equal, this approach may underestimate the benefits of PM2.5 and O3 exposure reductions
as individuals may be WTP to avoid specific risks where the evidence is insufficient to conclude
they are "likely to be caus[ed]" by exposure to these pollutants.7 At the same time, WTP may be
lower for those health outcomes for which causality has not been definitively established. This
approach treats relationships with ISA causality determinations of "likely to be causal" as if they
were known to be causal, and therefore benefits could be overestimated.
6	This decision criterion for selecting health effects to quantify and monetize PM2 5 and O3 is only applicable to
estimating the benefits of exposure of these two pollutants. This decision criterion may not be applicable or suitable
for quantifying and monetizing health and ecological effects of other pollutants.
7	EPA includes risk estimates for an example health endpoint with a causality determination of "suggestive, but not
sufficient to infer" that is associated with a potentially substantial economic value in the quantitative uncertainty
characterization {Estimating PM2.5- and Ozone-Attributable Health Benefits TSD section 6.2.3).
5-9
-------
Table 5-1. Health Effects of Ambient Ozone and PM2.5
Category
Effect
Effect
Quantified
Effect
Monetized
More
Information
Premature
mortality from
exposure to PM2 5
Adult premature mortality from long-term exposure
(age 65-99 or age 30-99)
~
~
PM ISA
Infant mortality (age <1)
~
~
PM ISA

Heart attacks (age >18)
~

PM ISA

Hospital admissions—cardiovascular (ages 65-99)
~
~
PM ISA

Emergency department visits— cardiovascular (age
0-99)
~
~
PM ISA

Hospital admissions—respiratory (ages 0-18 and 65-
99)
~
V
PM ISA

Emergency room visits—respiratory (all ages)
~
V
PM ISA

Cardiac arrest (ages 0-99; excludes initial hospital
and/or emergency department visits)
~
~ !
PM ISA

Stroke (ages 65-99)
~

PM ISA

Asthma onset (ages 0-17)
~
~
PM ISA

Asthma symptoms/exacerbation (6-17)
~
~
PM ISA

Lung cancer (ages 30-99)
~
~
PM ISA
Nonfatal
Allergic rhinitis (hay fever) symptoms (ages 3-17)
~
~
PM ISA
morbidity from
Lost work days (age 18-65)
~
~
PM ISA
exposure to PM2 5
Minor restricted-activity days (age 18-65)
~
~
PM ISA

Hospital admissions—Alzheimer's disease (ages 65-
99)
~
V
PM ISA

Hospital admissions—Parkinson's disease (ages 65-
99)
~
V
PM ISA

Other cardiovascular effects (e.g., other ages)
—
—
PM ISA2

Other respiratory effects (e.g., pulmonary function,
non-asthma ER visits, non-bronchitis chronic
diseases, other ages and populations)
—
—
PM ISA2

Other nervous system effects (e.g., autism, cognitive
decline, dementia)
—
—
PM ISA2

Metabolic effects (e.g., diabetes)
—
—
PM ISA2

Reproductive and developmental effects (e.g., low
birth weight, pre-term births, etc.)
—
—
PM ISA2

Cancer, mutagenicity, and genotoxicity effects
—
—
PM ISA2
Mortality from
Premature respiratory mortality from short-term
exposure (0-99)
~
~
Ozone ISA
exposure to ozone
Premature respiratory mortality from long-term
exposure (age 30-99)
~

Ozone ISA

Hospital admissions—respiratory (ages 65-99)
~
~
Ozone ISA

Emergency department visits—respiratory (ages 0-
99)
~
~
Ozone ISA

Asthma onset (0-17)
~
~
Ozone ISA

Asthma symptoms/exacerbation (asthmatics age 5-
17)
~
V
Ozone ISA
Nonfatal
Allergic rhinitis (hay fever) symptoms (ages 3-17)
~
V
Ozone ISA
morbidity from
Minor restricted-activity days (age 18-65)
~
V
Ozone ISA
exposure to ozone
School absence days (age 5-17)
~
V
Ozone ISA

Decreased outdoor worker productivity (age 18-65)
—
—
Ozone ISA2

Metabolic effects (e.g., diabetes)
—
—
Ozone ISA2

Other respiratory effects (e.g., premature aging of
lungs)
—
—
Ozone ISA2

Cardiovascular and nervous system effects
—
—
Ozone ISA2
5-10
-------
Reproductive and developmental effects	—	— Ozone ISA2
'Valuation estimate excludes initial hospital and/or emergency department visits.
2 Not quantified due to data availability limitations and/or because current evidence is only suggestive of causality.
5.2.1.2 Calculating Counts of Air Pollution Effects Using the Health Impact Function
We use BenMAP-CE to quantify individual risk and counts of estimated premature deaths
and illnesses attributable to photochemical modeled changes in summer season average ozone
concentrations for the year 2021, and summer season average ozone concentrations and annual
mean PM2.5 for the year 2024 using a health impact function. A health impact function combines
information regarding the: concentration-response relationship between air quality changes and
the risk of a given adverse outcome; population exposed to the air quality change; baseline rate
of death or disease in that population; and, air pollution concentration to which the population is
exposed.
The following provides an example of a health impact function, in this case for PM2.5
mortality risk. We estimate counts of PM2.5-related total deaths (yij) during each year i (i=l,.. .,1
where I is the total number of years analyzed) among adults aged 30 and older (a) in each county
in the contiguous U.S. j (j=l,.. ,,J where J is the total number of counties) as
yij— 2a yija
yija = moija x(ep ACij-l) x Pija, Eq[l]
where moija is the baseline all-cause mortality rate for adults aged a=30-99 in county j in year i
stratified in 10-year age groups, P is the risk coefficient for all-cause mortality for adults
associated with annual average PM2.5 exposure, Cij is the annual mean PM2.5 concentration in
county j in year i, and Pija is the number of county adult residents aged a=30-99 in county j in
year i stratified into 5-year age groups.8
The BenMAP-CE tool is pre-loaded with projected population from the Woods & Poole
company; cause-specific and age-stratified death rates from the Centers for Disease Control and
8 In this illustrative example, the air quality is resolved at the county level. For this RIA, we simulate air quality
concentrations at 12km by 12km grids. The BenMAP-CE tool assigns the rates of baseline death and disease stored
at the county level to the 12km by 12km grid cells using an area-weighted algorithm. This approach is described in
greater detail in the appendices to the BenMAP-CE user manual.
5-11
-------
Prevention, projected to future years; recent-year baseline rates of hospital admissions,
emergency department visits and other morbidity outcomes from the Healthcare Cost and
Utilization Program and other sources; concentration-response parameters from the published
epidemiologic literature cited in the ISAs for fine particles and ground-level ozone; and, cost of
illness or WTP unit values for each endpoint. Ozone and PM2.5 concentrations are taken from the
air pollution spatial surfaces described in Chapter 3.
5.2.1.3	Quantifying Ozone-Attributable Premature Mortality
In 2008, the National Academies of Science (NRC 2008) issued a series of
recommendations to EPA regarding the procedure for quantifying and valuing ozone-related
mortality due to short-term exposures. Chief among these was that "... short-term exposure to
ambient ozone is likely to contribute to premature deaths" and the committee recommended that
"ozone-related mortality be included in future estimates of the health benefits of reducing ozone
exposures..." The NAS also recommended that".. .the greatest emphasis be placed on the
multicity and [National Mortality and Morbidity Air Pollution Studies (NMMAPS)] ... studies
without exclusion of the meta-analyses" (NRC 2008). Prior to the 2015 Ozone NAAQS RIA, the
Agency estimated ozone-attributable premature deaths using an NMMAPS-based analysis of
total mortality (Bell et al. 2004), two multi-city studies of cardiopulmonary and total mortality
(Huang et al. 2004; Schwartz 2005) and effect estimates from three meta-analyses of non-
accidental mortality (Bell et al. 2005; Ito et al. 2005; Levy et al. 2005). Beginning with the 2015
Ozone NAAQS RIA, the Agency began quantifying ozone-attributable premature deaths using
two newer multi-city studies of non-accidental mortality (Smith et al. 2009; Zanobetti and
Schwartz 2008) and one long-term cohort study of respiratory mortality (Jerrett et al. 2009). The
2020 Ozone ISA included changes to the causality relationship determinations between short-
term exposures and total mortality, as well as including more recent epidemiologic analyses of
long-term exposure effects on respiratory mortality (U.E. EPA, 2020b). In this RIA, as described
in the corresponding TSD, two estimates of ozone-attributable respiratory deaths from short-term
exposures are estimated using the risk estimate parameters from Zanobetti et al. (2008) and
Katsouyanni et al. (2009). Ozone-attributable respiratory deaths from long-term exposures are
estimated using Turner et al. (2016). Due to time and resource limitations, we were unable to
5-12
-------
reflect the warm season defined by Zanobetti et al. (2008) as June-August. Instead, we apply this
risk estimate to our standard warm season of May-September.
5.2.1.4	Quantifying PM2.5-Attributable Premature Mortality
When quantifying PM-attributable cases of adult mortality, we use the effect coefficients
from two epidemiology studies examining two large population cohorts: the American Cancer
Society cohort (Turner et al. 2016) and the Medicare cohort (Di et al. 2017). The Integrated
Science Assessment for Particulate Matter (PM ISA) (U.S. EPA 2019) concluded that the
analyses of the ACS and Medicare cohorts provide strong evidence of an association between
long-term PM2.5 exposure and premature mortality with support from additional cohort studies.
There are distinct attributes of both the ACS and Medicare cohort studies that make them well-
suited to being used in a PM benefits assessment and so here we present PM2.5 related effects
derived using relative risk estimates from both cohorts.
The PM ISA, which was reviewed by the Clean Air Scientific Advisory Committee of
EPA's Science Advisory Board (SAB-CASAC) (EPA-SAB 2020a), concluded that there is a
causal relationship between mortality and both long-term and short-term exposure to PM2.5 based
on the entire body of scientific evidence. The PM ISA also concluded that the scientific literature
supports the use of a no-threshold log-linear model to portray the PM-mortality concentration-
response relationship while recognizing potential uncertainty about the exact shape of the
concentration-response relationship. The 2019 PM ISA, which informed the setting of the 2020
PM NAAQS, reviewed available studies that examined the potential for a population-level
threshold to exist in the concentration-response relationship. Based on such studies, the ISA
concluded that "evidence from recent studies reduce uncertainties related to potential copollutant
confounding and continues to provide strong support for a linear, no-threshold concentration-
response relationship" (U.S. EPA 2019) (section 11.2.7). Consistent with this evidence, the
Agency historically has estimated health impacts above and below the prevailing NAAQS (U.S.
EPA 2010c, 2010d, 2011c, 201 Id, 2012, 2013b, 2014a, 2014b, 2014c, 2015a, 2015b, 2015c,
2015d, 2015e, 2016b).
Following this approach, we report the estimated PM2.5-related benefits (in terms of both
health impacts and monetized values) calculated using a log-linear concentration-response
5-13
-------
function that quantifies risk from the full range of simulated PM2.5 exposures (NRC 2002; U.S.
EPA 2009). When setting the 2020 PM NAAQS, the EPA noted that".. .an important
consideration in characterizing the potential for additional public health improvements associated
with changes in PM2.5 exposure is whether concentration- response relationships are linear across
the range of concentrations or if nonlinear relationships exist along any part of this range.
Several recent studies examine this issue, and continue to provide evidence of linear, no-
threshold relationships between long-term PM2.5 exposures and all-cause and cause- specific
mortality (U.S. EPA, 2019, section 11.2.4). However, interpreting the shapes of these
relationships, particularly at PM2.5 concentrations near the lower end of the air quality
distribution, can be complicated by relatively low data density in the lower concentration range,
the possible influence of exposure measurement error, and variability among individuals with
respect to air pollution health effects (85 FR 82696, December 18, 2020)." Hence, we are most
confident in the size of the risks estimated from simulated PM2.5 concentrations that coincide
with the bulk of the observed PM concentrations in the epidemiological studies that are used to
estimate the benefits. Likewise, we are less confident in the risk we estimate from simulated
PM2.5 concentrations that fall below the bulk of the observed data in these studies.
Hence, we are most confident in the size of the risks estimated from simulated PM2.5
concentrations that coincide with the bulk of the observed PM concentrations in the
epidemiological studies that are used to estimate the benefits. Likewise, we are less confident in
the risk we estimate from simulated PM2.5 concentrations that fall below the bulk of the observed
data in these studies.
To give readers insight to the level of uncertainty in the estimated PM2.5 mortality
benefits at lower ambient concentrations, we report the estimated PM benefits as a distribution,
identifying points along this distribution corresponding to the Lowest Reported Levels of each
long-term exposure mortality study and the PM NAAQS (Figure 5-2). In addition to adult
mortality discussed above, we use effect coefficients from a multi-city study to estimate PM-
related infant mortality (Woodruff et al. 2008).
5-14
-------
5.2.2 Economic Valuation Methodology for Health Benefits
We next quantify the economic value of the ozone and PIVh.s-related deaths and illnesses
estimated above. Changes in ambient concentrations of air pollution generally yield small
changes in the risk of future adverse health effects for a large number of people. Therefore, the
appropriate economic measure is WTP for changes in risk of a health effect. For some health
effects, such as hospital admissions, WTP estimates are not generally available, so we use the
cost of treating or mitigating the effect. These cost-of-illness (COI) estimates are typically a
lower bound estimate of the true value of reducing the risk of a health effect because they reflect
the direct expenditures related to treatment, but not the value of avoided pain and suffering The
unit values applied in this analysis are provided in Table 21 of the Estimating PM2.5- and Ozone-
Attributable Health Benefits TSD.
The estimated value of avoided premature deaths account for over 95 percent of
monetized ozone-related benefits and over 98 percent of monetized PIVh.s-related benefits of this
rule. The economics literature concerning the appropriate method for valuing reductions in
premature mortality risk is still developing. The value for the projected reduction in the risk of
premature mortality is the subject of continuing discussion within the economics and public
policy analysis community. Following the advice of the SAB's Environmental Economics
Advisory Committee (SAB-EEAC), EPA currently uses the value of statistical life (VSL)
approach in calculating estimates of mortality benefits, because we believe this calculation
provides the most reasonable single estimate of an individual's willingness to trade off money
for changes in the risk of death (U.S. EPA-SAB 2000a). The VSL approach is a summary
measure for the value of small changes in the risk of death experienced by a large number of
people.
EPA continues work to update its guidance on valuing mortality risk reductions, and the
Agency consulted several times with the SAB-EEAC on this issue. Until updated guidance is
available, the Agency determined that a single, peer-reviewed estimate applied consistently, best
reflects the SAB-EEAC advice it has received. Therefore, EPA applies the VSL that was vetted
and endorsed by the SAB in the Guidelines for Preparing Economic Analyses (U.S. EPA 2016a)
while the Agency continues its efforts to update its guidance on this issue. This approach
calculates a mean value across VSL estimates derived from 26 labor market and contingent
5-15
-------
valuation studies published between 1974 and 1991. The mean VSL across these studies is $4.8
million (1990$). We then adjust this VSL to account for the currency year and to account for
income growth from 1990 to the analysis year. Specifically, the VSLs applied in this analysis in
2016$ after adjusting for income growth is $10.5 million for 2021 and $10.7 million for 2024.
The Agency is committed to using scientifically sound, appropriately reviewed evidence
in valuing changes in the risk of premature death and continues to engage with the SAB to
identify scientifically sound approaches to update its mortality risk valuation estimates. In 2016,
the Agency proposed new meta-analytic approaches for updating its estimates (U.S. EPA-SAB
2017), which were subsequently reviewed by the SAB-EEAC. EPA is taking the SAB's formal
recommendations under advisement.
In valuing PM2.5-related premature mortality, we discount the value of premature
mortality occurring in future years using rates of 3 percent and 7 percent (U.S. Office of
Management and Budget 2003). We assume that there is a multi-year "cessation" lag between
changes in PM exposures and the total realization of changes in health effects. Although the
structure of the lag is uncertain, EPA follows the advice of the SAB-HES to use a segmented lag
structure that assumes 30 percent of premature deaths are reduced in the first year, 50 percent
over years 2 to 5, and 20 percent over the years 6 to 20 after the reduction in PM2.5 (U.S. EPA-
SAB 2004). Changes in the cessation lag assumptions do not change the total number of
estimated deaths but rather the timing of those deaths.
Because short-term ozone-related premature mortality occurs within the analysis year, the
estimated ozone-related benefits are identical for all discount rates. When valuing changes in
ozone-attributable deaths using the Turner et al. (2016) study, we follow advice provided by the
Health Effects Subcommittee of the SAB, which found that".. .there is no evidence in the
literature to support a different cessation lag between ozone and particulate matter. The HES
therefore recommends using the same cessation lag structure and assumptions as for particulate
matter when utilizing cohort mortality evidence for ozone" (U.S. EPA-SAB 2010).
These estimated health benefits do not account for the influence of future changes in the
climate on ambient concentrations of pollutants (USGCRP 2016). For example, recent research
suggests that future changes to climate may create conditions more conducive to forming ozone;
5-16
-------
the influence of changes in the climate on PM2.5 concentrations are less clear (Fann et al. 2015).
The estimated health benefits also do not consider the potential for climate-induced changes in
temperature to modify the relationship between ozone and the risk of premature death (Jhun et al.
2014; Ren et al. 2008a, 2008b).
5.2.3 Characterizing Uncertainty in the Estimated Benefits
This analysis includes many data sources as inputs that are each subject to uncertainty.
Input parameters include projected emission inventories, projected compliance methods and
emissions from the electricity sector analysis, air quality data from models (with their associated
parameters and inputs), population data, population estimates, health effect estimates from
epidemiology studies, economic data, and assumptions regarding the future state of the world
(i.e., regulations, technology, and human behavior). When compounded, even small uncertainties
can greatly influence the size of the total quantified benefits.
Our estimate of the total monetized PM2.5 and ozone-attributable benefits is based on
EPA's interpretation of the best available scientific literature and methods and supported by the
SAB-HES and the National Academies of Science (NRC 2002). Below are key assumptions
underlying the estimates for PM2.5-related premature mortality, followed by key uncertainties
associated with estimating the number and value of ozone-related premature deaths.
We assume that all fine particles, regardless of their chemical composition, are equally
potent in causing premature mortality. This is an important assumption, the PM ISA concluded
that "many constituents of PM2.5 can be linked with multiple health effects, and the evidence is
not yet sufficient to allow differentiation of those constituents or sources that are more closely
related to specific outcomes" (U.S. EPA 2009).
As noted above, we assume that the health impact function for fine particles is log-linear
without a threshold. Thus, the estimates include health benefits from reducing fine particles in
areas with different concentrations of PM2.5, including both areas that do not meet the fine
particle standard and those areas that are in attainment and reflect the full distribution of PM2.5
air quality simulated above.
5-17
-------
Also, as noted above, we assume that there is a "cessation" lag between the change in PM
exposures and the total realization of changes in mortality effects. Specifically, we assume that
some of the incidences of premature mortality related to PM2.5 exposures occur in a distributed
fashion over the 20 years following exposure based on the advice of the SAB-HES (U.S. EPA-
SAB 2004), which affects the valuation of mortality benefits at different discount rates. The
above assumptions are subject to uncertainty.
In general, we are more confident in the magnitude of the risks we estimate from
simulated PM2.5 concentrations that coincide with the bulk of the observed PM concentrations in
the epidemiological studies that are used to estimate the benefits. Likewise, we are less confident
in the risk we estimate from simulated PM2.5 concentrations that fall below the bulk of the
observed data in these studies. There are uncertainties inherent in identifying any particular point
at which our confidence in reported associations decreases appreciably, and the scientific
evidence provides no clear dividing line. This relationship between the air quality data and our
confidence in the estimated risk is represented below in Figure 5-.
Less confident	More
confident
A_			
Below LRL of PM2.5 data
in epidemiology study
(extrapolation)
I standard deviation below the
mean PM2.5 observed in
epidemiology study
Mean of PM2.5 data in
epidemiology study
Figure 5-1. Stylized Relationship between the PM2.5 Concentrations Considered in
Epidemiology Studies and our Confidence in the Estimated PM-related
Premature Deaths
In this analysis, we build upon the concentration benchmark approach (also referred to as
the Lowest Reported Level (LRL) analysis) that has been featured in recent RIAs by reporting
the estimated PM-related deaths according to alternative concentration cut points.
Concentration benchmark analyses allow readers to determine the portion of population
exposed to annual mean PM2.5 levels at or above different concentrations, which provides some
5-18
-------
insight into the level of uncertainty in the estimated PM2.5 mortality benefits. EPA does not view
these concentration benchmarks as concentration thresholds below which we would not quantify
health benefits of air quality improvements.9 Rather, the PIVh.s-attributable benefits estimates
reported in this RIA are the most appropriate estimates because they reflect the full range of air
quality concentrations associated with the emission reduction strategies being evaluated in this
final rule. The 2019 PM ISA concluded that the scientific evidence collectively is sufficient to
conclude that there is a causal relationship between long-term PM2.5 exposures and mortality and
that overall the studies support the use of a no-threshold log-linear model to estimate mortality
attributed to long-term PM2.5 exposure (U.S. EPA 2009).
Figure 5-2 compares the percentage of the population and PM-related deaths, to the
annual mean PM2.5 concentrations in the final policy modeling for the year 2024. The figure
identifies the LRL for each of the major cohort studies and the annual mean PM2.5 NAAQS of 12
|ig/m3. For Turner et al. 2016, the LRL is 2.8 |ig/m3 and for Di et al. 2017, the LRL is 0.02
|ig/m3. As PM-related mortality quantified using risk estimates from the Di et al. (2017) and
Turner et al. (2016) are within 5% of one another, in the interest of clarity and simplicity, we
present the results estimated using the risk estimate from Turner et al. (2016) alone in Figure 5-2.
Additional information on low concentration exposures in Turner et al. 2016 and Di et al. 2017
can be found in section 6.1.2.1 of the Estimating PM2.5- and Ozone-Attributable Health Benefits
TSD. These results are sensitive to the annual mean PM2.5 concentration the air quality model
predicted in each 12km by 12km grid cell. The air quality modeling predicts PM2.5
concentrations to be at or below the PM2.5 NAAQS (12 |ig/m3) in nearly all locations. The
photochemical modeling we employ accounts for the suite of local, state and federal policies
expected to reduce PM2.5 and PM2.5 precursor emissions in future years, such that we project a
very small number of locations exceeding the annual standard. The results should be viewed in
the context of the air quality modeling technique we used to estimate PM2.5 concentrations. We
are more confident in our ability to use the air quality modeling technique described above to
estimate changes in annual mean PM2.5 concentrations than we are in our ability to estimate
absolute PM2.5 concentrations.
9 For a summary of the scientific review statements regarding the lack of a threshold in the PM2 5-mortality
relationship, see the TSD entitled Summary of Expert Opinions on the Existence of a Threshold in the
Concentration-Response Function for PM2.5-related Mortality (U.S. EPA, 2010b).
5-19
-------
LRL
Di (2017)
LRL
Turner (2016)
2012
PM2 5 Annual NAAQS
05
Population
Attributable deaths

7S04
02
CL
01
00
0
5
10
PM2 5 (fig/m3)
Figure 5-2. Estimated Percentage of PM2,5-Related Deaths and Number of Individuals
Exposed by Annual Mean PM2.5 Level in 2024
The estimated number and value of avoided ozone-attributable deaths are also subject to
uncertainty. When estimating the economic value of avoided premature mortality from long-term
exposure to ozone, we use a 20-year segment lag (as used for PM2.5) as there is no alternative
empirical estimate of the cessation lag for long-term exposure to ozone. The 20-year segmented
lag accounts for the onset of cardiovascular related mortality, an outcome which is not relevant
to the long-term respiratory mortality estimated here. We use a log-linear impact function
without a threshold in modeling short-term ozone-related mortality. However, we acknowledge
reduced confidence in specifying the shape of the concentration-response relationship in the
range of < 40ppb and below (2020 Ozone ISA, section 6.2.6). Thus, the estimates include health
benefits from reducing ozone in areas with varied concentrations of ozone down to the lowest
modeled concentrations.
5-20
-------
5.2.4 Estimated Number and Economic Value of Health Benefits
Below we report the estimated number of reduced premature deaths and illnesses in each year
relative to the baseline along with the 95% confidence interval (Table 5-2 and Table 5-3). The
number of reduced estimated deaths and illnesses from the final rule and more and less stringent
alternatives are calculated from the sum of individual reduced mortality and illness risk across
the population. Table 5-4 and Table 5-5 report the estimated economic value of avoided
premature deaths and illness in each year relative to the baseline along with the 95% confidence
interval. In each of these tables, for each discount rate and regulatory control alternative,
multiple benefits estimates are presented reflecting alternative ozone and PM2.5 mortality risk
estimates. We also report the stream of benefits from 2021 through 2040 for the final, more- and
less- stringent scenarios, using the monetized sums of long-term ozone and PM2.5 mortality and
morbidity impacts (Table 5-6 and Table 5-7).10
10 EPA continues to refine its approach for estimating and reporting PM-related effects at lower concentrations. The
Agency acknowledges the additional uncertainty associated with effects estimated at these lower levels and seeks to
develop quantitative approaches for reflecting this uncertainty in the estimated PM benefits.
5-21
-------
Table 5-2. Estimated Avoided Ozone-Related Premature Respiratory Mortality and
Illnesses for the Final and More and Less Stringent Alternatives for 2021
	(95% Confidence Interval) a,b	


Final
More Stringent
Alternative
Less Stringent
Alternative
Avoided premature respiratory mortality
Long-
term
Turner etal. (2016)°
190
190
18
exposure

(130 to 250)
(130 to 250)
(13 to 24)

Katsouyanni el al.
(2009)cd and Zanobetti el
9
9
1
Short-
al. (2008)d pooled
(4 to 14)
(0 to 14)
(0 to 1)
term
exposure
Katsouyanni el al.
(2009)cd
9
(-5 to 21)
9
(-5 to 21)
1
(0 to 2)

Zanobetti et al. (2008)e
9
(4 to 13)
9
(4 to 13)
1
(0 to 1)
Morbidity effects
Long-
term
exposure
Asthma onset"
1,300
(1,100 to 1,500)
1,300
(1,100 to 1,500)
130
(110 to 150)
Allergic rhinitis
symptoms8
7,700
(4,000 to 11,000)
7,700
(4,000 to 11,000)
750
(400 to 1,100)

Hospital admissions—
21
21
2

respiratoryd
(-5 to 46)
(-5 to 46)
(-1 to 5)
Short-
term
ED visits—respiratoryf
440
(120 to 920)
440
(120 to 920)
43
(12 to 90)
Asthma symptoms
240,000
(-30,000 to 500,000)
240,000
(-30,000 to 500,000)
24,000
(-2,900 to 49,000)
exposure
Minor restricted-activity
120,000
120,000
12,000

daysd-f
(49,000 to 200,000)
(49,000 to 200,000)
(4,800 to 19,000)

School absence days
91,000
(-13,000 to 190,000)
91,000
(-13,000 to 190,000)
8,900
(-1,300 to 19,000)
a Values rounded to two significant figures.
b We estimated changes in annual mean PM25 and PM2 5 -related benefits in 2024, but not 2021. As discussed in
Chapter 4, in 2021, the only control measure expected to be adopted for compliance in the regulatory control
alternatives analysis in this RIA is optimization of existing SCRs, and this measure will operate only during the
ozone season. As discussed in Chapter 3, NOx reductions in the ozone season provide minimal PM2 5 benefits since
PM2 5 nitrate concentrations, which result from conversion of NOx emissions to nitrate, are minimal during the
warmer temperatures during the ozone season. Conversely, the conversion of nitrates to PM2 5 is much greater in
cooler (non-ozone season) months, and thus it becomes worthwhile to estimate PM2 5 benefits from NOx reductions
in those months. In 2024, the presence of additional control measures that operate year-round and other changes in
market conditions as a result of the rule lead to notable NOx reductions in the winter months.
0 Applied risk estimate derived from April-September exposures to estimates of O3 across the standard May-
September warm season.
d Converted O3 risk estimate metric from MDA1 to MDA8.
e Applied risk estimate derived from June-August exposures to estimates of O3 across the standard May-September
warm season.
f Applied risk estimate derived from full year exposures to estimates of O3 across the standard May-September warm
season.
g Converted O3 risk estimate metric from DA24 to MDA8.
5-22
-------
Table 5-3. Estimated Avoided PM2.5 and Ozone-Related Mortality and Illnesses for the
Final and More and Less Stringent Alternatives for 2024 (95% Confidence
	Interval) a	


Final
More Stringent
Alternative
Less Stringent
Alternative
Avoided premature mortality



in
^ Long-term
Turner et al. (2016)b
4
(2 to 5)
4
(2 to 5)
	
^ exposure
Di et al. (2017)
4
(3 to 4)
4
(3 to 4)

Long-term
exposure
Turner etal. (2016)
230
(160 to 300)
410
(280 to 530)
19
(13 to 25)
ss
0
N
O Short-term -
exposure
Zanobetti et al (2008)° and
Katsouyanni et al (2009)b d
pooled
10
(4 to 16)
18
(7 to 29)
1
(0 to 1)
Katsouyanni et al (2009)b d
10
(-6 to 26)
18
(-10 to 46)
1
(-0 to 2)

Zanobetti et al (2008)°
10
(5 to 16)
18
(8 to 28)
1
(0 to 1)
PM2.5- related non-fatal heart attacks among adults


Short-term
Peters et al. (2001)
4
(1 to 6)
4
(1 to 6)

exposure
Pooled estimate
0
(0 to 1)
0
(0 to 1)

Morbidity effects
Asthma onsetbd
(PM2 5 & 03)
1,600
(1,400 to 1,800)
2,900
(2,500 to 3,300)
130
(110 to 150)

Allergic rhinitis
symptoms6
(PM2.5 & 03)
9,200
(4,900 to 13,000)
17,000
(8,700 to 24,000)
770
(400 to 1,100)
Long-term
exposure
Stroke (PM25)
0.2
(0 to 0.3)
0.2
(0 to 0.3)

Lung cancer (PM2 5)
0.4
(0.0 to 0.7)
0.4
(0.0 to 0.7)


Hospital Admissions -
Alzheimer's disease
(PM25)
2
(1 to 2)
2
(1 to 2)


Hospital Admissions-
Parkinson's disease
(PM2.5)
0.2
(0.1 to 0.3)
0.2
(0.1 to 0.3)


Hospital admissions-
cardiovascular (PM2 5)
0.5
(0.3 to 0.6)
0.5
(0.3 to 0.6)


ED visits- cardiovascular
(PM25)
1
(-0 to 2)
1
(0 to 2)

Short-term
Hospital admissions—
respiratory11 (PM2 5 & O3)
27
(-7 to 60)
49
(-12 to 110)
2
(-1 to 5)
exposure
ED visits —respiratory
(PM2 5 & 03)
530
(150 to 1,100)
950
(260 to 2,000)
44
(12 to 92)

Asthma symptomsf
(PM2 5 & 03)
290,000
(-37,000 to
610,000)
530,000
(-66,000 to
1,100,000)
25,000
(-3,000 to 51,000)
5-23
-------
Minor restricted-activity
days (PM2 5 & O3)
150,000
(60,000 to 230,000)
260,000
(110,000 to
410,000)
12,000
(4,800 to 19,000)
Cardiac arrest (PM2 5)
0.05
(-0.02 to 0.11)
0.05
(-0.02 to 0.11)
--
Lost work days
380
380
--
(PM25)
(320 to 440)
(320 to 440)

School absence days
(03)
110,000
(-15,000 to
200,000
(-28,000 to
9,100
230,000)
410,000)
(-1,300 to 19,000)
a Values rounded to two significant figures.
b Applied risk estimate derived from April-September exposures to estimates of O3 across the standard May-
September warm season.
0 Converted O3 risk estimate metric from MDA1 to MDA8
d Applied risk estimate derived from June-August exposures to estimates of O3 across the standard May-September
warm season.
e Converted O3 risk estimate metric from DA24 to MDA8
f Applied risk estimate derived from full year exposures to estimates of O3 across the standard May-September warm
season.
Table 5-4. Estimated Discounted Economic Value of Ozone-Attributable Premature
Mortality and Illnesses for the Final Policy Scenarios in 2021 (95% Confidence
	Interval; millions of 2016$)a'b	


Final

More Stringent Alternative
Less Stringent Alternative
3% Discount
Rate
$230
($58 to
$480)°
and
$1,900
($210 to
$5,000)d
| $230
| ($58 to
| $480)c
and
$1,900
($210 to
$5,000)d
) $22
1 ($6 to
| $47)°
and
$190
($20 to
$490)d
7% Discount
Rate
$200
($38 to
$460)°
and
$1,700
($170 to
$4,500)d
| $200
| ($38 to
1 $460)°
1
and
$1,700
($170 to
$4,500)d
|
|
1 $20
I ($4 to
1 $45)°
and
$170
($17 to
$440)d
a Values rounded to two significant figures. The two benefits estimates separated by the word "and" signify that they
are two separate estimates. The estimates do not represent lower- and upper-bound estimates and should not be
summed.
b We estimated changes in annual mean PM2 5 and PM2 5 -related benefits in 2024, but not 2021. As discussed in
Chapter 4, in 2021, the only control measure expected to be adopted for compliance in the regulatory control
alternatives analysis in this RIA is optimization of existing SCRs, and this measure will operate only during the
ozone season. As discussed in Chapter 3, NOx reductions in the ozone season provide minimal PM2 5 benefits since
PM2 5 nitrate concentrations, which result from conversion of NOx emissions to nitrate, are minimal during the
warmer temperatures during the ozone season. Conversely, the conversion of nitrates to PM2 5 is much greater in
cooler (non-ozone season) months, and thus it becomes worthwhile to estimate PM2 5 benefits from NOx reductions
in those months. In 2024, the presence of additional control measures that operate year-round and other changes in
market conditions as a result of the rule lead to notable NOx reductions in the winter months.
0 Sum of ozone mortality estimated using the pooled Katsouyanni et al. (2009) and Zanobetti and Schwartz (2008)
short-term risk estimate and the Di et al. (2017) long-term mortality risk estimate. As PM-related mortality
quantified using risk estimates from the Di et al. (2017) and Turner et al. (2016) are within 5% of one another, in the
interest of clarity and simplicity, we present the results estimated using the risk estimate from Di et al. (2017) alone.
d Sum of ozone mortality estimated using the long-term risk estimate and the Di et al. (2017) long-term mortality
risk estimate. PM-related mortality quantified using risk estimates from the Di et al. (2017) and Turner et al. (2016)
are within 5% of one another. In the interest of clarity and simplicity, we present the results estimated using the risk
estimate from Di et al. (2017) alone.
5-24
-------
Table 5-5. Estimated Discounted Economic Value of Avoided Ozone and PM2.5-
Attributable Premature Mortality and Illnesses for the Final Policy Scenario
	in 2024 (95% Confidence Interval; millions of 2016$)a	
Final
More Stringent Alternative
Less Stringent
Alternative1"
3%
$310
$2,400
$530

$4,200
$22
$190
Discount
($72 to
and ($250 to
($130 to
and
($450 to
($6 to and
($20 to
Rate
$680)°
$6,200)d
$l,100)c

$ll,000)d
$47)°
$490)d
7%
Discount
Rate
$280
($48 to
$640)°
$2,100
and ($210 to
$5,600)d
$470
($84 to
$l,100)c
and
$3,800
($370 to
$9,900)d
$20 and
($4 to
$45)°
$170
($17 to
$440)d
a Values rounded to two significant figures. The two benefits estimates separated by the word "and" signify that they
are two separate estimates. The estimates do not represent lower- and upper-bound estimates and should not be
summed.
b No PM-attributable benefits accrue for this scenario
0 Sum of ozone mortality estimated using the pooled Katsouyanni et al. (2009) and Zanobetti and Schwartz (2008)
short-term risk estimate and the Di et al. (2017) long-term mortality risk estimate. As PM-related mortality
quantified using risk estimates from the Di et al. (2017) and Turner et al. (2016) are within 5% of one another, in the
interest of clarity and simplicity, we present the results estimated using the risk estimate from Di et al. (2017). alone.
d Sum of ozone mortality estimated using the long-term risk estimate and the Di et al. (2017) long-term mortality
risk estimate. PM-related mortality quantified using risk estimates from the Di et al. (2017) and Turner et al. (2016)
are within 5% of one another. In the interest of clarity and simplicity, we present the results estimated using the risk
estimate from Di et al. (2017) alone.
5-25
-------
Table 5-6. Stream of Human Health Benefits from 2021 through 2040: Monetized
Benefits Quantified as Sum of Long-Term Ozone Mortality and Long-Term
PM2.5 Mortality (Discounted at 3%; 95% Confidence Interval; millions of
2016$)	

Final
More Stringent Alternative
Less Stringent Alternative
2021*
$1,900
$1,900
$190
2022
$2,000
$2,000
$190
2023
$2,000
$2,000
$200
2024*
$2,400
$4,200
$190
2025
$2,400
$4,200
$200
2026
$2,500
$4,300
$200
2027
$2,400
$4,200
$190
2028
$2,500
$4,300
$200
2029
$2,500
$4,400
$210
2030
$2,600
$4,600
$210
2031
$2,600
$4,600
$210
2032
$2,700
$4,800
$220
2033
$2,600
$4,700
$220
2034
$2,700
$4,800
$220
2035
$2,800
$4,900
$230
2036
$2,800
$5,000
$230
2037
$2,900
$5,100
$230
2038
$2,800
$5,000
$230
2039
$2,800
$5,000
$230
2040
$2,900
$5,100
$240
Net Present Value
$37,000
$61,000
$3,100
*Year in which air quality was simulated. Ozone air quality was simulated in 2021 and 2024 while the formation of
PM2.5 was simulated only in 2024. Health benefits for all other years were linearly extrapolated or interpolated from
model-simulated air quality in these years. This method assumes that ozone and PM2.5 formation reaches a steady
state beyond 2024. Benefits calculated as value of avoided: PM2 5-attributable deaths (quantified using a
concentration-response relationship from the Di et al. 2017 study); Ozone-attributable deaths (quantified using a
concentration-response relationship from the Turner et al. 2017 study); and, PM2 5 and ozone-related morbidity
effects.
5-26
-------
Table 5-7. Stream of Human Health Benefits from 2021 through 2040: Monetized
Benefits Quantified as Sum of Short-Term Ozone Mortality and Long-Term
PM2.5 Mortality (Discounted at 7%; 95% Confidence Interval; millions of
2016$)	

Final
More Stringent Alternative
Less Stringent Alternative
2021*
$200
$200
$20
2022
$210
$210
$21
2023
$210
$210
$20
2024*
$280
$470
$21
2025
$290
$490
$21
2026
$290
$500
$22
2027
$300
$520
$22
2028
$310
$530
$23
2029
$320
$550
$24
2030
$330
$560
$25
2031
$340
$580
$25
2032
$350
$600
$26
2033
$360
$620
$27
2034
$370
$630
$28
2035
$380
$650
$28
2036
$390
$670
$29
2037
$400
$690
$30
2038
$410
$710
$31
2039
$430
$730
$32
2040
$440
$750
$33
Net Present Value
$3,200
$5,100
$250
*Year in which air quality was simulated. Ozone air quality was simulated in 2021 and 2024 while the formation of
PM2.5 was simulated only in 2024. Health benefits for all other years were linearly extrapolated or interpolated from
model-simulated air quality in these years. This method assumes that ozone and PM2.5 formation reaches a steady
state beyond 2024. Benefits calculated as value of avoided: PM2 5-attributable deaths (quantified using a
concentration-response relationship from the Di et al. 2017 study); Ozone-attributable deaths (quantified using a
concentration-response relationship from the Turner et al. 2017 study); and, PM2 5 and ozone-related morbidity
effects.
5.2 Estimated Climate Benefits from Reducing CO2
We estimate the climate benefits for this final rulemaking using a measure of the social
cost of carbon (SC-CO2). The SC-CO2 is the monetary value of the net harm to society
associated with a marginal increase in CO2 emissions in a given year, or the benefit of avoiding
that increase. In principle, SC-CO2 includes the value of all climate change impacts, including
(but not limited to) changes in net agricultural productivity, human health effects, property
damage from increased flood risk and natural disasters, disruption of energy systems, risk of
conflict, environmental migration, and the value of ecosystem services. The SC-CO2, therefore,
5-27
-------
reflects the societal value of reducing emissions of the gas in question by one metric ton. The
SC-CO2 is the theoretically appropriate values to use in conducting benefit-cost analyses of
policies that affect CO2 emissions.
We estimate the global social benefits of CO2 emission reductions expected from this
final rule using the SC-CO2 estimates presented in the Technical Support Document: Social Cost
of Carbon, Methane, and Nitrous Oxide Interim Estimates under Executive Order 13990 (IWG
2021). These SC-CO2 estimates are interim values developed under Executive Order (E.O.)
13990 for use in benefit-cost analyses until an improved estimate of the impacts of climate
change can be developed based on the best available science and economics. These SC-CO2
estimates are the same as those used in the 2016 Final CSAPR Update RIA.
The SC-CO2 estimates presented here were developed over many years, using
transparent process, peer-reviewed methodologies, the best science available at the time of that
process, and with input from the public. Specifically, in 2009, an interagency working group
(IWG) that included the EPA and other executive branch agencies and offices was established to
ensure that agencies were using the best available science and to promote consistency in the SC-
CO2 values used across agencies. The IWG published SC-CO2 estimates in 2010 that were
developed from an ensemble of three widely cited integrated assessment models (IAMs) that
estimate global climate damages using highly aggregated representations of climate processes
and the global economy combined into a single modeling framework. The three IAMs were run
using a common set of input assumptions in each model for future population, economic, and
CO2 emissions growth, as well as equilibrium climate sensitivity (ECS) - a measure of the
globally averaged temperature response to increased atmospheric CO2 concentrations. These
estimates were updated in 2013 based on new versions of each IAM.11 In 2015, as part of the
response to public comments received to a 2013 solicitation for comments on the SC-CO2
estimates, the IWG announced a National Academies of Sciences, Engineering, and Medicine
review of the SC-CO2 estimates to offer advice on how to approach future updates to ensure that
the estimates continue to reflect the best available science and methodologies. In January 2017,
the National Academies released their final report, Valuing Climate Damages: Updating
11 Dynamic Integrated Climate and Economy (DICE) 2010 (Nordhaus 2010), Climate Framework for Uncertainty,
Negotiation, and Distribution (FUND) 3.8 (Anthoff and Tol 2013a, 2013b), and Policy Analysis of the Greenhouse
Gas Effect (PAGE) 2009 (Hope 2013).
5-28
-------
Estimation of the Social Cost of Carbon Dioxide, and recommended specific criteria for future
updates to the SC-CO2 estimates, a modeling framework to satisfy the specified criteria, and both
near-term updates and longer-term research needs pertaining to various components of the
estimation process (National Academies 2017). Shortly thereafter, in March 2017, President
Trump issued Executive Order 13783, which disbanded the IWG, withdrew the previous TSDs,
and directed agencies to ensure SC-CO2 estimates used in regulatory analyses are consistent with
the guidance contained in OMB's Circular A-4, "including with respect to the consideration of
domestic versus international impacts and the consideration of appropriate discount rates" (E.O.
13783, Section 5(c)). Benefit-cost analyses following E.O. 13783, including the benefit-cost
analysis in the proposal RIA,12 used SC-CO2 estimates that attempted to focus on the domestic
impacts of climate change as estimated by the models to occur within U.S. borders and were
calculated using two discount rates recommended by Circular A-4, 3 percent and 7 percent. All
other methodological decisions and model versions used in SC- CO2 calculations remained the
same as those used by the IWG in 2010 and 2013, respectively.
On January 20, 2021, President Biden issued Executive Order 13990, which re-
established the IWG and directed it to ensure that the U.S. Government's estimates of the social
cost of carbon, methane, and nitrous oxide (collectively referred to as SC-GHG) reflect the best
available science and the recommendations of the National Academies (2017). The IWG was
tasked with first reviewing the SC-GHG estimates currently used in Federal analyses and
publishing interim estimates within 30 days of the E.O. that reflect the full impact of GHG
emissions, including by taking global damages into account. The interim SC-CO2 estimates
published in February 2021 are used here to estimate the climate benefits for this final
rulemaking. The E.O. instructs the IWG to undertake a fuller update of the SC-GHG estimates
by January 2022 that takes into consideration the advice of the National Academies (2017) and
other recent scientific literature.
The February 2021 TSD provides a complete discussion of the IWG's initial review
conducted under E.O. 13990. In particular, the IWG found that the SC-GHG estimates used since
E.O. 13783 fail to reflect the full impact of GHG emissions in multiple ways. First, the IWG
12 The values used in the proposal RIA were interim values developed under E.O. 13783 for use in regulatory
analyses. EPA followed E.O. 13783 by using SC-C02 estimates reflecting impacts occurring within U.S. borders
and 3% and 7% discount rates in our central analysis for the proposal RIA.
5-29
-------
found that a global perspective is essential for SC-GHG estimates because climate impacts
occurring outside U.S. borders can directly and indirectly affect the welfare of U.S. citizens and
residents. Thus, U.S. interests are affected by the climate impacts that occur outside U.S.
borders. Examples of affected interests include: direct effects on U.S. citizens and assets located
abroad, international trade, and tourism, and spillover pathways such as economic and political
destabilization and global migration. In addition, assessing the benefits of U.S. GHG mitigation
activities requires consideration of how those actions may affect mitigation activities by other
countries, as those international mitigation actions will provide a benefit to U.S. citizens and
residents by mitigating climate impacts that affect U.S. citizens and residents. Therefore, in this
final rule EPA centers attention on a global measure of SC-GHG. This approach is the same as
that taken in EPA regulatory analyses over 2009 through 2016. As noted in the February 2021
TSD, the IWG will continue to review developments in the literature, including more robust
methodologies for estimating SC-GHG values based on purely domestic damages, and explore
ways to better inform the public of the full range of carbon impacts, both global and domestic.
As a member of the IWG, EPA will likewise continue to follow developments in the literature
pertaining to this issue.
Second, the IWG found that the use of the social rate of return on capital (7 percent under
current OMB Circular A-4 guidance) to discount the future benefits of reducing GHG emissions
inappropriately underestimates the impacts of climate change for the purposes of estimating the
SC-GHG. Consistent with the findings of the National Academies (2017) and the economic
literature, the IWG continued to conclude that the consumption rate of interest is the theoretically
appropriate discount rate in an intergenerational context (IWG 2010, 2013, 2016a, 2016b), and
recommended that discount rate uncertainty and relevant aspects of intergenerational ethical
considerations be accounted for in selecting future discount rates. As a member of the IWG
involved in the development of the February 2021 TSD, EPA agrees with this assessment, and
will continue to follow developments in the literature pertaining to this issue.
While the IWG works to assess how best to incorporate the latest, peer reviewed science
to develop an updated set of SC-GHG estimates, it set the interim estimates to be the most recent
estimates developed by the IWG prior to the group being disbanded in 2017. The estimates rely
on the same models and harmonized inputs and are calculated using a range of discount rates. As
explained in the February 2021 TSD, the IWG has determined that it is appropriate for agencies
5-30
-------
to revert to the same set of four values drawn from the SC-GHG distributions based on three
discount rates as were used in regulatory analyses between 2010 and 2016 and subject to public
comment. For each discount rate, the IWG combined the distributions across models and
socioeconomic emissions scenarios (applying equal weight to each) and then selected a set of
four values for use in benefit-cost analyses: an average value resulting from the model runs for
each of three discount rates (2.5%, 3%, and 5%), plus a fourth value, selected as the 95th
percentile of estimates based on a 3 percent discount rate. The fourth value was included to
provide information on potentially higher-than-expected economic impacts from climate change,
conditional on the 3% estimate of the discount rate. As explained in the February 2021 TSD, this
update reflects the immediate need to have an operational SC-GHG for use in regulatory benefit-
cost analyses and other applications that was developed using a transparent process, peer-
reviewed methodologies, and the science available at the time of that process. Those estimates
were subject to public comment in the context of dozens of proposed rulemakings as well as in a
dedicated public comment period in 2013.
Table 5-8 summarizes the interim global SC-CO2 estimates for the years 2015 to 2050.
These estimates are reported in 2016 dollars but are otherwise identical to those presented in the
IWG's 2016 TSD (IWG 2016a). For purposes of capturing uncertainty around the SC-CO2
estimates in analyses, the IWG's February 2021 TSD emphasizes the importance of considering
all four of the SC-CO2 values. The SC-CO2 increases over time within the models - i.e., the
societal harm from one metric ton emitted in 2030 is higher than the harm caused by one metric
ton emitted in 2025 - because future emissions produce larger incremental damages as physical
and economic systems become more stressed in response to greater climatic change, and because
GDP is growing over time and many damage categories are modeled as proportional to GDP.
Table 5-8. Interim Global Social Cost of Carbon Values, 2020-2050 (2016$/Metric Tonne
CO2)
Emissions

Discount Rate and Statistic

Year





5%
3%
2.5%
3%

Average
Average
Average
95th Percentile
2020
$13
$47
$71
$140
2025
$15
$52
$77
$160
2030
$18
$57
$83
$170
2035
$20
$63
$90
$190
2040
$23
$67
$95
$210
2045
$26
$73
$100
$220
2050
$29
$78
$110
$240
5-31
-------
Note: These SC-CO2 values are identical to those reported in the 2016 TSD (IWG 2016a) adjusted for inflation to
2016 dollars using the annual GDP Implicit Price Deflator values in the U.S. Bureau of Economic Analysis'
(BEA) NIPA Table 1.1.9 (U.S. BEA 2021). The values are stated in $/metric tonne CO2 (1 metric tonne equals
1.102 short tons) and vary depending on the year of CO2 emissions. This table displays the values rounded to the
nearest dollar; the annual unrounded values used in the calculations in this RIA are available on OMB's website:
.
Source: Technical Support Document: Social Cost of Carbon, Methane, and Nitrous Oxide Interim Estimates
under Executive Order 13990 (IWG 2021)
There are a number of limitations and uncertainties associated with the SC-CO2 estimates
presented in Table 5-8. Some uncertainties are captured within the analysis, while other areas of
uncertainty have not yet been quantified in a way that can be modeled. Figure 5-3 presents the
quantified sources of uncertainty in the form of frequency distributions for the SC-CO2 estimates
for emissions in 2030. The distributions of SC-CO2 estimates reflect uncertainty in key model
parameters such as the equilibrium climate sensitivity, as well as uncertainty in other parameters
set by the original model developers. To highlight the difference between the impact of the
discount rate and other quantified sources of uncertainty, the bars below the frequency
distributions provide a symmetric representation of quantified variability in the SC-CO2
estimates for each discount rate. As illustrated by the figure, the assumed discount rate plays a
critical role in the ultimate estimate of the SC-CO2. This is because CO2 emissions today
continue to impact society far out into the future, so with a higher discount rate, costs that accrue
to future generations are weighted less, resulting in a lower estimate. As discussed in the
February 2021 TSD, there are other sources of uncertainty that have not yet been quantified and
are thus not reflected in these estimates.
5-32
-------
CO
•4—
o
o
LO
CM —
°S
O
LO
O
LO
O _J
5% Average = $18
Discount Rate
~	5.0%
~	3.0%
~	2.5%
3% Average = $57
2.5% Average = $83
3%
95th Pet. = $174
nl
fiQEBBBBi
IBBBBBdhh L
~

~
}
I I
I I I
20
40
I I I
60
of Simulations
Ml ii ii
I ii I i i I I ii ii ii ii I i i I i I I ill
80 100 120 140 160 180 200 220 240 260 280 300
Social Cost of Carbon in 2030 [2016$ / metric ton C02]
Figure 5-3. Frequency Distribution of SC-CO2 Estimates for 203013
The interim SC-CO2 estimates presented in Table 5-8 have a number of limitations. First,
the current scientific and economic understanding of discounting approaches suggests discount
rates appropriate for intergenerational analysis in the context of climate change are likely to be
less than 3 percent, near 2 percent or lower (IWG 2021). Second, the IAMs used to produce
these interim estimates do not include all of the important physical, ecological, and economic
impacts of climate change recognized in the climate change literature and the science underlying
their "damage functions" - i.e., the core parts of the IAMs that map global mean temperature
changes and other physical impacts of climate change into economic (both market and
nonmarket) damages - lags behind the most recent research. For example, limitations include the
incomplete treatment of catastrophic and non-catastrophic impacts in the integrated assessment
models, their incomplete treatment of adaptation and technological change, the incomplete way
in which inter-regional and intersectoral linkages are modeled, uncertainty in the extrapolation of
damages to high temperatures, and inadequate representation of the relationship between the
discount rate and uncertainty in economic growth over long time horizons. Likewise, the
13 Although the distributions and numbers in Figure 5-3 are based on the full set of model results (150,000 estimates
for each discount rate), for display purposes the horizontal axis is truncated with 0.78 percent of the estimates falling
below the lowest bin displayed and 3.64 percent of the estimates falling above the highest bin displayed.
5-33
-------
socioeconomic and emissions scenarios used as inputs to the models do not reflect new
information from the last decade of scenario generation or the full range of projections.
The modeling limitations do not all work in the same direction in terms of their influence
on the SC-CO2 estimates. However, the IWG has recommended that, taken together, the
limitations suggest that the interim SC-CO2 estimates used in this final rule likely underestimate
the damages from CO2 emissions. In particular, the Intergovernmental Panel on Climate Change
(IPCC) Fourth Assessment Report (IPCC 2007), which was the most current IPCC assessment
available at the time when the IWG decision over the ECS input was made, concluded that SC-
CO2 estimates "very likely.. .underestimate the damage costs" due to omitted impacts. Since
then, the peer-reviewed literature has continued to support this conclusion, as noted in the
IPCC's Fifth Assessment report (IPCC 2014) and other recent scientific assessments (e.g., IPCC
2018, 2019a, 2019b; U.S. Global Change Research Program (USGCRP) 2016, 2018; and
National Academies 2016b, 2019). These assessments confirm and strengthen the science,
updating projections of future climate change and documenting and attributing ongoing changes.
For example, sea level rise projections from the IPCC's Fourth Assessment report ranged from
18 to 59 centimeters by the 2090s relative to 1980-1999, while excluding any dynamic changes
in ice sheets due to the limited understanding of those processes at the time (IPCC 2007). A
decade later, the Fourth National Climate Assessment projected a substantially larger sea level
rise of 30 to 130 centimeters by the end of the century relative to 2000, while not ruling out even
more extreme outcomes (USGCRP 2018). The February 2021 TSD briefly previews some of the
recent advances in the scientific and economic literature that the IWG is actively following and
that could provide guidance on, or methodologies for, addressing some of the limitations with the
interim SC-CO2 estimates. The IWG will be taking comment on how to incorporate the
recommendations of the National Academies (2017) and other recent science including the
advances discussed in the February 2021 TSD in the development of the fully updated SC-GHG
estimates to be released by January of 2022 under E.O. 13990. To complement the IWG
process, and as an active member of the IWG, the EPA will also be soliciting comment in
forthcoming proposed rules that use the interim SC-CO2 presented in this RIA.
Table 5-9 shows the estimated monetary value of the estimated changes in CO2 emissions
expected to occur over 2021-2040 for the Revised CSAPR Update, the more-stringent
alternative, and the less-stringent alternative. EPA estimated the dollar value of the C02-related
5-34
-------
effects for each analysis year between 2021 and 2040 by applying the SC-CO2 estimates, shown
in Table 5-8, to the estimated changes in CO2 emissions in the corresponding year under the
regulatory options.14 EPA then calculated the present value and annualized benefits from the
perspective of 2020 by discounting each year-specific value to the year 2020 using the same
discount rate used to calculate the SC-CO2.15
Table 5-9. Estimated Global Climate Benefits from Changes in CO2 Emissions 2021 -
	2040 (Millions of2016$)a	
Discount Rate and Statistic
Regulatory Alternative
Year
5%
Average
3%
Average
2.5%
Average
3%
95th
Percentile

2021
0
1
1
2

2022
46
143
206
434

2023
94
290
417
882
Finalized Option
2024
2025
102
109
311
331
444
473
946
1,011

2030
128
373
525
1,146

2035
98
273
380
838

2040
127
340
467
1,043

2021
1
2
3
7

2022
76
237
341
720

2023
156
480
689
1,460
More-Stringent
2024
204
623
892
1,898
Alternative
2025
254
771
1,100
2,350

2030
323
939
1,322
2,885

2035
316
878
1,222
2,698

2040
383
1,025
1,410
3,146

2021
0
1
1
3
14 Under the baseline, CO2 emissions are projected to rise through 2025 and then taper off through 2035 and rise
during the rest of the period, reflecting increasing demand growth, changing generation mix patterns and the impact
of retiring capacity. CO2 emissions reductions as a result of the modeled policies follow a similar trend, which
causes total climate benefit estimates to oscillate over time.
15According to OMB's Circular A-4 (2003), an "analysis should focus on benefits and costs that accrue to citizens
and residents of the United States", and international effects should be reported separately. To correctly assess the
total climate damages to U.S. citizens and residents, an analysis must account for impacts that occur within U.S.
borders, climate impacts occurring outside U.S. borders that directly and indirectly affect the welfare of U.S. citizens
and residents, and spillover effects from climate action elsewhere. The SC-CO2 estimates used in regulatory analysis
under revoked E.O. 13783, including in the R1A for the proposed rule, were an approximation of the climate
damages occurring within U.S. borders only (e.g., $7/mtC02 (2016 dollars) and $9/mtC02 using a 3% discount rate
for emissions occurring in 2021 and 2040, respectively). Applying the same estimate (based on a 3% discount rate)
to the CO2 emission reduction expected under the finalized option in this final rule would yield benefits from
climate impacts within U.S borders of $0.1 million in 2021, increasing to $39.8 million in 2040. However, as
discussed at length in the IWG's February 2021 TSD, estimates focusing on the climate impacts occurring solely
within U.S. borders are an underestimate of the benefits of CO2 mitigation accruing to U.S. citizens and residents, as
well as being subject to a considerable degree of uncertainty due to the manner in which they are derived.
5-35
-------

2022
39
122
176
371

2023
80
248
356
754
Less-Stringent
2024
81
248
355
755
2025
82
248
353
755
Alternative
2030
93
271
381
831

2035
73
203
282
623

2040
91
242
333
743
a Climate benefits are based on changes (reductions) in CO2 emissions and are calculated using four different
estimates of the social cost of carbon (SC-CO2) (model average at 2.5 percent, 3 percent, and 5 percent discount
rates; 95th percentile at 3 percent discount rate). We emphasize the importance and value of considering the
benefits calculated using all four SC-CO2 estimates. As discussed in the Technical Support Document: Social
Cost of Carbon, Methane, and Nitrous Oxide Interim Estimates under Executive Order 13990 (IWG 2021), a
consideration of climate benefits calculated using discount rates below 3 percent, including 2 percent and lower,
are also warranted when discounting intergenerational impacts.	
5.3 Total Benefits
Tables 5-10 through 5-12 present the total health and climate benefits for the final rule
and the more and less stringent alternatives for 2021, 2025, and 2030. In each of these tables, for
each discount rate and regulatory control alternative, multiple benefits estimates are presented
reflecting alternative ozone and PM2.5 mortality risk estimates.
Table 5-10. Combined Health Benefits and Climate Benefits for the Final Rule and More
and Less Stringent Alternatives for 2021 (millions of 2016$)a

Health and Climate Benefits
Climate Benefits
Onlyb
SC-CO2 Discount
Rate and Statistic
(Discount Rate Applied to Health
Benefits)

3%
7%

Final Rule
5% (average)
$230 and $1,900
$200 and $1,700
$0
3% (average)
$230 and $1,900
$200 and $1,700
$1
2.5% (average)
$230 and $1,900
$200 and $1,700
$1
3% (95th percentile)
$230 and $1,900
$200 and $1,700
$2
More Stringent Alternative
5% (average)
$260 and $1,900
$200 and $1,700
$1
3% (average)
$260 and $1,900
$200 and $1,700
$2
2.5% (average)
$260 and $1,900
$200 and $1,700
$3
3% (95th percentile)
$270 and $1,900
$210 and $1,700
$7
Less Stringent Alternative
5% (average)
$20 and $190
$20 and $170
$0
3% (average)
$20 and $190
$20 and $170
$1
2.5% (average)
$20 and $190
$20 and $170
$1
5-36
-------
3% (95th percentile) $20 and $190	$20 and $170
$3
a The two benefits estimates are separated by the word "and" to signify that they are two separate estimates. The
estimates do not represent lower- and upper-bound estimates and should not be summed.
b Climate benefits are based on changes (reductions) in CO2 emissions and are calculated using four different
estimates of the social cost of carbon (SC-CO2) (model average at 2.5 percent, 3 percent, and 5 percent discount
rates; 95th percentile at 3 percent discount rate). For purposes of this table, we show the benefits associated with the
model average at a 3 percent discount rate. However, we emphasize the importance and value of considering the
benefits calculated using all four SC-CO2 estimates; the additional benefit estimates range from $0.24 million to
$2.31 million in 2021 for the finalized option. As discussed in Chapter 5, a consideration of climate benefits
calculated using discount rates below 3 percent, including 2 percent and lower, are also warranted when discounting
intergenerational impacts.
Table 5-11. Combined Health Benefits and Climate Benefits for the Final Rule and More
and Less Stringent Alternatives for 2025 (millions of 2016$)a

Health and Climate Benefits
Climate
Benefits Onlyb
SC-CO2 Discount
Rate and Statistic
(Discount Rate Applied to Health
Benefits)

3%
7%

Final Rule
5% (average)
$430 and $2,500
$400 and $2,300
$110
3% (average)
$650 and $2,700
$620 and $2,500
$330
2.5% (average)
$790 and $2,900
$760 and $2,700
$470
3% (95th percentile)
$1,300 and $3,400
$1,300 and $3,200
$1,000
More Stringent Alternative
5% (average)
$790 and $4,500
$740 and $4,000
$250
3% (average)
$1,300 and $5,000
$1,300 and $4,600
$770
2.5% (average)
$1,600 and $5,300
$1,600 and $4,900
$1,100
3% (95th percentile)
$2,900 and $6,600
$2,900 and $6,200
$2,400
Less Stringent Alternative
5% (average)
$100 and $280
$100 and $250
$80
3% (average)
$270 and $450
$270 and $420
$250
2.5% (average)
$370 and $550
$370 and $520
$350
3% (95th percentile)
$780 and $960
$780 and $930
$760
" The two benefits estimates are separated by the word "and" to signify that they are two separate estimates. The
estimates do not represent lower- and upper-bound estimates and should not be summed.
b Climate benefits are based on changes (reductions) in CO2 emissions and are calculated using four different
estimates of the social cost of carbon (SC-CO2) (model average at 2.5 percent, 3 percent, and 5 percent discount
rates; 95th percentile at 3 percent discount rate). For purposes of this table, we show the benefits associated with the
model average at a 3 percent discount rate. However, we emphasize the importance and value of considering the
benefits calculated using all four SC-CO2 estimates; the additional benefit estimates range from $ 109 million to
$1,011 million in 2025 for the finalized option. As discussed in Chapter 5, a consideration of climate benefits
calculated using discount rates below 3 percent, including 2 percent and lower, are also warranted when discounting
intergenerational impacts.
5-37
-------
Table 5-12. Combined Health Benefits and Climate Benefits for the Final Rule and More
	and Less Stringent Alternatives for 2030 (millions of 2016$)a
Health and Climate Benefits	Climate
SC-CO2 Discount	(Discount Rate Applied to Health Benefits Onlvb
Rate and Statistic	Benefits)
	3%	7%	
Final Rule
5% (average)
$470 and $2,700
$460 and $2,600
$130
3% (average)
$710 and $3,000
$700 and $2,900
$370
2.5% (average)
$870 and $3,100
$860 and $3,000
$530
3% (95th percentile)
$1,400 and $3,700
$1430 and $3,600
$1,100
More Stringent Alternative
5% (average)
$910 and $4,900
$880 and $4,200
$320
3% (average)
$1,500 and $5,500
$1,500 and $4,800
$940
2.5% (average)
$1,900 and $5,900
$1,900 and $5,200
$1,300
3% (95th percentile)
$3,500 and $7,500
$3,500 and $6,800
$2,900
Less Stringent Alternative
5% (average)
$120 and $300
$110 and $270
$90
3% (average)
$300 and $480
$290 and $450
$270
2.5% (average)
$410 and $590
$400 and $560
$380
3% (95th percentile)
$860 and $1,040
$850 and $1,010
$830
a The two benefits estimates are separated by the word "and" to signify that they are two separate estimates. The
estimates do not represent lower- and upper-bound estimates and should not be summed.
b Climate benefits are based on changes (reductions) in CO2 emissions and are calculated using four different
estimates of the social cost of carbon (SC-CO2) (model average at 2.5 percent, 3 percent, and 5 percent discount
rates; 95th percentile at 3 percent discount rate). For purposes of this table, we show the benefits associated with the
model average at a 3 percent discount rate. However, we emphasize the importance and value of considering the
benefits calculated using all four SC-CO2 estimates; the additional benefit estimates range from $128 million to
$1,146 million in 2030 for the finalized option. As discussed in Chapter 5, a consideration of climate benefits
calculated using discount rates below 3 percent, including 2 percent and lower, are also warranted when discounting
intergenerational impacts.
5.4 Unquantified Benefits
In the presence of sunlight, NOx and volatile organic compounds (VOCs) can undergo a
chemical reaction in the atmosphere to form ozone. Reducing NOx emissions generally reduces
human exposure to ozone and the incidence of ozone-related health effects, though the degree to
which ozone is reduced will depend in part on local levels of VOCs as discussed in Chapter 3.
The final rule would also reduce emissions of NOx throughout the year. Because NOx is also a
precursor to formation of ambient PM2.5, reducing these emissions would reduce human
exposure to ambient PM2.5 throughout the year and would thus reduce the incidence of PM2.5-
5-38
-------
attributable health effects.16 Reducing emissions of NOx would also reduce ambient exposure to
NO2 and its associated health effects.
Data, time, and resource limitations prevented EPA from quantifying the estimated
impacts or monetizing estimated benefits associated with exposure to NO2 (independent of the
role NO2 plays as precursors to PM2.5), as well as ecosystem effects, and visibility impairment
due to the absence of air quality modeling data for these pollutants in this analysis. Lack of
quantification does not imply that there are no additional benefits associated with reductions in
exposures to ozone, PM2.5, or NO2. In this section, we provide a qualitative description of these
benefits, which are listed in Table 5-13.
Table 5-13. Unquantified Health and Welfare Benefits Categories
Category
Effect
Effect
Quantified
Effect
Monetized
More
Information
Improved Human Health

Asthma hospital admissions
—
—
NO2 ISA2

Chronic lung disease hospital admissions
—
—
NO2 ISA2

Respiratory emergency department visits
—
—
NO2 ISA2
Reduced incidence of
Asthma exacerbation
—
—
NO2 ISA2
morbidity from
exposure to NO2
Acute respiratory symptoms
—
—
NO2 ISA2
Premature mortality
—
—
NO2 ISA2'3'4

Other respiratory effects (e.g., airway
hyperresponsiveness and inflammation, lung
function, other ages and populations)
—
—
NO2 ISA3-4
Improved Environment
Reduced visibility
Visibility in Class 1 areas
—
—
PM ISA2
impairment
Visibility in residential areas
—
—
PM ISA2
Reduced effects on
Household soiling
—
—
PM ISA2-3
materials
Materials damage (e.g., corrosion, increased wear)
—
—
PM ISA3
Reduced effects from




PM deposition (metals
and organics)
Effects on Individual organisms and ecosystems
—
—
PM ISA3

Visible foliar injury on vegetation
—
—
Ozone ISA2

Reduced vegetation growth and reproduction
—
—
Ozone ISA2
Reduced vegetation
Yield and quality of commercial forest products and
crops
—
—
Ozone ISA2
and ecosystem effects
from exposure to
ozone
Damage to urban ornamental plants
—
—
Ozone ISA3
Carbon sequestration in terrestrial ecosystems
—
—
Ozone ISA2

Recreational demand associated with forest


Ozone ISA3

aesthetics



Other non-use effects


Ozone ISA3
16 This RIA does not quantify PM2 5-related benefits associated with direct PM2 5 and SO2 emission reductions. As
discussed in Chapter 4, EPA does not estimate significant direct PM2 5 and SO2 emission reductions as a result of
this rule.
5-39
-------
Ecosystem functions (e.g., water cycling,
biogeochemical cycles, net primary productivity, — — Ozone ISA3
leaf-gas exchange, community composition)	

Recreational fishing
—
—
NOx
SOxISA2

Tree mortality and decline
—
—
NOx
SOxISA3
Reduced effects from
acid deposition
Commercial fishing and forestry effects
—
—
NOx
SOxISA3
Recreational demand in terrestrial and aquatic
ecosystems
—
—
NOx
SOxISA3

Other non-use effects


NOx
SOxISA3

Ecosystem functions (e.g., biogeochemical cycles)
—
—
NOx
SOxISA3

Species composition and biodiversity in terrestrial


NOx
SOxISA3

and estuarine ecosystems



Coastal eutrophication
—
—
NOx
SOxISA3
Reduced effects from
nutrient enrichment
Recreational demand in terrestrial and estuarine
ecosystems
—
—
NOx
SOxISA3

Other non-use effects


NOx
SOxISA3

Ecosystem functions (e.g., biogeochemical cycles,


NOx
SOx ISA3

fire regulation)




Reduced vegetation
Injury to vegetation from SO2 exposure
—
—
NOx
SOxISA3
effects from ambient





exposure to SO2 and
NOx
Injury to vegetation from NOx exposure
—
—
NOx
SOxISA3
1 These endpoints are generally quantified and monetized when EPA quantitatively characterizes the benefits of changes in PM2.5
and Ozone.
2	We assess these benefits qualitatively due to data and resource limitations for this RIA.
3	We assess these benefits qualitatively because we do not have sufficient confidence in available data or methods.
4	We assess these benefits qualitatively because current evidence is only suggestive of causality or there are other significant
concerns over the strength of the association.
5.4.1 NO2 Health Benefits
In addition to being a precursor to PM2.5 and ozone, NOx emissions are also linked to a
variety of adverse health effects associated with direct exposure. We were unable to estimate the
health benefits associated with reduced NO2 exposure in this analysis. Following a
comprehensive review of health evidence from epidemiologic and laboratory studies, the
Integrated Science Assessment for Oxides of Nitrogen —Health Criteria (NOx ISA) (U.S. EPA,
2016c) concluded that there is a likely causal relationship between respiratory health effects and
short-term exposure to NO2. These epidemiologic and experimental studies encompass a number
of endpoints including emergency department visits and hospitalizations, respiratory symptoms,
airway hyperresponsiveness, airway inflammation, and lung function. The NOx ISA also
concluded that the relationship between short-term NO2 exposure and premature mortality was
"suggestive but not sufficient to infer a causal relationship," because it is difficult to attribute the
mortality risk effects to NO2 alone. Although the NOx ISA stated that studies consistently
5-40
-------
reported a relationship between NO2 exposure and mortality, the effect was generally smaller
than that for other pollutants such as PM.
5.4.2	Ozone Welfare Benefits
Exposure to ozone has been associated with a wide array of vegetation and ecosystem
effects in the published literature (U.S. EPA, 2020b). Sensitivity to ozone is highly variable
across species, with over 65 plant species identified as "ozone-sensitive", many of which occur
in state and national parks and forests. These effects include those that damage or impair the
intended use of the plant or ecosystem. Such effects can include reduced growth and/or biomass
production in sensitive plant species, including forest trees, reduced yield and quality of crops,
visible foliar injury, species composition shift, and changes in ecosystems and associated
ecosystem services.
5.4.3	NO2 Welfare Benefits
As described in the Integrated Science Assessment for Oxides of Nitrogen and Sulfur —
Ecological Criteria (NOx/SOx ISA) (U.S. EPA, 2008b), NOx emissions also contribute to a
variety of adverse welfare effects, including those associated with acidic deposition, visibility
impairment, and nutrient enrichment. Deposition of nitrogen causes acidification, which can
cause a loss of biodiversity of fishes, zooplankton, and macro invertebrates in aquatic
ecosystems, as well as a decline in sensitive tree species, such as red spruce (Picea rubens) and
sugar maple (Acer saccharum) in terrestrial ecosystems. In the northeastern U.S., the surface
waters affected by acidification are a source of food for some recreational and subsistence
fishermen and for other consumers and support several cultural services, including aesthetic and
educational services and recreational fishing. Biological effects of acidification in terrestrial
ecosystems are generally linked to aluminum toxicity, which can cause reduced root growth,
restricting the ability of the plant to take up water and nutrients. These direct effects can, in turn,
increase the sensitivity of these plants to stresses, such as droughts, cold temperatures, insect
pests, and disease leading to increased mortality of canopy trees. Terrestrial acidification affects
several important ecological services, including declines in habitat for threatened and endangered
species (cultural), declines in forest aesthetics (cultural), declines in forest productivity
5-41
-------
(provisioning), and increases in forest soil erosion and reductions in water retention (cultural and
regulating). (U.S. EPA, 2008b)
Deposition of nitrogen is also associated with aquatic and terrestrial nutrient enrichment.
In estuarine waters, excess nutrient enrichment can lead to eutrophication. Eutrophication of
estuaries can disrupt an important source of food production, particularly fish and shellfish
production, and a variety of cultural ecosystem services, including water-based recreational and
aesthetic services. Terrestrial nutrient enrichment is associated with changes in the types and
number of species and biodiversity in terrestrial systems. Excessive nitrogen deposition upsets
the balance between native and nonnative plants, changing the ability of an area to support
biodiversity. When the composition of species changes, then fire frequency and intensity can
also change, as nonnative grasses fuel more frequent and more intense wildfires. (U.S. EPA,
2008b)
5.4.4 Visibility Impairment Benefits
Reducing secondary formation of PM2.5 would improve levels of visibility in the U.S.
because suspended particles and gases degrade visibility by scattering and absorbing light (U.S.
EPA, 2009). Fine particles with significant light-extinction efficiencies include sulfates, nitrates,
organic carbon, elemental carbon, and soil (Sisler, 1996). Visibility has direct significance to
people's enjoyment of daily activities and their overall sense of wellbeing. Good visibility
increases the quality of life where individuals live and work, and where they engage in
recreational activities. Particulate sulfate is the dominant source of regional haze in the eastern
U.S. and particulate nitrate is an important contributor to light extinction in California and the
upper Midwestern U.S., particularly during winter (U.S. EPA, 2009). Previous analyses (U.S.
EPA, 201 la) show that visibility benefits can be a significant welfare benefit category. Without
air quality modeling, we are unable to estimate visibility-related benefits, and we are also unable
to determine whether the emission reductions associated with the final emission guidelines
would be likely to have a significant impact on visibility in urban areas or Class I areas.
Reductions in emissions of NO2 will improve the level of visibility throughout the United
States because these gases (and the particles of nitrate and sulfate formed from these gases)
impair visibility by scattering and absorbing light (U.S. EPA, 2009). Visibility is also referred to
5-42
-------
as visual air quality (VAQ), and it directly affects people's enjoyment of a variety of daily
activities (U.S. EPA, 2009). Good visibility increases quality of life where individuals live and
work, and where they travel for recreational activities, including sites of unique public value,
such as the Great Smoky Mountains National Park (U. S. EPA, 2009).
5.5 References
Amorim MIM, Mergler D, Bahia MO, Dubeau H, Miranda D, Lebel J, et al. 2000. Cytogenetic
damage related to low levels of methyl mercury contamination in the Brazilian Amazon.
An Acad Bras Cienc 72:497-507; doi:10.1590/S0001- 37652000000400004.
Anthoff, D. and Tol, R.S.J. 2013a. The uncertainty about the social cost of carbon: a
decomposition analysis using FUND. Climatic Change 117: 515-530.
Anthoff, D. and Tol, R.S.J. 2013b. Erratum to: The uncertainty about the social cost of carbon: A
decomposition analysis using FUND. Climatic Change 121(2): 413.
ATSDR. 1999. ATSDR - Toxicological Profile: Mercury.
Bell, M.L., A. McDermott, S.L. Zeger, J.M. Sarnet, and F. Dominici. 2004. "Ozone and Short-
Term Mortality in 95 U.S. Urban Communities, 1987-2000." Journal of the American
Medical Association. 292(19):2372-8.
Bell, M.L., F. Dominici, and J.M. Samet. 2005. "A Meta-Analysis of Time-Series Studies of
Ozone and Mortality with Comparison to the National Morbidity, Mortality, and Air
Pollution Study." Epidemiology. 16(4):436-45.
Di, Q, Wang, Y, Zanobetti, A, Wang, Y, Koutrakis, P, Choirat, C, Dominici, F and Schwartz, JD
(2017). Air pollution and mortality in the Medicare population. New Engl J Med 376(26):
2513-2522.
Executive Order (E.O.) 13783. Promoting Energy Independence and Economic Growth March
28, 2017. Available at: https://www.federalregister.gov/documents/2017/03/31/2017-
06576/promoting-energy-independence-and-economic-growth (accessed February 5,
2021).
Executive Order (E.O.) 13990. Protecting Public Health and the Environment and Restoring
Science to Tackle the Climate Crisis. January 20, 2021. Available at:
https://www.federalregister.gOv/documents/2021/01/25/2021-01765/protecting-public-
health-and-the-environment-and-restoring-science-to-tackle-the-climate-crisis (accessed
February 5, 2021).
Fann N, Nolte CG, Dolwick P, Spero TL, Brown AC, Phillips S, et al. 2015. The geographic
distribution and economic value of climate change-related ozone health impacts in the
United States in 2030. J Air Waste Manag Assoc 65;
doi: 10.1080/10962247.2014.996270.
5-43
-------
Gwinn, M.R., J. Craig, D.A. Axelrad, R. Cook, C. Dockins, N. Fann, R. Fegley, D.E. Guinnup,
G. Helfand, B. Hubbell, S.L. Mazur, T. Palma, R.L. Smith, J. Vandenberg, and
B. Sonawane. 2011. "Meeting report: Estimating the benefits of reducing hazardous air
pollutants—summary of 2009 workshop and future considerations." Environmental
Health Perspectives. 119(1): 125-30.
Hope, C. 2013. Critical issues for the calculation of the social cost of CO2: why the estimates
from PAGE09 are higher than those from PAGE2002. Climatic Change, 117: 531-543.
Huang Y, Dominici F, Bell M. 2004. Bayesian Hierarchical Distributed Lag Models for Summer
Ozone Exposure and Cardio-Respiratory Mortality. Johns Hopkins Univ Dept Biostat
Work Pap Ser.
IARC. 1994. Mercury Carcinogenicity.
Interagency Working Group on Social Cost of Carbon (IWG). 2010. Technical Support
Document: Social Cost of Carbon for Regulatory Impact Analysis under Executive Order
12866. February. United States Government.
Interagency Working Group on Social Cost of Carbon (IWG). 2013. Technical Support
Document: Technical Update of the Social Cost of Carbon for Regulatory Impact
Analysis Under Executive Order 12866. May. United States Government.
Interagency Working Group on Social Cost of Greenhouse Gases (IWG). 2016a. Technical
Support Document: Technical Update of the Social Cost of Carbon for Regulatory Impact
Analysis Under Executive Order 12866. August. United States Government.
Interagency Working Group on the Social Cost of Greenhouse Gases. 2016b. Addendum to
Technical Support Document on Social Cost of Carbon for Regulatory Impact Analysis
under Executive Order 12866: Application of the Methodology to Estimate the Social
Cost of Methane and the Social Cost of Nitrous Oxide. August. United Stated
Government. Available at: https://www.epa.gov/sites/production/files/2016-
12/documents/addendum_to_sc-ghg_tsd_august_2016.pdf (accessed February 5, 2021).
Interagency Working Group on Social Cost of Greenhouse Gases (IWG). 2021. Technical
Support Document: Social Cost of Carbon, Methane, and Nitrous Oxide Interim
Estimates under Executive Order 13990. February. United States Government. Available
at: 
-------
Jhun I, Fann N, Zanobetti A, Hubbell B. 2014. Effect modification of ozone-related mortality
risks by temperature in 97 US cities. Environment International. 73:128-34.
Katsouyanni, K, Samet, JM, Anderson, HR, Atkinson, R, Le Tertre, A, Medina, S, Samoli, E,
Touloumi, G, Burnett, RT, Krewski, D, Ramsay, T, Dominici, F, Peng, RD, Schwartz, J,
Zanobetti, A and Committee, HEIHR (2009). Air pollution and health: a European and
North American approach (APHENA). Res Rep Health Eff Inst(142): 5-90.
Kivi, PA and Shogren, JF (2010). Second-order ambiguity in very low probability risks: food
safety valuation. J Agric Resour Econ: 443-456.
Krewski D., M. Jerrett, R.T. Burnett, R. Ma, E. Hughes, Y. Shi, et al. 2009. Extended Follow-Up
and Spatial Analysis of the American Cancer Society Study Linking Particulate Air
Pollution and Mortality. HEI Research Report, 140, Health Effects Institute, Boston, MA.
Lepeule, J., F. Laden, D. Dockery, and J. Schwartz. 2012. "Chronic Exposure to Fine Particles
and Mortality: An Extended Follow-Up of the Harvard Six Cities Study from 1974 to
2009." Environmental Health Perspectives. 120(7):965-70.
Levy, J.I., S.M. Chemerynski, and J.A. Sarnat. 2005. "Ozone Exposure and Mortality: An
Empiric Bayes Metaregression Analysis." Epidemiology. 16(4):458-68.
McGartland, A, Revesz, R, Axelrad, DA, Dockins, C, Sutton, P and Woodruff, TJ (2017).
Estimating the health benefits of environmental regulations. Science 357(6350): 457-
458.
National Academies of Sciences, Engineering, and Medicine (National Academies). 2017.
Valuing Climate Damages: Updating Estimation of the Social Cost of Carbon Dioxide.
Washington, D.C.: National Academies Press.
National Research Council (NRC). 2000. Toxicological Effects of Methylmercury: Committee
on the Toxicological Effects of Methylmercury." Board on Environmental Studies and
Toxicology. National Academies Press. Washington, DC.
National Research Council (NRC). 2002. Estimating the Public Health Benefits of Proposed Air
Pollution Regulations. National Academies Press. Washington, DC.
National Research Council (NRC). 2008. Estimating Mortality Risk Reduction and Economic
Benefits from Controlling Ozone Air Pollution. National Academies Press. Washington,
DC.
Nordhaus, W. 2010. Economic aspects of global warming in a post-Copenhagen environment.
Proceedings of the National Academy of Sciences, 107(26): 11721-11726.
Office of Management and Budget (OMB). 2003. Circular A-4: Regulatory Analysis.
Washington, DC. Available at: < http://www.whitehouse.gov/omb/circulars/a004/a-
4.html>.
5-45
-------
Pizer, W., M. Adler, J. Aldy, D. Anthoff, M. Cropper, K. Gillingham, M. Greenstone, B. Murray,
R. Newell, R. Richels, A. Rowell, S. Waldhoff, J. Wiener. 2014. "Using and improving
the social cost of carbon." Science, Vol. 346, No. 6214, 12/05/14, pp 1189-1190.
Ren, C., G.M. Williams, K. Mengersen, L. Morawska, and S. Tong. 2008a. "Does Temperature
Modify Short-Term Effects of Ozone on Total Mortality in 60 Large Eastern U.S.
Communities? An Assessment Using the NMMAPS Data." Environment International.
34:451-458.
Ren, C., G.M. William, L. Morawska, K. Mengensen, and S. Tong. 2008b. "Ozone Modifies
Associations between Temperature and Cardiovascular Mortality: Analysis of the
NMMAPS Data." Occupational and Environmental Medicine. 65:255-260.
Sisler, J.F. 1996. Spatial and Seasonal Patterns and Long-Term Variability of the Composition of
the Haze in the United States: An analysis of data from the IMPROVE network. CIRA
Report, ISSN 0737-5352-32, Colorado State University.
Schwartz, J. 2005. "How Sensitive is the Association between Ozone and Daily Deaths to
Control for Temperature?" American Journal of Respiratory and Critical Care Medicine.
171(6): 627-31.
Smith RL, Xu B, Switzer P. 2009. Reassessing the relationship between ozone and short-term
mortality in U.S. urban communities. Inhal Toxicol 21 Suppl 2:37-61;
doi: 10.1080/08958370903161612.
Turner, MC, Jerrett, M, Pope, A, III, Krewski, D, Gapstur, SM, Diver, WR, Beckerman, BS,
Marshall, JD, Su, J, Crouse, DL and Burnett, RT (2016). Long-term ozone exposure and
mortality in a large prospective study. Am J Respir Crit Care Med 193(10): 1134-1142.
U.S. Environmental Protection Agency (U.S. EPA). 2002. Methylmercury (MeHg) CASRN
22967-92-6 | IRIS | US EPA, ORD
U.S. Environmental Protection Agency (U.S. EPA). 2008b. Integrated Science Assessment for
Oxides of Nitrogen and Sulfur-Ecological Criteria National (Final Report). National
Center for Environmental Assessment - RTP Division, Research Triangle Park, NC.
EPA/600/R-08/139. December. Available at:
.
U.S. Environmental Protection Agency (U.S. EPA). 2009. Integrated Science Assessment for
Particulate Matter (Final Report). EPA-600-R-08-139F. National Center for
Environmental Assessment - RTP Division, Research Triangle Park, NC. December.
Available at: .
U.S. Environmental Protection Agency (U.S. EPA). 2010a. Integrated Science Assessment for
Carbon Monoxide. National Center for Environmental Assessment, Research Triangle
Park, NC. EPA/600/R-09/019F. January. Available at:
.
5-46
-------
U.S. Environmental Protection Agency (U.S. EPA). 2010b. Technical Support Document:
Summary of Expert Opinions on the Existence of a Threshold in the Concentration-
Response Function for PIVh.s-related Mortality. Research Triangle Park, NC. June.
Available at: .
U.S. Environmental Protection Agency (U.S. EPA). 2010c. Regulatory Impact Analysis (RIA)
for Existing Stationary Compression Ignition Engines NESHAP Final Draft.
U.S. Environmental Protection Agency (U.S. EPA). 2010d. Regulatory Impact Analysis for the
Proposed Federal Transport Rule.
U.S. Environmental Protection Agency (U.S. EPA). 201 la. Policy Assessment for the Review of
the Particulate Matter National Ambient Air Quality Standards. EPA-452/D-11-003.
Office of Air Quality Planning and Standards, Health and Environmental Impacts
Division. April. Available at:
.
U.S. Environmental Protection Agency (U.S. EPA). 201 lb. The Benefits and Costs of the Clean
Air Act from 1990 to 2020. Office of Air and Radiation, Washington, DC. March.
Available at: .
U.S. Environmental Protection Agency (U.S. EPA). 201 lc. Regulatory Impact Analysis for the
Federal Implementation Plans to Reduce Interstate Transport of Fine Particulate Matter
and Ozone in 27 States; Correction of SIP Approvals for 22 States.
U.S. Environmental Protection Agency (U.S. EPA). 201 Id. Regulatory Impact Analysis for the
Final Mercury and Air Toxics Standards.
U.S. Environmental Protection Agency (U.S. EPA). 2012. Regulatory Impact Analysis for the
Final Revisions to the National Ambient Air Quality Standards for Particulate Matter.
EPA-452/R-12-003. Office of Air Quality Planning and Standards, Health and
Environmental Impacts Division, Research Triangle Park, NC. Available at: <
http://www.epa.gov/ttnecasl/regdata/RIAs/finalria.pdf>.
U.S. Environmental Protection Agency (U.S. EPA). 2013a. Integrated Science Assessment of
Ozone and Related Photochemical Oxidants (Final Report). EPA/600/R-10/076F.
National Center for Environmental Assessment - RTP Division, Research Triangle Park.
Available at: .
U.S. Environmental Protection Agency (U.S. EPA). 2013b. Regulatory Impact Analysis for the
Final Revisions to the National Ambient Air Quality Standards for Particulate Matter.
U.S. Environmental Protection Agency (U.S. EPA). 2014a. Regulatory Impact Analysis (RIA)
for Proposed Residential Wood Heaters NSPS Revision.
U.S. Environmental Protection Agency (U.S. EPA). 2014b. Regulatory Impact Analysis for the
Proposed Carbon Pollution Guidelines for Existing Power Plants and Emission Standards
for Modified and Reconstructed Power Plants.
5-47
-------
U.S. Environmental Protection Agency (U.S. EPA). 2014c. Regulatory Impact Analysis of the
Proposed Revisions to the National Ambient Air Quality Standards for Ground-Level
Ozone.
U.S. Environmental Protection Agency (U.S. EPA). 2015a. Regulatory Impact Analysis for
Residential Wood Heaters NSPS Revision: Final Report.
U.S. Environmental Protection Agency (U.S. EPA). 2015b. Regulatory Impact Analysis for the
Clean Power Plan Final Rule.
U.S. Environmental Protection Agency (U.S. EPA). 2015c. Regulatory Impact Analysis for the
Proposed Cross-State Air Pollution Rule (CSAPR) Update for the 2008 Ozone National
Ambient Air Quality Standards (NAAQS).
U.S. Environmental Protection Agency (U.S. EPA). 2015d. Regulatory Impact Analysis for the
Proposed Federal Plan Requirements for Greenhouse Gas Emissions from Electric Utility
Generating Units Constructed on or Before January 8, 2014; Model Trading Rules;
Amendments to Framework Regulations. 4-52
U.S. Environmental Protection Agency (U.S. EPA). 2015e. Regulatory Impact Analysis of the
Final Revisions to the National Ambient Air Quality Standards for Ground-Level Ozone,
EPA-452/R-15-07. EPA-452/R-12-003. Office of Air Quality Planning and Standards,
Health and Environmental Impacts Division, Research Triangle Park, NC.
(https://www.epa.gOv/sites/production/files/2016-02/documents/20151001ria.pdf).
U.S. Environmental Protection Agency (U.S. EPA). 2016a. Guidelines for Preparing Economic
Analyses.
U.S. Environmental Protection Agency (U.S. EPA). 2016b. Regulatory Impact Analysis of the
Cross-State Air Pollution Rule (CSAPR) Update for the 2008 National Ambient Air
Quality Standards for Ground-Level Ozone.
U.S. Environmental Protection Agency (U.S. EPA). 2016c. Integrated Science Assessment for
Oxides of Nitrogen - Health Criteria (Final Report). National Center for Environmental
Assessment, Research Triangle Park, NC. July. Available at: <
https://cfpub.epa.gov/ncea/isa/recordisplay.cfm?deid=310879>.
U.S. Environmental Protection Agency (U.S. EPA). 2017. Integrated Science Assessment for
Sulfur Oxides—Health Criteria (Final Report). National Center for Environmental
Assessment - RTP Division, Research Triangle Park, NC. September. Available at:
.
U.S. Environmental Protection Agency (U.S. EPA). 2018. Environmental Benefits Mapping and
Analysis Program - Community Edition. User's Manual. Office of Air Quality Planning
and Standards, Health and Environmental Impacts Division, Research Triangle Park, NC.
Available at: < https://www.epa.gov/sites/production/files/2015-04/documents/benmap-
ce_user_manual_march_2015,pdf>
5-48
-------
U.S. Environmental Protection Agency (U.S. EPA). 2019a. Integrated Science Assessment (ISA)
for Particulate Matter (Final Report, 2019). U.S. Environmental Protection Agency,
Washington, DC, EPA/600/R-19/188, 2019.
U.S. Environmental Protection Agency (U.S. EPA). 2019b. Regulatory Impact Analysis for
Repeal of the Clean Power Plan; Emission Guidelines for Greenhouse Gas Emissions
from Existing Electric Utility Generating Units; Revisions to Emission Guidelines
Implementing Regulations. EPA-452/R19-003. June 2019
U.S. Environmental Protection Agency (U.S. EPA). 2020a. Benefit and Cost Analysis for
Revisions to the Effluent Limitations Guidelines and Standards for the Steam Electric
Power Generating Point Source Category. U.S. Environmental Protection Agency,
Washington, DC, EPA-821-R-20-003.
U.S. Environmental Protection Agency (U.S. EPA). 2020b. Integrated Science Assessment (ISA)
for Ozone and Related Photochemical Oxidants (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-20/012, 2020.
U.S. Environmental Protection Agency (U.S. EPA). 2020c. Regulatory Impact Analysis for the
Proposed Cross-State Air Pollution Rule (CSAPR) Revised Update for the 2008 Ozone
National Ambient Air Quality Standards (NAAQS).
U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 2000. An
SAB Report on EPA's White Paper Valuing the Benefits of Fatal Cancer Risk Reduction.
EPA-SAB-EEAC-00-013. July. Available at:
.
U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 2004.
Advisory Council on Clean Air Compliance Analysis Response to Agency Request on
Cessation Lag. EPA-COUNCIL-LTR-05-001. December. Available at:
.
U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 2008.
Subject: Benefits of Reducing Benzene Emissions in Houston, 1990- 2020.
U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 2009a.
Review of EPA's Integrated Science Assessment for Particulate Matter (First External
Review Draft, December 2008). EPA-COUNCIL-09-008. May. Available at:
.
5-49
-------
U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 2009b.
Review of Integrated Science Assessment for Particulate Matter (Second External
Review Draft, July 2009). EPA-CASAC-10-001. November. Available at:
-------
USGCRP. 2016. The Impacts of Climate Change on Human Health in the United States: A
Scientific Assessment.; doi:http://dx.doi.org/10.7930/J0R49NQX.
Woodruff, T.J., J. Grillo, and K.C. Schoendorf. 2008. "Air pollution and postneonatal infant
mortality in the United States, 1999-2002." Environmental Health Perspectives. 116(1):
110-115
Woods & Pool (2015). Complete Demographic Database.
Zanobetti A, Schwartz J. 2008. Mortality displacement in the association of ozone with
mortality: an analysis of 48 cities in the United States. Am J Respir Crit Care Med
177:184-9; doi: 10.1164/rccm.200706-8230C.
5-51
-------
CHAPTER 6: STATUTORY AND EXECUTIVE ORDER REVIEWS
Overview
This chapter presents the statutory and executive orders applicable to EPA rules, and
discusses EPA's actions taken pursuant to these orders.
6.1	Executive Order 12866: Regulatory Planning and Review
This action is an economically significant regulatory action that was submitted to the
Office of Management and Budget (OMB) for review. Any changes made in response to OMB
recommendations have been documented in the docket. EPA prepared an analysis of the
potential costs and benefits associated with this final action.
6.2	Paperwork Reduction Act
This final action will not impose any new information collection burden under the PRA.
This final action relocates certain existing information collection requirements for certain
sources from subpart EEEEE of 40 CFR part 97 to a new subpart GGGGG of 40 CFR part 97,
but neither changes the inventory of sources subject to information collection requirements nor
changes any existing information collection requirements for any source. OMB has previously
approved the information collection activities contained in the existing regulations and has
assigned OMB control number 2060-0667.
6.3	Regulatory Flexibility Act
EPA certifies that this action will not have a significant economic impact on a substantial
number of small entities under the Regulatory Flexibility Act (RFA). The small entities subject
to the requirements of this action are small businesses, small organizations, and small
governmental jurisdictions. EPA has determined that no small entities potentially affected by the
rule will have compliance costs greater than 1 percent of annual revenues in 2021. Details of this
analysis are presented below.
The Regulatory Flexibility Act (5 U.S.C. 601 et seq.), as amended by the Small Business
Regulatory Enforcement Fairness Act (Public Law No. 104-121), provides that whenever an
agency is required to publish a general notice of proposed rulemaking, it must prepare and make
6-1
-------
available an initial regulatory flexibility analysis, unless it certifies that the proposed rule, if
promulgated, will not have a significant economic impact on a substantial number of small
entities (5 U.S.C. 605(b)). Small entities include small businesses, small organizations, and small
governmental jurisdictions.
EPA conducted regulatory flexibility analysis at the ultimate (i.e., highest) level of
ownership, evaluating parent entities with the largest share of ownership in at least one
potentially-affected EGU included in EPA's base case using the IPM v.6, used in this RIA.1 This
analysis draws on the "parsed" unit-level estimates using IPM results for 2021, as well as
ownership, employment, and financial information for the potentially affected small entities
drawn from other resources described in more detail below. This analysis is focused on
estimating impacts in 2021 because implementation of the illustrative EGU controls occurs in the
2021 ozone season.
EPA identified the size of ultimate parent entities by using the Small Business
Administration (SBA) size standard guidelines.2 The criteria for size determination vary by the
organization/operation category of the ultimate parent entity, as follows:
• Privately-owned (non-government) entities (see Table 6-1)
o Privately-owned entities include investor-owned utilities, non-utility entities,
and entities with a primary business other than electric power generation.
o For entities with electric power generation as a primary business, small entities
are those with less than the threshold number of employees specified by SBA
for each of the relevant North American Industry Classification System
(NAICS) sectors (NAICS 2211).
1	Detailed documentation for IPM v.6 is available at: http://www.epa.gov/airmarkets/powersectormodeling.html.
2	U.S. Small Business Administration (SBA). 2019. Small Business Size Standards, effective as of August 19, 2019
and available at the following link: https://www.sba.gov/sites/default/files/2019-
08/SBA%20Table%20of%20Size%20Standards_Effective%20Aug%2019%2C%202019.pdf.
6-2
-------
o For entities with a primary business other than electric power generation, the
relevant size criteria are based on revenue, assets, or number of employees by
NAICS sector.3
•	Publicly-owned entities
o Publicly-owned entities include federal, state, municipal, and other political
subdivision entities.
o The federal and state governments are considered to be large. Municipalities
and other political units with populations fewer than 50,000 ae considered to be
small.
•	Rural Electric Cooperatives
o Small entities are those with fewer than the threshold level of employees or
revenue specified by SB A for each of the relevant NAICS sectors.
6.3.1 Identification of Small Entities
In this analysis, EPA considered EGUs that meet the following five criteria: 1) EGU is
represented in NEEDS v6; 2) EGU is fossil fuel-fired; 3) EGU is located in a state covered by
this rule; 4) EGU is neither a cogeneration unit nor solid waste incineration unit; and 5) EGU
capacity is 25 MW or larger. EPA next refined this list of EGUs, narrowing it to those that
exhibit at least one of the following changes, in comparison to the baseline.
•	Summer fuel use (BTUs) changes by +/- 1 percent or more
•	Summer generation (GWh) changes by +/- 1 percent or more
•	NOx summer emissions (tons) changes by +/- 1 percent or more
Based on these criteria, EPA identified a total of 97 potentially affected EGUs warranting
examination in this RFA analysis. Next, we determined power plant ownership information,
including the name of associated owning entities, ownership shares, and each entity's type of
ownership. We primarily used data from Ventyx, supplemented by limited research using
3 Certain affected EGUs are owned by ultimate parent entities whose primary business is not electric power
generation.
6-3
-------
publicly available data.4 Majority owners of power plants with affected EGUs were categorized
as one of the seven ownership types.5 These ownership types are:
1.	Investor-Owned Utility (IOU): Investor-owned assets (e.g., a marketer, independent
power producer, financial entity) and electric companies owned by stockholders, etc.
2.	Cooperative (Co-Op): Non-profit, customer-owned electric companies that generate
and/or distribute electric power.
3.	Municipal: A municipal utility, responsible for power supply and distribution in a small
region, such as a city.
4.	Sub-division: Political subdivision utility is a county, municipality, school district,
hospital district, or any other political subdivision that is not classified as a municipality
under state law.
5.	Private: Similar to an investor-owned utility, however, ownership shares are not openly
traded on the stock markets.
6.	State: Utility owned by the state.
7.	Federal: Utility owned by the federal government.
Next, EPA used both the D&B Hoover's online database and the Ventyx database to
identify the ultimate owners of power plant owners identified in the Ventyx database. This was
necessary, as many majority owners of power plants (listed in Ventyx) are themselves owned by
other ultimate parent entities (listed in D&B Hoover's).6 In these cases, the ultimate parent entity
was identified via D&B Hoover's, whether domestically or internationally owned.
EPA followed SB A size standards to determine which non-government ultimate parent
entities should be considered small entities in this analysis. These SBA size standards are
specific to each industry, each having a threshold level of either employees, revenue, or assets
4	The Ventyx Energy Velocity Suite database consists of detailed ownership and corporate affiliation information at
the EGU level. For more information, see: www.ventyx.com.
5	Throughout this analysis, EPA refers to the owner with the largest ownership share as the "majority owner" even
when the ownership share is less than 51 percent.
6	The D&B Hoover's online platform includes company records that can contain NAICS codes, number of
employees, revenues, and assets. For more information, see: https://www.dnb.com/products/marketing-sales/dnb-
hoovers.html.
6-4
-------
below which an entity is considered small. SBA guidelines list all industries, along with their
associated NAICS code and SBA size standard. Therefore, it was necessary to identify the
specific NAICS code associated with each ultimate parent entity in order to understand the
appropriate size standard to apply. Data from D&B Hoover's was used to identify the NAICS
codes for most of the ultimate parent entities. In many cases, an entity that is a majority owner of
a power plant is itself owned by an ultimate parent entity with a primary business other than
electric power generation. Therefore, it was necessary to consider SBA entity size guidelines for
the range of NAICS codes listed in Table 6-1. This table represents the range of NAICS codes
and areas of primary business of ultimate parent entities that are majority owners of potentially
affected EGUs in EPA's IPM base case.
6-5
-------
Table 6-1. SBA Size Standards by NAICS Code


Size
Size


Standards
Standards
NAICS

(millions of
(number of
Codes
NAICS U.S. Industry Title
dollars)
employees)
221111
Hydroelectric Power Generation

500
221112
Fossil Fuel Electric Power Generation

750
221113
Nuclear Electric Power Generation

750
221114
Solar Electric Power Generation

250
221115
Wind Electric Power Generation

250
221116
Geothermal Electric Power Generation

250
221117
Biomass Electric Power Generation

250
221118
Other Electric Power Generation
Electric Bulk Power Transmission and

250
221121
Control

500
221122
Electric Power Distribution

1000
221210
Natural Gas Distribution

1000
221310
Water Supply and Irrigation Systems
$30

221320
Sewage Treatment Facilities
$22

221330
Steam and Air-Conditioning Supply
$16

Note: Based on size standards effective at the time EPA conducted this analysis (SBA size standards, effective
August 19, 2019. Available at the following link: https://www.sba.gov/document/support--table-size-standards).
Source: SBA, 2019
EPA compared the relevant entity size criterion for each ultimate parent entity to the SBA
size standard noted in Table 6-1. We used the following data sources and methodology to
estimate the relevant size criterion values for each ultimate parent entity:
1. Employment, Revenue, and Assets: EPA used the D&B Hoover's database as the
primary source for information on ultimate parent entity employee numbers, revenue, and
assets.7 In parallel, EPA also considered estimated revenues from affected EGUs based
on analysis of parsed-file estimates for the rule. EPA assumed that the ultimate parent
entity revenue was the larger of the two revenue estimates. In limited instances,
supplemental research was also conducted to estimate an ultimate parent entity's number
of employees, revenue, or assets.
7 Estimates of sales were used in lieu of revenue estimates when revenue data was unavailable.
6-6
-------
2. Population: Municipal entities are defined as small if they serve populations of less than
50,000. EPA primarily relied on data from the Ventyx database and the U.S. Census
Bureau to inform this determination.
Ultimate parent entities for which the relevant measure is less than the SBA size standard
were identified as small entities and carried forward in this analysis. In total EPA identified 89
potentially affected EGUs, owned by 15 entities. Of these, EPA identified 7 potentially affected
EGUs owned by 2 small entities8 included in EPA's Base Case.
6.3.2 Overview of Analysis and Results
This section presents the methodology and results for estimating the impact of the
Revised CSAPR Update on small entities in 2021 based on the following endpoints:
•	annual economic impacts of the Revised CSAPR Update on small entities, and
•	ratio of small entity impacts to revenues from electricity generation.
6.3.2.1 Methodology for Estimating Impacts of the Revised CSAPR Update on Small Entities
An entity can comply with the Revised CSAPR Update through some combination of the
following: optimizing existing SCRs, turning on idled SCR controls, optimizing existing SNCR
controls, using allocated allowances, purchasing allowances, or reducing emissions through a
reduction in generation. Additionally, units with more allowances than needed can sell these
allowances in the market. The chosen compliance strategy will be primarily a function of the
unit's marginal control costs and its position relative to the marginal control costs of other units.
To attempt to account for each potential control strategy, EPA estimates compliance costs
as follows:
Ccompliance A Coperating+Retrofit A CFuel A CAllowances A (^Transaction A R
8 Both of these small entities are in NAICS 221118, which is defined as establishments primarily engaged in
operating electric power generation facilities (except hydroelectric, fossil fuel, nuclear, solar, wind, geothermal,
biomass). These facilities convert other forms of energy, such as tidal power, into electric energy.
6-7
-------
where C represents a component of cost as labeled, and A R represents the value of foregone
electricity generation, calculated as the difference in revenues between the base case and the
Revised CSAPR Update in 2021.
Realistically, compliance choices and market conditions can combine such that an entity
may actually experience a savings in any of the individual components of cost. Under the
Revised CSAPR Update, some units will forgo some level of electricity generation (and thus
revenues) to comply and this impact will be lessened on these entities by the projected increase
in electricity prices under the Revised CSAPR Update. On the other hand, those increasing
generation levels will see an increase in electricity revenues and as a result, lower net compliance
costs. If entities are able to increase revenue more than an increase in fuel cost and other
operating costs, ultimately, they will have negative net compliance costs (or savings). Overall,
small entities are not projected to install relatively costly emissions control retrofits but may
choose to do so in some instances. Because this analysis evaluates the total costs along each of
the compliance strategies laid out above for each entity, it inevitably captures savings or gains
such as those described. As a result, what we describe as cost is really more of a measure of the
net economic impact of the rule on small entities.
For this analysis, EPA used IPM-parsed output to estimate costs based on the parameters
above, at the unit level. These impacts were then summed for each small entity, adjusting for
ownership share. Net impact estimates were based on the following: operating and retrofit costs,
sale or purchase of allowances, and the change in fuel costs or electricity generation revenues
under the Revised CSAPR Update relative to the base case. These individual components of
compliance cost were estimated as follows:
(1) Operating and retrofit costs: Using engineering analytics, EPA identified which
compliance option was selected by each EGU in 2021 (i.e., SCR optimization or
turning on existing SCR controls) and applied the appropriate cost to this choice.
EPA assumes that state of the art combustion controls may be installed in 2022
and are not part of the controls available in 2021. As part of the illustrative
emission budgets modeled for this rule, SNCR optimization was included from
2022 forward.
6-8
-------
(2)	Sale or purchase of allowances: To estimate the value of allowances holdings,
allocated allowances were subtracted from projected emissions, and the difference
was then multiplied by $1,600 (2016$) per ton, which is the marginal cost of NOx
reductions used to set the modeled budgets in the Revised CSAPR Update. While
this is a reasonable approximation, it is possible that the actual allowance price
could be lower. Units were assumed to purchase or sell allowances to exactly
cover their projected emissions under the Revised CSAPR Update.
(3)	Fuel costs: The change in fuel expenditures under the Revised CSAPR Update
was estimated by taking the difference in projected fuel expenditures between the
IPM estimates for the Revised CSAPR Update and the base case.
(4)	Value of electricity generated: To estimate the value of electricity generated, the
projected level of electricity generation is multiplied by the regional-adjusted
retail electricity price ($/MWh) estimate, for all entities except those categorized
as private in Ventyx. For private entities, EPA used the wholesale electricity price
instead of the retail electricity price because most of the private entities are
independent power producers (IPP). IPPs sell their electricity to wholesale
purchasers and do not own transmission facilities. Thus, their revenue was
estimated with wholesale electricity prices.
(5)	Administrative costs: Because most affected units are already monitored as a
result of other regulatory requirements, EPA considered the primary
administrative cost to be transaction costs related to purchasing or selling
allowances. EPA assumed that transaction costs were equal to 1.5 percent of the
total absolute value of the difference between a unit's allocation and projected
NOx emissions. This assumption is based on market research by ICF.
6.3.2.2 Results
The potential impacts of the Revised CSAPR Update on small entities are summarized in
Table 6-2. All costs are presented in 2016$. EPA estimated the annual net compliance cost to
small entities to be approximately $0.04 million in 2021.
6-9
-------
Table 6-2. Projected Impact of the Revised CSAPR Update on Small
EGU
Ownership
Type
Number of
Potentially
Affected
Entities
Total Net
Compliance
Cost
($2016
millions)
Number of Small
Entities with
Compliance Costs
>1% of Generation
Revenues
Number of Small
Entities with
Compliance Costs
>3% of Generation
Revenues
Cooperative
1
0.04
0
0
Private
1
0.00
0
0
Total
2
0.04
0
0
Entities in 2021
Source: IPM analysis
EPA assessed the economic and financial impacts of the rule using the ratio of compliance
costs to the value of revenues from electricity generation, focusing in particular on entities for
which this measure is greater than 1 percent. Although this metric is commonly used in EPA
impact analyses, it makes the most sense when as a general matter an analysis is looking at small
businesses that operate in competitive environments.9 However, small businesses in the electric
power industry often operate in a price-regulated environment where they are able to recover
expenses through rate increases. Given this, EPA considers the 1 percent measure in this case a
crude measure of the price increases these small entities will be asking of rate commissions or
making at publicly owned companies. Of the 2 small entities considered in this analysis, neither
is projected to experience compliance costs greater than 1 percent of generation revenues in
2021. EPA has concluded that there is no significant economic impact on a substantial number of
small entities (no SISNOSE) for this rule.
The separate components of annual costs to small entities under the Revised CSAPR
Update are summarized in Table 6-3. The most significant components of incremental cost to the
cooperative category under the Revised CSAPR Update are due to higher operating costs
(reflecting the cost of controls). Among the private category, however, reduced generation is the
key driver. Total impacts to the private category are well below $10,000.
9 U.S. EPA. EPA's Action Development Process. Final Guidance for EPA Rulewriters: Regulatory Flexibility Act as
Amended by the Small Business Regulatory Enforcement Fairness Act. September 2006. Available at
https://www.epa.gov/sites/production/files/2015-06/documents/guidance-regflexact.pdf.
6-10
-------
Table 6-3. Incremental Annual Costs under the Revised CSAPR Update Summarized by
Ownership Group and Cost Category in 2021 (2016$ millions)
EGU
Ownership
Type
Operating
Cost
Net Purchase
of Allowances
Fuel Cost
Lost
Electricity
Revenue
Administrative
Cost
Cooperative
0.06
0.00
-0.02
0.00
0.00
Private
0.00
0.00
0.00
0.00
0.00
Source: IPM analysis
6.3.3 Summary of Small Entity Impacts
EPA examined the potential economic impacts to small entities associated with this rule
based on assumptions of how the affected states will implement control measures to meet their
emissions. To summarize, of the 2 small entities potentially affected, none are projected to
experience compliance costs in excess of 1 percent of revenues in 2021, based on assumptions of
how the affected states implement control measures to meet their emissions budgets as set forth
in this rule.
EPA has lessened the impacts for small entities by excluding all units smaller than 25 MW.
This exclusion, in addition to the exemptions for cogeneration units and solid waste incineration
units, eliminates the burden of higher costs for a substantial number of small entities located in
the 12 states for which EPA is promulgating FIPs.
6.4 Unfunded Mandates Reform Act
Title II of the Unfunded Mandates Reform Act of 1995 (Public Law 104-4) (UMRA)
establishes requirements for federal agencies to assess the effects of their regulatory actions on
State, local, and Tribal governments and the private sector. Under section 202 of the UMRA, 2
U.S.C. 1532, EPA generally must prepare a written statement, including a cost-benefit analysis,
for any proposed or final rule that includes any Federal mandate that may result in the
expenditure by State, local, and Tribal governments, in the aggregate, or by the private sector, of
$100 million or more in any one year. A Federal mandate is defined under section 421(6) of the
UMRA, 2 U.S.C. 658(6), to be either a Federal intergovernmental mandate or a Federal private
sector mandate, as defined by the UMRA. A Federal intergovernmental mandate, in turn, is
6-11
-------
defined to include a regulation that would impose an enforceable duty upon State, Local, or
Tribal governments, UMRA section 421(5)(A)(i), 2 U.S.C. 658(5)(A)(i), except for, among other
things, a duty that is a condition of Federal assistance, UMRA section 421(5)(A)(i)(I). A Federal
private sector mandate includes a regulation that would impose an enforceable duty upon the
private sector, with certain exceptions, UMRA section 421(7)(A), 2 U.S.C. 658(7)(A).
This final action does not contain an unfunded mandate of $100 million or more as
described in UMRA, 2 U.S.C. 1531-1538, and will not significantly or uniquely affect small
governments. Note that EPA expects the final rule to potentially have an impact on only one
category of government-owned entities (municipality-owned entities). This analysis does not
examine potential indirect economic impacts associated with the final rule, such as employment
effects in industries providing fuel and pollution control equipment, or the potential effects of
electricity price increases on government entities.
6.5	Executive Order 13132: Federalism
This final action does not have federalism implications. As finalized, this final action will
not have substantial direct effects on the states, on the relationship between the national
government and the states, or on the distribution of power and responsibilities among the various
levels of government.
6.6	Executive Order 13175: Consultation and Coordination with Indian Tribal
Governments
This action has tribal implications. However, it will neither impose substantial direct
compliance costs on federally recognized tribal governments, nor preempt tribal law.
This final action implements EGUNOx ozone season emission reductions in 12 eastern
states (Illinois, Indiana, Kentucky, Louisiana, Maryland, Michigan, New Jersey, New York,
Ohio, Pennsylvania, Virginia, and West Virginia.). However, at this time, none of the existing or
planned EGUs affected by this rule are owned by tribes or located in Indian country. This action
may have tribal implications if a new affected EGU is built in Indian country. Additionally,
tribes have a vested interest in how this rule affects air quality.
6-12
-------
In developing the CSAPR, which was promulgated on July 6, 2011, to address interstate
transport of ozone pollution under the 1997 ozone NAAQS, EPA consulted with tribal officials
under the EPA Policy on Consultation and Coordination with Indian Tribes early in the process
of developing that regulation to allow for meaningful and timely tribal input into its
development. A summary of that consultation is provided at 76 FR 48346.
In that rulemaking, EPA received comments from several tribal commenters regarding the
availability of the CSAPR allowance allocations to new units in Indian country. EPA responded
to these comments by instituting Indian country new unit set-asides in the final CSAPR. In order
to protect tribal sovereignty, these set-asides are managed and distributed by the federal
government regardless of whether the CSAPR in the adjoining or surrounding state is
implemented through a FIP or SIP. While there are no existing affected EGUs in Indian country
covered by this action, the Indian country set-asides will ensure that any future new units built in
Indian country will be able to obtain the necessary allowances. This rule maintains the Indian
country new unit set-aside and adjusts the amounts of allowances in each set-aside according to
the same methodology of the CSAPR and the CSAPR Update.
EPA consulted with tribal officials early in the process of developing this rule in
accordance with the EPA Policy on Consultation and Coordination with Indian Tribes (May
2011). Before proposing this rule, EPA informed tribes of the rule's development on a National
Tribal Air Association (NTAA) monthly air policy conference call that took place on June 25,
2020. In a separate NTAA call on October 20, 2020, EPA gave an overview of the proposed rule.
In order to permit tribes to have meaningful and timely input into the development of the final
rule, EPA offered consultation to tribal leaders. On October 30, 2020, EPA sent out letters via
electronic mail to all 574 federally recognized tribes informing them of this action, offering
consultation and requesting comment on this rulemaking. Courtesy copies of the letters were also
sent via email to tribal air staff and tribal environmental professionals. EPA also sent courtesy
copies to EPA's Regional Tribal Air Coordinators for notification to their tribes. To further
provide tribes with the resources that they might require to engage in effective consultation, EPA
also held an informational webinar on the rule on November 9, 2020. EPA did not receive any
requests for consultation on this rule.
6-13
-------
6.7	Executive Order 13045: Protection of Children from Environmental Health &
Safety Risks
EPA interprets Executive Order 13045 as applying only to those regulatory actions that
concern environmental health or safety risks that the EPA has reason to believe may
disproportionately affect children, per the definition of "covered regulatory action" in section 2-
202 of the Executive Order. This action is not subject to Executive Order 13045 because it
implements a previously promulgated health-based federal standard. This action's health and risk
assessments are contained in Chapter 5 of the accompanying RIA. EPA believes that the ozone
reductions, PM2.5 reductions, and CO2 reductions from this final rule will further improve
children's health.
6.8	Executive Order 13211: Actions that Significantly Affect Energy Supply,
Distribution, or Use
This action is not a "significant energy action" because it is not likely to have a significant
adverse effect on the supply, distribution, or use of energy. EPA has prepared a Statement of
Energy Effects for the regulatory control alternative as follows. The Agency estimates a much
less than 1 percent change in retail electricity prices on average across the contiguous U.S. in
2021, and a much less than 1 percent reduction in coal-fired electricity generation in 2021 as a
result of this rule. EPA projects that utility power sector delivered natural gas prices will change
by less than 1 percent in 2021. For more information on the estimated energy effects, please see
Chapter 4 of this RIA.
6.9	National Technology Transfer and Advancement Act
The rulemaking does not involve technical standards.
6.10	Executive Order 12898: Federal Actions to Address Environmental Justice in
Minority Populations and Low-Income Populations
Because of the need to meet the court-ordered signature deadline on this action, EPA did
not have sufficient time to undertake a definitive assessment of the impacts of this final rule on
minority populations, low-income populations and/or indigenous peoples, as specified in
Executive Order 12898 (59 FR 7629, February 16, 1994). EPA does not have information at this
time that would suggest that this rule has the potential to result in disproportionately high and
6-14
-------
adverse human health or environmental impacts on vulnerable populations or overburdened
communities; however, EPA is also not currently in a position to make a determination to this
effect. For a longer discussion of a framework for assessing potential EJ concerns for future
rulemakings, please see Section X.J. of the final rule preamble.
Ozone pollution from power plants has both local and regional components: part of the
pollution in a given location—even in locations near emission sources—is due to emissions from
nearby sources and part is due to emissions that are transported in the atmosphere over large
distances and mix with emissions from other sources. Undertaken to implement CAA section
110(a)(2)(D), this action addresses that "significant" portion of contribution from upwind states
to a nonattainment or maintenance receptor. As a result, the rule will reduce exposures to ozone
in areas that are struggling to attain or maintain the 2008 ozone NAAQS. By addressing
maintenance receptors, this rule reduces the likelihood that areas close to the level of the
standard will exceed the current health-based standards in the future. The rule will result in
incidental reductions in ozone in other areas, as well as reducing emissions of PM and other
pollutants from EGUs that have both localized and distant impacts.
At the same time, this action alone cannot fully resolve any disproportionate impacts of
ozone levels in downwind areas. Rather, it eliminates upwind state "significant contribution,"
thus ameliorating those conditions and improving downwind air quality. While this rule is
expected to reduce interstate ozone transport and thus to yield overall health and environmental
benefits, further analysis would be required to assess potential environmental justice concerns -
including, for example, whether the downwind air quality benefits are equitably distributed.10
It is important to note that nothing in this final rule allows sources to violate their title V
permit or any other federal, state, or local emissions or air quality requirements. Moreover, CAA
section 110(a)(2)(D) addresses transport of criteria pollutants between states and is only one of
many provisions of the CAA that provide EPA, states, and local governments with authorities to
reduce exposure to ozone in communities. These legal authorities work together to reduce
10 A potential environmental justice concern is "the actual or potential lack of fair treatment or meaningful
involvement of minority populations, low-income populations, tribes, and indigenous peoples in the development,
implementation and enforcement of environmental laws, regulations and policies" EPA, Guidance on Considering
Environmental Justice During the Development of Regulatory Actions (May 2015).
6-15
-------
exposure to these pollutants in communities, including for minority, low-income, and tribal
populations, and provide substantial health benefits to both the general public and sensitive sub-
populations.
EPA informed tribal communities of its development of this rule on a National Tribal Air
Association - EPA air policy conference call on June 25, 2020. EPA also held two informational
webinars for tribes and environmental justice communities on November 9, 2020 and November
10, 2020, respectively, where EPA presented an overview of the rule and provided tribes and
communities with resources that they might require to engage in the public comment process.
6-16
-------
CHAPTER 7: COMPARISON OF BENEFITS AND COSTS
Overview
EPA performed an analysis to estimate the costs and benefits of compliance with the
Revised CSAPR Update and more and less stringent alternatives. EPA is promulgating electric
generating unit (EGU) oxides of nitrogen (NOx) ozone season emissions budgets for 12 states.1
This action finds that for these states, their projected 2021 ozone season NOx emissions
significantly contribute to downwind states' nonattainment and/or maintenance problems for the
2008 ozone national ambient air quality standards (NAAQS). For these 12 states, EPA amends
their federal implementation plans (FIPs) to revise the existing Cross-State Air Pollution Rule
(CSAPR) NOx Ozone Season Group 2 emissions budgets for EGUs and implement the revised
budgets beginning in the 2021 ozone season (May 1, 2021 - September 30, 2021) via a new
CSAPR NOx Ozone Season Group 3 Trading Program.
The Revised CSAPR Update state budgets reflect the optimization of existing selective
catalytic reduction (SCR), selective non-catalytic reduction (SNCR) controls and installation of
state-of-the-art NOx combustion controls, at an estimated representative cost of $1,800 per ton
(2016$). For the RIA, in order to implement the OMB Circular A-4 requirement for fulfilling
Executive Order (E.O.) 12866 to assess one less stringent and one more stringent alternative to
the rule, EPA is also analyzing EGU NOx ozone season emissions budgets reflecting NOx
reduction strategies that are widely available at a uniform cost of $9,600 per ton (2016$) and
strategies that are widely available at a uniform cost of $500 per ton (2016$). These alternatives
are used to illustrate the monetized cost and benefit impacts of varying program stringency. They
are designed to show the effects of more stringent and less stringent NOx reduction requirements
in a regulatory structure that is otherwise the same as the final NOx emissions budgets. We show
the results for 2021 to reflect the year in which implementation of this rule begins, for 2025 to
reflect full implementation of the rule, and 2030 to show the continued costs and benefits of the
rule. This RIA evaluates how the EGUs covered by the rule are expected to reduce their
emissions in response to the requirements and flexibilities provided by the remedy implemented
1 The 12 states are Illinois, Indiana, Kentucky, Louisiana, Maryland, Michigan, New Jersey, New York, Ohio,
Pennsylvania, Virginia, and West Virginia.
7-1
-------
by the Revised CSAPR Update and the benefits, costs and impacts of their expected compliance
behavior. This chapter summarizes these results.
7.1 Results
The rule and regulatory control alternatives' compliance costs are estimated using the IPM
model and an evaluation of control technologies evaluated outside of IPM. As shown in Chapter
4, the estimated annual compliance costs to implement the rule, as described in this document,
are approximately $5 million in 2021 and $2 million in 2025 (2016$). As described in Section
4.5, this RIA uses compliance costs as a proxy for social costs. As shown in Chapter 5, the
estimated monetized health benefits from implementation of the rule are approximately $230 and
$1,900 million in 2021 (2016$, based on a real discount rate of 3 percent). For 2025, the
estimated monetized health benefits from implementation of the rule are approximately $320 and
$2,400 million (2016$, based on a real discount rate of 3 percent). The two estimates of the
benefits and net-benefits for each discount rate reflect alternative ozone and PM2.5 mortality risk
estimates. The estimated monetized climate benefits are $1 million in 2021 (using a 3 percent
discount rate) and $330 million in 2025 (using a 3 percent discount rate). We present the costs
and benefits for the years 2021 through 2040 at real discount rates of 3 and 7 percent in Table 7-
4.
EPA calculates the net benefits of the rule by subtracting the estimated compliance costs
from the estimated benefits in 2021, 2025, and 2030. The benefits include those to public health
and climate. The annual net benefits of the rule in 2021 (in 2016$) are approximately $230 and
$1,900 million using a 3 percent real discount rate. The annual net benefits of the rule in 2025
are approximately $650 and $2,700 using a 3 percent real discount rate. The annual net benefits
of the rule in 2030 are approximately $650 and $2,900 million using a 3 percent real discount
rate. Table 7-1 presents a summary of the health benefits, climate benefits, costs, and net benefits
of the rule and the more and less stringent alternatives for 2021. Table 7-2 presents a summary of
these impacts for the Revised CSAPR Update and the more and less stringent alternatives for
2025. Table 7-3 presents a summary of these impacts for the rule and the more and less stringent
alternatives for 2030.
7-2
-------
Table 7-1. Benefits, Costs, and Net Benefits of the Final Rule and More and Less Stringent
	Alternatives for 2021 for the U.S. (millions of 2016$)a'b'c	
, _ .	More Stringent Less Stringent
Final Rule	... ..s	... .?
Alternative	Alternative
Health Benefits (3%)	$230 and $1,900	$260 and $1,900	$20 and $190
Climate Benefits (3%)	$1	$2	$1
Total Benefits	$230 and $1,900	$260 and $1,900	$20 and $190
Costs	$5	$5	$2
Net Benefits	$230 and $1,900	$260 and $1,900	$20 and $190
Health Benefits (7%)	$200 and $1,700	$200 and $1,700	$20 and $170
Climate Benefits (3%)	$1	$2	$1
Total Benefits	$200 and $1,700	$200 and $1,700	$20 and $170
Costs	$5	$5	$2
Net Benefits	$200 and $1,700	$200 and $1,700	$20 and $170
a We focus results to provide a snapshot of costs and benefits in 2021, using the best available information to
approximate social costs and social benefits recognizing uncertainties and limitations in those estimates. The two
benefits estimates are separated by the word "and" to signify that they are two separate estimates. The estimates do
not represent lower- and upper-bound estimates and should not be summed.
b Benefits include those related to public health and climate. The health benefits are associated with several point
estimates and are presented at real discount rates of 3 and 7 percent. Climate benefits are based on changes
(reductions) in CO2 emissions and are calculated using four different estimates of the social cost of carbon (SC-CO2)
(model average at 2.5 percent, 3 percent, and 5 percent discount rates; 95th percentile at 3 percent discount rate). For
the presentational purposes of this table, we show the benefits associated with the average SC-CO2 at a 3 percent
discount rate, but the Agency does not have a single central SC-CO2point estimate. We emphasize the importance
and value of considering the benefits calculated using all four SC-CO2 estimates; the additional benefit estimates
range from $0.24 million to $2.31 million in 2021 for the finalized option. Please see Table 5-9 for the full range of
SC-CO2 estimates. As discussed in Chapter 5, a consideration of climate benefits calculated using discount rates
below 3 percent, including 2 percent and lower, are also warranted when discounting intergenerational impacts. The
costs presented in this table are 2021 annual estimates for each alternative analyzed.
0 Rows may not appear to add correctly due to rounding.
Table 7-2. Benefits, Costs, and Net Benefits of the Final Rule and More and Less Stringent
	Alternatives for 2025 for the U.S. (millions of 2016$)a'b'c
, _ .	More Stringent Less Stringent
Final Rule	... ,.s	... ,?
Alternative	Alternative
Health Benefits (3%)	$320 and $2,400	$540 and $4,200	$20 and $200
Climate Benefits (3%) $330	$770	$250
Total Benefits	$650 and $2,700	$1,300 and $5,000	$270 and $450
	Costs	$2	$4	-$15
Net Benefits	$650 and $2,700	$1,300 and $5,000	$280 and $460
Health Benefits (7%)	$290 and $2,200	$490 and $3,800	$20 and $170
Climate Benefits (3%) $330	$770	$250
Total Benefits	$620 and $2,500	$1,300 and $4,600	$270 and $420
Costs $2	$4	-$15
Net Benefits	$620 and $2,500	$1,300 and $4,500	$280 and $430
7-3
-------
" We focus results to provide a snapshot of costs and benefits in 2025, using the best available information to
approximate social costs and social benefits recognizing uncertainties and limitations in those estimates. The two
benefits estimates are separated by the word "and" to signify that they are two separate estimates. The estimates do
not represent lower- and upper-bound estimates and should not be summed.
b Benefits include those related to public health and climate. The health benefits are associated with several point
estimates and are presented at real discount rates of 3 and 7 percent. Climate benefits are based on changes
(reductions) in CO2 emissions and are calculated using four different estimates of the social cost of carbon (SC-CO2)
(model average at 2.5 percent, 3 percent, and 5 percent discount rates; 95th percentile at 3 percent discount rate). For
the presentational purposes of this table, we show the benefits associated with the average SC-CO2 at a 3 percent
discount rate, but the Agency does not have a single central SC-CO2point estimate. We emphasize the importance
and value of considering the benefits calculated using all four SC-CO2 estimates; the additional benefit estimates
range from $109 million to $1,011 million in 2025 for the finalized option. Please see Table 5-9 for the full range of
SC-CO2 estimates. As discussed in Chapter 5, a consideration of climate benefits calculated using discount rates
below 3 percent, including 2 percent and lower, are also warranted when discounting intergenerational impacts. The
costs presented in this table are 2025 annual estimates for each alternative analyzed.
0 Rows may not appear to add correctly due to rounding.
Table 7-3. Benefits, Costs, and Net Benefits of the Final Rule and More and Less Stringent
	Alternatives for 2030 for the U.S. (millions of 2016$)a'b'c	
Final Rule
More Stringent
Alternative
Less Stringent
Alternative
Health Benefits (3%)
Climate Benefits (3%)
Total Benefits
Costs
$340 and $2,600
$370
$710 and $3,000
$64
$590 and $4,600
$940
$1,500 and $5,500
$32
$30 and $210
$270
$300 and $480
$67
Net Benefits
$650 and $2,900 $1,500 and $5,500 $230 and $410
Health Benefits (7%)
Climate Benefits (3%)
Total Benefits
Costs
$330 and $2,500
$370
$700 and $2,900
$64
$560 and $3,900
$940
$1500 and $4,800
$32
$20 and $180
$270
$290 and $450
$67
Net Benefits
$640 and $2,800 $1,500 and $4,800 $220 and $380
a We focus results to provide a snapshot of costs and benefits in 2030, using the best available information to
approximate social costs and social benefits recognizing uncertainties and limitations in those estimates. The two
benefits estimates are separated by the word "and" to signify that they are two separate estimates. The estimates do
not represent lower- and upper-bound estimates and should not be summed.
b Benefits include those related to public health and climate. The health benefits are associated with several point
estimates and are presented at real discount rates of 3 and 7 percent. Climate benefits are based on changes
(reductions) in CO2 emissions and are calculated using four different estimates of the social cost of carbon (SC-CO2)
(model average at 2.5 percent, 3 percent, and 5 percent discount rates; 95th percentile at 3 percent discount rate). For
the presentational purposes of this table, we show the benefits associated with the average SC-CO2 at a 3 percent
discount rate, but the Agency does not have a single central SC-CO2point estimate. We emphasize the importance
and value of considering the benefits calculated using all four SC-CO2 estimates; the additional benefit estimates
range from $128 million to $1,146 million in 2030 for the finalized option. Please see Table 5-9 for the full range of
SC-CO2 estimates. As discussed in Chapter 5, a consideration of climate benefits calculated using discount rates
below 3 percent, including 2 percent and lower, are also warranted when discounting intergenerational impacts. The
costs presented in this table are 2030 annual estimates for each alternative analyzed.
0 Rows may not appear to add correctly due to rounding.
7-4
-------
As part of fulfilling analytical guidance with respect to E.O. 12866, EPA presents
estimates of the present value (PV) of the benefits and costs over the twenty-year period 2021 to
2040. To calculate the present value of the social net-benefits of the Revised CSAPR Update,
annual benefits and costs are discounted to 2021 at 3 percent and 7 discount rates as directed by
OMB's Circular A-4. The EPA also presents the equivalent annualized value (EAV), which
represents a flow of constant annual values that, had they occurred in each year from 2021 to
2040, would yield a sum equivalent to the PV. The EAV represents the value of a typical cost or
benefit for each year of the analysis, in contrast to the year-specific estimates mentioned earlier
in the RIA.
For the twenty-year period of 2021 to 2040, the PV of the net benefits, in 2016$ and
discounted to 2021, is $8,800 and $41,000 million when using a 3 percent discount rate and
$7,300 and $29,000 million when using a 7 percent discount rate. The EAV is $590 and $2,800
million per year when using a 3 percent discount rate and $570 and $2,700 million when using a
7 percent discount rate. The comparison of benefits and costs in PV and EAV terms for the rule
can be found in Table 7-4. Estimates in the table are presented as rounded values and based on
air quality simulations run for years 2021 and 2024.
7-5
-------
Table 7-4. Summary of Annual Values, Present Values and Equivalent Annualized Values for the 2021-2040 Timeframe for
Estimated Compliance Costs, Benefits, and Net Benefits for the Final Rule (millions of 2016$, discounted to 2021)a'b

Health Benefits
Climate
Benefits0
Cost"
Net Benefits

3%
7%
3%
3%
7%
3%
7%
2021*
$230 and $1,900
$200 and $1,700
$1
$5
$230 and $1,900
$200 and $1,700
2022
$230 and $2,000
$210 and $1,600
$140
$19
$350 and $2,100
$330 and $1,700
2023
$230 and $2,000
$210 and $1,600
$290
$19
$500 and $2,300
$480 and $1,900
2024*
$310 and $2,400
$280 and $2,100
$310
$2
$620 and $2,700
$590 and $2,400
2025
$320 and $2,400
$290 and $2,200
$330
$2
$650 and $2,700
$620 and $2,500
2026
$330 and $2,500
$290 and $2,200
$340
$1
$670 and $2,800
$630 and $2,500
2027
$320 and $2,400
$300 and $2,300
$350
$0
$670 and $2,800
$650 and $2,700
2028
$330 and $2,500
$310 and $2,400
$360
$66
$620 and $2,800
$600 and $2,700
2029
$330 and $2,500
$320 and $2,400
$360
$65
$630 and $2,800
$620 and $2,700
2030
$340 and $2,600
$330 and $2,500
$370
$64
$650 and $2,900
$640 and $2,800
2031
$350 and $2,600
$340 and $2,600
$350
$64
$640 and $2,900
$630 and $2,900
2032
$360 and $2,700
$350 and $2,600
$330
$63
$630 and $3,000
$620 and $2,900
2033
$350 and $2,600
$360 and $2,700
$310
$18
$640 and $2,900
$650 and $3,000
2034
$360 and $2,700
$370 and $2,800
$290
$18
$630 and $3,000
$640 and $3,100
2035
$370 and $2,800
$380 and $2,800
$270
$18
$620 and $3,100
$630 and $3,100
2036
$370 and $2,800
$390 and $2,900
$290
$18
$640 and $3,100
$660 and $3,200
2037
$380 and $2,900
$400 and $3,000
$300
$18
$660 and $3,200
$680 and $3,300
2038
$370 and $2,800
$410 and $3,100
$310
$s

$670 and $3,100
$710 and $3,400
2039
$380 and $2,800
$430 and $3,200
$330
$s

$700 and $3,100
$750 and $3,500
2040
$380 and $2,900
$440 and $3,200
$340
$s

$710 and $3,200
$770 and $3,500
PV
2021-2040
$4,800 and $37,000
$3,200 and $25,000
$4,400
$370
$260
$8,800 and $41,000
$7,300 and $29,000
EAV
2021 - 2040
$320 and $2,500
$300 and $2,400
$290
$25
$25
$590 and $2,800
$570 and $2,700
aRows may not appear to add correctly due to rounding. The two benefits estimates are separated by the word "and" to signify that they are two separate
estimates. The estimates do not represent lower- and upper-bound estimates and should not be summed.
b The annualized present value of costs and benefits are calculated over a 20-year period from 2021 to 2040.
0 Climate benefits are based on changes (reductions) in CO2 emissions and are calculated using four different estimates of the social cost of carbon (SC-CO2)
(model average at 2.5 percent, 3 percent, and 5 percent discount rates; 95th percentile at 3 percent discount rate). For purposes of this table, we show the benefits
7-6
-------
associated with the model average at a 3 percent discount rate. However, we emphasize the importance and value of considering the benefits calculated using all
four SC-CO2 estimates. As discussed in Chapter 5, a consideration of climate benefits calculated using discount rates below 3 percent, including 2 percent and
lower, are also warranted when discounting intergenerational impacts.
d The costs presented in this table are consistent with the costs presented in Chapter 4, Table 4-6. To estimate these annualized costs,
EPA uses a conventional and widely accepted approach that applies a capital recovery factor (CRF) multiplier to capital investments
and adds that to the annual incremental operating expenses. Annual costs were calculated using a 4.25% real discount rate consistent with the rate used in IPM's
objective function for cost-minimization.
*Year in which air quality was simulated. Ozone air quality was simulated in 2021 and 2024 while the formation of PM2.5 was simulated only in 2024. Health
benefits for all other years were linearly extrapolated or interpolated from model-simulated air quality in these years. This method assumes that ozone and PM2 5
formation reaches a steady state beyond 2024 and may create increasing uncertainty in the benefits estimates the farther into the future estimates are extrapolated.
Benefits calculated as value of avoided: PM2 5-attributable deaths (quantified using a concentration-response relationship from the Di et al. 2017 study); Ozone-
attributable deaths (quantified using a concentration-response relationship from the Turner et al. 2017 study); and PM2 5 and ozone-related morbidity effects.
7-7
-------
United States	Office of Air Quality Planning and Standards	Publication No. EPA-452/R-21-002
Environmental Protection	Health and Environmental Impacts Division	March 2021
Agency	Research Triangle Park, NC
-------